file
stringlengths 6
44
| content
stringlengths 38
162k
|
|---|---|
openai_embedding.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import openai
import tiktoken
import numpy as np
import pandas as pd
from pathlib import Path
from openai.embeddings_utils import get_embedding
from sklearn.base import BaseEstimator, TransformerMixin
class embedding(BaseEstimator, TransformerMixin):
def __init__(self, embedding_engine='Text-Embedding', embedding_ctx_size=8191, encoding_method='cl100k_base'):
self.embedding_engine = embedding_engine
self.embedding_ctx_size = embedding_ctx_size
self.encoding_method = encoding_method
self.number_of_features = 1536
def fit(self,X,y=None):
return self
def transform(self, X):
setup_openai()
X = map(lambda text: self.len_safe_get_embedding( text), X)
return list(X)
def split_large_text(self, large_text):
encoding = tiktoken.get_encoding( self.encoding_method)
tokenized_text = encoding.encode(large_text)
chunks = []
current_chunk = []
current_length = 0
for token in tokenized_text:
current_chunk.append(token)
current_length += 1
if current_length >= self.embedding_ctx_size:
chunks.append(encoding.decode(current_chunk).rstrip(' .,;'))
current_chunk = []
current_length = 0
if current_chunk:
chunks.append(encoding.decode(current_chunk).rstrip(' .,;'))
return chunks
def len_safe_get_embedding(self, text):
chunk_embeddings = []
chunk_lens = []
for chunk in self.split_large_text(text):
chunk_embeddings.append( get_embedding(chunk, engine=self.embedding_engine))
chunk_lens.append(len(chunk))
chunk_embeddings = np.average(chunk_embeddings, axis=0, weights=None)
chunk_embeddings = chunk_embeddings / np.linalg.norm(chunk_embeddings) # normalizes length to 1
chunk_embeddings = chunk_embeddings.tolist()
return chunk_embeddings
def get_feature_names_out(self):
return [str(x) for x in range(self.number_of_features)]
def get_feature_names(self):
return self.get_feature_names_out()
"""
Open AI initialization has to be done separately as follows:
1. During training read the parameters from user
a. from config
b. SQLite database
c. From Json file
"""
class setup_openai():
def __init__( self, config=None):
param_keys = ['api_type','api_key','api_base','api_version']
if isinstance(config, dict):
valid_params = {x:y for x,y in config.items() if x in param_keys}
self._update_params(valid_params)
elif self._is_sqlite():
self._update_params( self._get_cred_from_sqlite())
elif ((Path(__file__).parent.parent/'etc')/'openai.json').exists():
with open(((Path(__file__).parent.parent/'etc')/'openai.json'), 'r') as f:
import json
params = json.load(f)
valid_params = {x:y for x,y in params.items() if x in param_keys}
self._update_params(valid_params)
else:
raise ValueError('Open AI credentials are not provided.')
def _is_sqlite(self):
try:
from AION.appbe.sqliteUtility import sqlite_db
from AION.appbe.dataPath import DATA_DIR
db_dir = Path(DATA_DIR)/'sqlite'
db_file = 'config.db'
if (db_dir/db_file).exists():
sqlite_obj = sqlite_db(db_dir,db_file)
if sqlite_obj.table_exists('openai'):
return True
return False
except:
return False
def _get_cred_from_sqlite(self):
from AION.appbe.sqliteUtility import sqlite_db
from AION.appbe.dataPath import DATA_DIR
db_dir = Path(DATA_DIR)/'sqlite'
db_file = 'config.db'
sqlite_obj = sqlite_db(db_dir,db_file)
data = sqlite_obj.read_data('openai')[0]
param_keys = ['api_type','api_key','api_base','api_version']
return dict((x,y) for x,y in zip(param_keys,data))
def _update_params(self, valid_params):
for key, value in valid_params.items():
if key == 'api_type':
openai.api_type = value
elif key == 'api_key':
openai.api_key = value
elif key == 'api_base':
openai.api_base = value
elif key == 'api_version':
openai.api_version = value
|
eda.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import sys
import logging
from collections import Counter
import spacy
import numpy as np
import pandas as pd
import nltk
from nltk.corpus import stopwords
from nltk import pos_tag
from nltk.tokenize import word_tokenize
from nltk.stem.wordnet import WordNetLemmatizer
from textblob import TextBlob
from sklearn.feature_extraction.text import CountVectorizer
'''
nltk.download("punkt")
nltk.download("wordnet")
'''
stopWords = stopwords.words("english")
class ExploreTextData:
def __init__(self, logEnabled=False):
self.logEnabled = logEnabled
def __Log(self, logType="info", text=None):
if logType.lower() == "exception":
logging.exception( text)
elif self.logEnabled:
if logType.lower() == "info":
logging.info( text)
elif logType.lower() == "debug":
logging.debug( text)
def Describe(self, inputCorpus):
""" Generate descriptive statistics for length of documents.
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
dict
Summary statistics of the Series or Dataframe provided.
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
stat = {}
word_count = self.DocumentWordCount(inputCorpus)
stat['count'] = float(len(word_count))
stat['mean'] = float(word_count.mean())
stat['std'] = float(word_count.std())
stat['max'] = float(word_count.max())
stat['min'] = float(word_count.min())
return pd.DataFrame.from_dict(stat, orient='index')
except:
self.__Log("exception", sys.exc_info())
raise
def DocumentLength(self, inputCorpus):
""" Calculate the length of each document in corpus
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
pandas.Series of {int}
series of length of documents
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
return inputCorpus.str.len()
except:
self.__Log("exception", sys.exc_info())
raise
def DocumentWordCount(self, inputCorpus):
""" Calculate the number of words in each document in corpus
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
pandas.Series of {int}
series of number of words in documents
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
return inputCorpus.str.split().map(lambda x: len(x))
except:
self.__Log("exception", sys.exc_info())
raise
def AverageWordLength(self, inputCorpus):
""" Calculate the average length of words in each document in corpus
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
pandas.Series of {double}
series of average length of words in documents
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
return inputCorpus.str.split()\
.apply(lambda x: [len(i) for i in x])\
.map(lambda x: np.mean(x))
except:
self.__Log("exception", sys.exc_info())
raise
def StopWordsCount(self, inputCorpus):
""" Calculate the number of stopwords in each document in corpus
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
pandas.Series of {int}
series of count of stopwords in documents
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
stopWordsCount = []
inputCorpus = list(inputCorpus)
for doc in inputCorpus:
count = 0
for word in doc.split():
if word in stopWords:
count += 1
stopWordsCount.append(count)
return pd.Series(stopWordsCount)
except:
self.__Log("exception", sys.exc_info())
raise
def MostCommonWords(self, inputCorpus, num_of_words=40):
""" get the most common words in corpus
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
Pandas.DataFrame{string, int}
Dataframe with columns "most_common_words" and "freq"
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
new = inputCorpus.str.split()
new = new.values.tolist()
corpus = [word for i in new for word in i if word not in stopWords]
counter = Counter(corpus)
most = counter.most_common()
x, y = [], []
for word, count in most[: num_of_words + 1]:
x.append(word)
y.append(count)
return pd.DataFrame([x, y],index=['most_common_words', 'freq']).T
except:
self.__Log("exception", sys.exc_info())
raise
def NullCount(self, inputCorpus):
""" Calculate the number of null entries in corpus
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
int
count of null entries in corpus
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
return pd.Series(inputCorpus.isnull().sum())
except:
self.__Log("exception", sys.exc_info())
raise
def TopNgram(self, inputCorpus, ngram, num_of_words=10):
""" Get the top words from the ngrams
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
ngram: int
ngram required
num_of_words:int, optional
numbers of words to be returned
Returns
-------
Pandas.DataFrame{string, int}
Dataframe with columns "ngram_words" and "freq"
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
words = []
for doc in inputCorpus:
word = [w for w in word_tokenize(doc) if (w not in stopWords)]
words.append(" ".join(word))
vec = CountVectorizer(ngram_range=(ngram, ngram)).fit(words)
bag_of_words = vec.transform(inputCorpus)
sum_words = bag_of_words.sum(axis=0)
words_freq = [(word, sum_words[0, idx]) for word, idx in vec.vocabulary_.items()]
words_freq = sorted(words_freq, key=lambda x: x[1], reverse=True)[:num_of_words]
words = []
frequency = []
for word, freq in words_freq:
words.append(word)
frequency.append(freq)
return pd.DataFrame([words, frequency],index=['ngram_words', 'freq']).T
except:
self.__Log("exception", sys.exc_info())
raise
def Polarity(self, inputCorpus):
""" Get the polarity of the text
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
pandas.Series {double}
series of calculated polarity of the documents
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
return inputCorpus.apply(lambda x: TextBlob(x).sentiment.polarity)
except:
self.__Log("exception", sys.exc_info())
raise
def ReadabilityScore(self, inputCorpus):
""" Get the Readability Score of the text
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
pandas.Series {double}
series of calculated Readability Score of the documents
"""
import textstat
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
if isinstance(inputCorpus, pd.Series):
return pd.Series([textstat.flesch_reading_ease(text) for text in inputCorpus])
else:
return [textstat.flesch_reading_ease(inputCorpus)]
except:
self.__Log("exception", sys.exc_info())
raise
def TagEntityCount(self, inputCorpus):
""" Calculate the frequency of each entity present in documents
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
Pandas.DataFrame{string, int}
Dataframe with columns "entity" and "freq"
"""
def ner(text):
doc = nlp(text)
return [X.label_ for X in doc.ents]
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
nlp = spacy.load("en_core_web_sm")
ent = inputCorpus.apply(lambda x: ner(x))
ent = [x for sub in ent for x in sub]
counter = Counter(ent)
count = counter.most_common()
x, y = map(list, zip(*count))
return pd.DataFrame([x, y],index=['entity', 'freq']).T
except:
self.__Log("exception", sys.exc_info())
raise
def MostCommonTokenPerEntity(self, inputCorpus, entity="GPE"):
""" Get the frequency of most common words corresponding to the specified entity in documents
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
entity: string, optional
name of the entity corresponding to which words are counted
Returns
-------
Pandas.DataFrame{string, int}
Dataframe with columns "token" and "freq"
"""
def ner(text, ent):
doc = nlp(text)
return [X.text for X in doc.ents if X.label_ == ent]
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
nlp = spacy.load("en_core_web_sm")
gpe = inputCorpus.apply(lambda x: ner(x, entity.upper()))
gpe = [i for x in gpe for i in x]
counter = Counter(gpe)
x, y = map(list, zip(*counter.most_common(10)))
return pd.DataFrame([x, y],index=['token', 'freq']).T
except:
self.__Log("exception", sys.exc_info())
raise
def MostCommonPosTag(self, inputCorpus):
""" Get the frequency of most common POS tag present in documents
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
Returns
-------
Pandas.DataFrame{string, int}
Dataframe with columns "postag" and "freq"
"""
def pos(text):
pos = pos_tag(word_tokenize(text))
pos = list(map(list, zip(*pos)))[1]
return pos
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
tags = inputCorpus.apply(lambda x: pos(x))
tags = [x for l in tags for x in l]
counter = Counter(tags)
x, y = list(map(list, zip(*counter.most_common(7))))
return pd.DataFrame([x, y],index=['postag', 'freq']).T
except:
self.__Log("exception", sys.exc_info())
raise
def MostCommonWordsInPOSTag(self, inputCorpus, tag="NN"):
""" Get the frequency of most common words related to specified POS tag present in documents
Parameters
----------
inputCorpus: sequence of input documents where each document consists of paragraphs or sentences
tag: string, optional
POS tag corresponding to which words frequency will be calculated
Returns
-------
Pandas.DataFrame{string, int}
Dataframe with columns "words" and "freq"
"""
def get_POSTag(text, tag):
adj = []
pos = pos_tag(word_tokenize(text))
for word, tg in pos:
if tg == tag:
adj.append(word)
return adj
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
words = inputCorpus.apply(lambda x: get_POSTag(x, tag.upper()))
words = [x for l in words for x in l]
counter = Counter(words)
x = []
y = []
if len(counter):
x, y = list(map(list, zip(*counter.most_common(7))))
return pd.DataFrame([x, y],index=['words', 'freq']).T
except:
self.__Log("exception", sys.exc_info())
raise
def __preprocessData(self, inputCorpus):
""" Prepare the data for topic modelling
"""
try:
self.__Log("info", "Start of {} function".format(sys._getframe().f_code.co_name))
corpus = []
lem = WordNetLemmatizer()
for doc in inputCorpus:
words = [w for w in word_tokenize(doc) if (w not in stopWords)]
words = [lem.lemmatize(w) for w in words if len(w) > 2]
corpus.append(words)
return corpus
except:
self.__Log("exception", sys.exc_info())
raise |
TextCleaning.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import re
import string
import sys
import demoji
#demoji.download_codes()
import nltk
import spacy
from nltk.corpus import stopwords
from bs4 import BeautifulSoup
from text_unidecode import unidecode
from textblob import TextBlob
from spellchecker import SpellChecker
from nltk import pos_tag
from nltk.tokenize import word_tokenize
from nltk.corpus import wordnet
from nltk.stem.wordnet import WordNetLemmatizer
from nltk.stem.porter import PorterStemmer
from spacy.lang.en import English
from collections import defaultdict
import contractions
spacy_nlp = None
def WordTokenize( inputText, tokenizationLib = 'nltk'):
tokenList = []
if inputText is not None and inputText != "":
tokenizationLib = tokenizationLib.lower()
if tokenizationLib == 'nltk':
tokenList = word_tokenize(inputText)
elif tokenizationLib == 'textblob':
tbObj = TextBlob(inputText)
tokenList = tbObj.words
elif tokenizationLib == 'spacy':
nlp = English()
nlpDoc = nlp(inputText)
for token in nlpDoc:
tokenList.append(token.text)
elif tokenizationLib == 'keras':
from tensorflow.keras.preprocessing.text import text_to_word_sequence
tokenList = text_to_word_sequence(inputText)
else:
tokenList = word_tokenize(inputText)
return tokenList
def SentenceTokenize( inputText):
sentenceList = []
if inputText is not None and inputText != "":
sentenceList = sent_tokenize(inputText)
return sentenceList
def Lemmatize(inputTokensList, lemmatizationLib = 'nltk'):
global spacy_nlp
lemmatized_list= []
lemmatizationLib = lemmatizationLib.lower()
if (inputTokensList is not None) and (len(inputTokensList)!=0):
if (lemmatizationLib == 'textblob'):
inputText = " ".join(inputTokensList)
sent = TextBlob(inputText)
tag_dict = {"J": 'a',
"N": 'n',
"V": 'v',
"R": 'r'}
words_and_tags = [(w, tag_dict.get(pos[0], 'n')) for w, pos in sent.tags]
lemmatized_list = [wd.lemmatize(tag) for wd, tag in words_and_tags]
if (lemmatizationLib == 'spacy'):
inputText = " ".join(inputTokensList)
if spacy_nlp == None:
spacy_nlp = spacy.load('en_core_web_sm')
doc = spacy_nlp(inputText)
for token in doc:
if token.text != token.lemma_:
if token.lemma_ != "-PRON-":
lemmatized_list.append(token.lemma_)
else:
lemmatized_list.append(token.text)
else:
lemmatized_list.append(token.text)
else:
tag_map = defaultdict(lambda : wordnet.NOUN)
tag_map['J'] = wordnet.ADJ
tag_map['V'] = wordnet.VERB
tag_map['R'] = wordnet.ADV
wnLemmatizer = WordNetLemmatizer()
token_tags = pos_tag(inputTokensList)
lemmatized_list = [wnLemmatizer.lemmatize(token, tag_map[tag[0]]) for token, tag in token_tags]
return lemmatized_list
def Stemmize(inputTokensList):
stemmedTokensList= []
if (inputTokensList is not None) and (len(inputTokensList)!=0):
porterStemmer = PorterStemmer()
stemmedTokensList = [porterStemmer.stem(token) for token in inputTokensList]
return stemmedTokensList
def ToLowercase(inputText):
resultText = ""
if inputText is not None and inputText != "":
resultText = inputText.lower()
return resultText
def ToUppercase(inputText):
resultText = ""
if inputText is not None and inputText != '':
resultText = inputText.upper()
return resultText
def RemoveNoise(inputText,
removeNoise_fHtmlDecode = True,
removeNoise_fRemoveHyperLinks = True,
removeNoise_fRemoveMentions = True,
removeNoise_fRemoveHashtags = True,
removeNoise_RemoveOrReplaceEmoji = 'remove',
removeNoise_fUnicodeToAscii = True,
removeNoise_fRemoveNonAscii = True):
if inputText is not None and inputText != "":
if removeNoise_fHtmlDecode == True:
inputText = BeautifulSoup(inputText, "html.parser").text
if removeNoise_fRemoveHyperLinks == True:
inputText = re.sub(r'https?:\/\/\S*', '', inputText, flags=re.MULTILINE)
if removeNoise_fRemoveMentions == True:
inputText = re.sub('[@]+\S+','', inputText)
if removeNoise_fRemoveHashtags == True:
inputText = re.sub('[#]+\S+','', inputText)
if removeNoise_RemoveOrReplaceEmoji == 'remove':
inputText = demoji.replace(inputText, "")
elif removeNoise_RemoveOrReplaceEmoji == 'replace':
inputText = demoji.replace_with_desc(inputText, " ")
if removeNoise_fUnicodeToAscii == True:
inputText = unidecode(inputText)
if removeNoise_fRemoveNonAscii == True:
inputText= re.sub(r'[^\x00-\x7F]+',' ', inputText)
inputText = re.sub(r'\s+', ' ', inputText)
inputText = inputText.strip()
return inputText
def RemoveStopwords(inputTokensList, stopwordsRemovalLib='nltk', stopwordsList = None, extend_or_replace='extend'):
resultTokensList = []
if (inputTokensList is not None) and (len(inputTokensList)!=0):
stopwordsRemovalLib= stopwordsRemovalLib.lower()
if stopwordsRemovalLib == 'spacy':
nlp = English()
stopwordRemovalList = nlp.Defaults.stop_words
else:
stopwordRemovalList = set(stopwords.words('english'))
if extend_or_replace == 'replace':
if stopwordsList is not None:
stopwordRemovalList = set(stopwordsList)
else:
if stopwordsList:
stopwordRemovalList = stopwordRemovalList.union(set(stopwordsList))
resultTokensList = [word for word in inputTokensList if word not in stopwordRemovalList]
return resultTokensList
def RemoveNumericTokens(inputText, removeNumeric_fIncludeSpecialCharacters=True):
resultText = ""
if inputText is not None and inputText != "":
if removeNumeric_fIncludeSpecialCharacters == True:
#Remove tokens having numbers and punctuations
resultText = re.sub(r'\b\d+[^a-zA-Z]*\d*\b',' ', inputText)
else:
#Remove only numeric tokens
resultText = re.sub(r'\b\d+\b','', inputText)
# convert consecutive whitespaces to single space in the results
resultText = re.sub(r'\s+', ' ', resultText)
return resultText
def RemovePunctuation(inputText, fRemovePuncWithinTokens=False):
resultText = ""
if inputText is not None and len(inputText) != 0:
if fRemovePuncWithinTokens == True:
resultText = inputText.translate(str.maketrans("","", string.punctuation))
else:
punctuationList = list(string.punctuation)
tokensList = WordTokenize(inputText)
resultTokensList = [word for word in tokensList if word not in punctuationList]
resultText = " ".join(resultTokensList)
resultText = re.sub(r'\s+', ' ', resultText)
return resultText
def CorrectSpelling(inputTokensList):
correctedTokensList = []
if (inputTokensList is not None) and (len(inputTokensList)!=0):
spell = SpellChecker()
for word in inputTokensList:
word = word.lower()
if word not in spell:
word = spell.correction(word)
if word:
correctedTokensList.append(word)
return correctedTokensList
def ReplaceAcronym(inputTokensList, acrDict=None):
resultTokensList = []
if (inputTokensList is not None) and (len(inputTokensList)!=0):
if ((acrDict is not None) and (len(acrDict) != 0)):
acrDictLowercase = dict((key.lower(), value.lower()) for key, value in acrDict.items())
resultTokensList = [acrDictLowercase.get(token.lower(), token.lower()) for token in inputTokensList]
else:
resultTokensList = inputTokensList
return resultTokensList
def ExpandContractions(inputText):
resultText = ""
if inputText != '':
resultText = contractions.fix(inputText)
return resultText
def cleanText( inputText,
functionSequence = ['RemoveNoise','ExpandContractions','Normalize','ReplaceAcronym',
'CorrectSpelling','RemoveStopwords','RemovePunctuation','RemoveNumericTokens'],
fRemoveNoise = True,
fExpandContractions = False,
fNormalize = True,
fReplaceAcronym = False,
fCorrectSpelling = False,
fRemoveStopwords = True,
fRemovePunctuation = True,
fRemoveNumericTokens = True,
removeNoise_fHtmlDecode = True,
removeNoise_fRemoveHyperLinks = True,
removeNoise_fRemoveMentions = True,
removeNoise_fRemoveHashtags = True,
removeNoise_RemoveOrReplaceEmoji = 'remove',
removeNoise_fUnicodeToAscii = True,
removeNoise_fRemoveNonAscii = True,
tokenizationLib='nltk',
normalizationMethod = 'Lemmatization',
lemmatizationLib = 'nltk',
acronymDict = None,
stopwordsRemovalLib = 'nltk',
stopwordsList = None,
extend_or_replace_stopwordslist = 'extend',
removeNumeric_fIncludeSpecialCharacters = True,
fRemovePuncWithinTokens = False
):
if inputText is not None and inputText != "":
for function in functionSequence:
if function == 'RemoveNoise':
if (fRemoveNoise == True):
inputText = RemoveNoise(inputText,
removeNoise_fHtmlDecode,
removeNoise_fRemoveHyperLinks,
removeNoise_fRemoveMentions,
removeNoise_fRemoveHashtags,
removeNoise_RemoveOrReplaceEmoji,
removeNoise_fUnicodeToAscii,
removeNoise_fRemoveNonAscii)
if function == 'ExpandContractions':
if (fExpandContractions == True):
inputText = ExpandContractions(inputText)
if function == 'Normalize':
if (fNormalize == True):
inputTokens = WordTokenize(inputText, tokenizationLib)
if (normalizationMethod == 'Stemming'):
inputTokens = Stemmize(inputTokens)
else:
inputTokens = Lemmatize(inputTokens, lemmatizationLib)
inputText = " ".join(inputTokens)
if function == 'ReplaceAcronym':
if fReplaceAcronym == True and (acronymDict is not None) and acronymDict != 'None':
inputText = ToLowercase(inputText)
inputTokens = WordTokenize(inputText, tokenizationLib)
inputTokens= ReplaceAcronym(inputTokens, acronymDict)
inputText = " ".join(inputTokens)
if function == 'CorrectSpelling':
if (fCorrectSpelling == True):
try:
inputTokens = WordTokenize(inputText, tokenizationLib)
inputTokens = CorrectSpelling(inputTokens)
inputText = " ".join(inputTokens)
except Exception as e:
print(e)
pass
if function == 'RemoveStopwords':
if (fRemoveStopwords == True):
inputText = ToLowercase(inputText)
inputTokens = WordTokenize(inputText, tokenizationLib)
inputTokens = RemoveStopwords(inputTokens, stopwordsRemovalLib, stopwordsList, extend_or_replace_stopwordslist)
inputText = " ".join(inputTokens)
if function == 'RemovePunctuation':
if (fRemovePunctuation == True):
inputText = RemovePunctuation(inputText, fRemovePuncWithinTokens)
if function == 'RemoveNumericTokens':
if (fRemoveNumericTokens == True):
inputText = RemoveNumericTokens(inputText, removeNumeric_fIncludeSpecialCharacters)
inputText = ToLowercase(inputText)
return inputText
|
timeseries.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import json
#Python sklearn & std libraries
import numpy as np
import pandas as pd
from time_series.ts_arima_eion import eion_arima
from time_series.aion_fbprophet import aion_fbprophet
from time_series.timeseriesDLUnivariate import timeseriesDLUnivariate
from time_series.timeseriesDLMultivariate import timeseriesDLMultivariate
from time_series.tsDLMultiVrtInUniVrtOut import tsDLMultiVrtInUniVrtOut
from statsmodels.tsa.vector_ar.vecm import coint_johansen
from statsmodels.tsa.vector_ar.var_model import VAR
from math import *
from sklearn.metrics import mean_squared_error
from sklearn.metrics import mean_absolute_error
from math import sqrt
import logging
import os
import sys
import time
import pickle
from statsmodels.tsa.arima_model import ARIMA
from sklearn.metrics import mean_squared_error
from statsmodels.tsa.stattools import adfuller
import pmdarima as pm
from statsmodels.tsa.stattools import grangercausalitytests
from statsmodels.stats.stattools import durbin_watson
from time_series.ts_modelvalidation import timeseriesModelTests
from sklearn.utils import check_array
from time_series.tsStationarySeasonalityTest import tsStationarySeasonalityTest
class timeseries():
def __init__(self,tsConfig,modelconfig,modelList,data,targetFeature,dateTimeFeature,modelName,trainPercentage,usecasename,version,deployLocation,scoreParam):
self.tsConfig = tsConfig
self.modelconfig = modelconfig
self.modelList = modelList
self.data = data
self.data1=data
self.pred_freq = ''
self.additional_regressors=''
self.trainPercentage = trainPercentage
self.targetFeature = targetFeature
self.dateTimeFeature = dateTimeFeature
self.modelName=modelName
self.usecasename=usecasename
self.model_fit=None
self.selectedColumns = ''
self.version=version
self.deployLocation=deployLocation
self.dictDiffCount={}
self.log = logging.getLogger('eion')
self.scoreParam=str(scoreParam)
try:
##For bug:12280
self.data.dropna(how='all',axis=1,inplace=True)
except Exception as e:
self.data.fillna(0)
self.log.info("data empty feature process error info:, check any text column contain empty records. if yes, please remove the column and upload the data for time series forecasting. \n"+str(e))
def var_prediction(self,no_of_prediction):
tdata = self.data.drop([self.dateTimeFeature], axis=1)
tdata.index = self.data[self.dateTimeFeature]
lag_order = self.model_fit.k_ar
predictions = self.model_fit.forecast(tdata.values[-lag_order:],steps=no_of_prediction)
predictions = predictions.round(2)
col = self.targetFeature.split(",")
pred = pd.DataFrame(index=range(0,len(predictions)),columns=col)
for j in range(0,len(col)):
for i in range(0, len(predictions)):
pred.iloc[i][j] = predictions[i][j]
predictions = pred
pred=self.invertTransformation(tdata,self.targetFeature,predictions,self.dictDiffCount)
return pred
def save_dl_model(self,smodel,scaler_model):
try:
saved_model = self.usecasename+'_'+self.version
filename = os.path.join(self.deployLocation,'model',saved_model)
smodel.save(filename)
if scaler_model != 'NA' and scaler_model != '':
scaler_filename = os.path.join(self.deployLocation,'model',saved_model+'_scaler.pkl')
with open(scaler_filename, 'wb') as f:
pickle.dump(scaler_model,f)
f.close()
else:
scaler_filename = 'NA'
return filename,saved_model,scaler_filename
except Exception as e:
print(e)
def save_model(self,smodel):
try:
saved_model = self.usecasename+'_'+self.version+'.sav'
filename = os.path.join(self.deployLocation,'model',saved_model)
with open(filename, 'wb') as f:
pickle.dump(smodel,f)
f.close()
return filename,saved_model
except Exception as e:
print(e)
def mean_absolute_percentage_error(self,y_true, y_pred):
try:
y_true, y_pred=np.array(y_true), np.array(y_pred)
mape=np.mean(np.abs((y_true - y_pred) / y_true+sys.float_info.epsilon)) * 100
return mape
except Exception as inst:
self.log.info('------------- mean_absolute_percentage_error ---------------')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
## Fbprophet model
def getfbprophetmodel(self,predicted_data_file,dataFolderLocation,tFeature):
try:
modelName='fbprophet'
modelconfig = self.modelconfig['fbprophet']
self.targetFeature=tFeature[0]
X_Train = pd.DataFrame(self.data[self.targetFeature])
try:
# self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce')
##For bug 13513 - If the datetime needs UTC timestamp process, except part will handle.
try:
#for non utc timestamp
self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce')
except:
#for utc timestamp
self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce',utc=True)
self.data = self.data.dropna()
except:
pass
aion_prophet_obj = aion_fbprophet(modelconfig,self.trainPercentage,self.data,self.targetFeature,self.dateTimeFeature)
self.log.info('Status:- |... TimeSeries Algorithm applied: FBPROPHET')
self.model_fit,mae,rmse_prophet,mse,mape,r2,pred_freq,additional_regressors,prophet_df_new = aion_prophet_obj.aion_probhet(X_Train,self.dateTimeFeature,predicted_data_file,dataFolderLocation)
## Added for additional scoring params
if (self.scoreParam.lower() == "r2"):
scoringparam_v=r2
self.log.info("fbprophet User selected scoring parameter is r2. r2 value: "+str(r2))
elif (self.scoreParam.lower() == "rmse"):
scoringparam_v=rmse_prophet
self.log.info("fbprophet User selected scoring parameter is RMSE. RMSE value: "+str(rmse_prophet))
elif (self.scoreParam.lower() == "mse"):
scoringparam_v=mse
self.log.info("fbprophet User selected scoring parameter is MSE. MSE value: "+str(mse))
elif (self.scoreParam.lower() == "mae"):
scoringparam_v=mae
self.log.info("fbprophet User selected scoring parameter is MAE. MAE value: "+str(mae))
else:
scoringparam_v=rmse_prophet
self.log.info('Status:- |... Score '+self.scoreParam.capitalize()+': '+str(round(scoringparam_v,2))) #task 11997 displaying user selected scoring parameter in status logs
error_matrix = '"RMSE":"'+str(round(rmse_prophet,2))+'","MAPE":"'+str(round(mape,2))+'","R2":"'+str(round(r2,2))+'","MAE":"'+str(round(mae,2))+'","MSE":"'+str(round(mse,2))+'"'
self.log.info("fbprophet all scoring parameter results: "+str(error_matrix))
scoredetails = '{"Model":"FBProphet ","Score":'+str(scoringparam_v)+',"Scoring Param": "'+str(self.scoreParam.lower())+'"}'
self.selectedColumns = self.targetFeature+','+self.dateTimeFeature
self.selectedColumns = self.selectedColumns.split(",")
self.pred_freq = pred_freq
self.additional_regressors=additional_regressors
self.log.info('------------- End FBPROPHET Model -------------\n')
return('Success',modelName.upper(),self.scoreParam.lower(),scoringparam_v,self.model_fit,self.selectedColumns,error_matrix,scoredetails,self.dictDiffCount,self.pred_freq,self.additional_regressors,prophet_df_new)
except Exception as e:
self.log.info("FBProphet operation failed. error: "+str(e))
return('Error',modelName.upper(),self.scoreParam.lower(),0,None,self.selectedColumns,'','{}',self.dictDiffCount,self.pred_freq,self.additional_regressors,pd.DataFrame())
## Arima model
def get_arima_values(self):
try:
tFeature = self.targetFeature.split(',')
if(len(tFeature) == 1):
model_name = 'arima'
else:
self.log.info("Note: ARIMA model is going to perform only on first feature of provided target features due to data not met the VAR model constraints")
self.targetFeature=tFeature[0]
sesonalityChecks=True
stationaryChecks=False
#start checking sessonality using ch test and ocsb
self.log.info(self.data.head(5))
res = pm.arima.nsdiffs(self.data[self.targetFeature], m=355, max_D=5, test="ch") # 365 since daily
self.log.info('-------> Seasonality checks: %f' % res)
if res >=4:
self.log.info("-----------> Data is following Seasonality ")
self.log.info('Status:- |... Seasonality Check Done. Data is following Seasonality ')
sesonalityChecks=True
else:
self.log.info("-----------> Data is not following Seasonality ")
self.log.info('Status:- |... Seasonality Check Done. Data is not following Seasonality')
sesonalityChecks=False
# end checking sessonality using ch test and ocsb
# start checking stationary data for time Series
series=self.data[self.targetFeature]
adf_test = pm.arima.ADFTest(alpha=0.05)
resultSt = adfuller(self.data[self.targetFeature])
self.log.info('ADF Statistic: %f' % resultSt[0])
self.log.info('p-value: %f' % resultSt[1])
if resultSt[1]<= 0.05:
stationaryChecks=True
self.log.info("the data does not have a unit root and is stationary.")
self.log.info('Status:- |... Stationary Check Done. Data is stationary')
else:
stationaryChecks=False
self.log.info("the data has a unit root and is non-stationary.")
self.log.info('Status:- |... Stationary Check Done. Data is non-stationary')
# End of stationary checks
self.log.info('\n------------- Start Arima Model -------------')
self.log.info('-------> Top 5 Rows: ')
self.log.info(self.data.head(5))
eion_arima_obj = eion_arima(self.modelconfig['arima'],self.trainPercentage,sesonalityChecks,stationaryChecks)
return 'Success',eion_arima_obj
except Exception as e:
self.log.info('<!------------- Get ARIMA Values ---------------> '+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return 'Error',None
def getEncDecLSTMMultVrtInUniVrtOut(self):
try:
self.log.info('Status:- |... TimeSeries Algorithm applied: Encoder Decoder LSTM')
modelName='encoder_decoder_lstm_mvi_uvo'
modelconfig = self.modelconfig['encoder_decoder_lstm_mvi_uvo']
df = self.data
targetFeature = list(self.targetFeature.split(","))
try:
# df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
##For bug 13513 - If the datetime needs UTC timestamp process, except part will handle.
try:
#for non utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
except:
#for utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce',utc=True)
df = df.dropna()
except:
pass
df = df.groupby(self.dateTimeFeature).mean()
df = df.reset_index()
tdata = df.drop([self.dateTimeFeature], axis=1)
tdata.index = df[self.dateTimeFeature]
#tdata = tdata[tdata.columns[tdata.columns.isin(targetFeature)]]
#selectedColumns = self.targetFeature+','+self.dateTimeFeature
#selectedColumns = selectedColumns.split(",")
selectedColumns = tdata.columns
df_predicted=None
aion_dlts_obj = tsDLMultiVrtInUniVrtOut(modelconfig,self.trainPercentage,targetFeature,self.dateTimeFeature)
status,mse,rmse,r2,mae,model,df_predicted,lag_order,scaler = aion_dlts_obj.lstm_encdec_mvin_uvout(tdata)
if status.lower() == 'success':
## Added for additional scoring params
if (self.scoreParam.lower() == "r2"):
scoringparam_v=r2
self.log.info('Status:- |... Score R2(Avg) '+str(r2))
elif (self.scoreParam.lower() == "rmse"):
scoringparam_v=rmse
self.log.info("Status:- |... Score RMSE(Avg) "+str(rmse))
elif (self.scoreParam.lower() == "mse"):
scoringparam_v=mse
self.log.info("Status:- |... Score MSE(Avg) "+str(mse))
elif (self.scoreParam.lower() == "mae"):
scoringparam_v=mae
self.log.info("Status:- |... Score MAE(Avg) : "+str(mae))
else:
scoringparam_v=rmse
error_matrix = '"RMSE":"'+str(round(rmse,2))+'","MSE":"'+str(round(mse,2))+'"'
error_matrix=error_matrix+',"R2":"'+str(round(r2,2))+'","MAE":"'+str(round(mae,2))+'"'
self.log.info("LSTM Multivariant Input Univariate Output all scoring param results: "+str(error_matrix))
self.log.info('Status:- |... Score '+self.scoreParam.capitalize()+': '+str(round(scoringparam_v,2))) #task 11997 displaying user selected scoring parameter in status logs
scoredetails = '{"Model":"LSTM Multivariant","Score":'+str(scoringparam_v)+',"Scoring Param": "'+str(self.scoreParam.lower())+'"}'
else:
return 'Error',modelName.upper(),self.scoreParam.lower(),'NA',None,selectedColumns,'','{}',pd.DataFrame(),lag_order,None
except Exception as e:
self.log.info("getEncDecLSTMMultVrtInUniVrtOut method error. Error msg: "+str(e))
return 'Error',modelName.upper(),self.scoreParam.lower(),'NA',None,selectedColumns,'','{}',pd.DataFrame(),lag_order,None
return 'Success',modelName.upper(),self.scoreParam.lower(),scoringparam_v,model,selectedColumns,error_matrix,scoredetails,df_predicted,lag_order,scaler
def getLSTMMultivariate(self):
try:
self.log.info('Status:- |... TimeSeries Algorithm applied: LSTM')
modelName='lstm'
modelconfig = self.modelconfig['lstm']
df = self.data
targetFeature = list(self.targetFeature.split(","))
try:
# df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
##For bug 13513 - If the datetime needs UTC timestamp process, except part will handle.
try:
#for non utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
except:
#for utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce',utc=True)
df = df.dropna()
except:
pass
df = df.groupby(self.dateTimeFeature).mean()
df = df.reset_index()
tdata = df.drop([self.dateTimeFeature], axis=1)
tdata.index = df[self.dateTimeFeature]
tdata = tdata[tdata.columns[tdata.columns.isin(targetFeature)]]
selectedColumns = self.targetFeature+','+self.dateTimeFeature
selectedColumns = selectedColumns.split(",")
df_predicted=None
aion_dlts_obj = timeseriesDLMultivariate(modelconfig,self.trainPercentage,targetFeature,self.dateTimeFeature)
status,mse,rmse,r2,mae,model,df_predicted,lag_order,scaler = aion_dlts_obj.lstm_multivariate(tdata)
if status.lower() == 'success':
## Added for additional scoring params
if (self.scoreParam.lower() == "r2"):
scoringparam_v=r2
self.log.info('Status:- |... Score R2(Avg) '+str(r2))
elif (self.scoreParam.lower() == "rmse"):
scoringparam_v=rmse
self.log.info("Status:- |... Score RMSE(Avg) "+str(rmse))
elif (self.scoreParam.lower() == "mse"):
scoringparam_v=mse
self.log.info("Status:- |... Score MSE(Avg) "+str(mse))
elif (self.scoreParam.lower() == "mae"):
scoringparam_v=mae
self.log.info("Status:- |... Score MAE(Avg) : "+str(mae))
else:
scoringparam_v=rmse
error_matrix = '"RMSE":"'+str(round(rmse,2))+'","MSE":"'+str(round(mse,2))+'"'
error_matrix=error_matrix+',"R2":"'+str(round(r2,2))+'","MAE":"'+str(round(mae,2))+'"'
self.log.info("LSTM Multivariant all scoring param results: "+str(error_matrix))
self.log.info('Status:- |... Score '+self.scoreParam.capitalize()+': '+str(round(scoringparam_v,2))) #task 11997 displaying user selected scoring parameter in status logs
scoredetails = '{"Model":"LSTM Multivariant","Score":'+str(scoringparam_v)+',"Scoring Param": "'+str(self.scoreParam.lower())+'"}'
else:
return 'Error',modelName.upper(),self.scoreParam.lower(),'NA',None,selectedColumns,'','{}',pd.DataFrame(),lag_order,None
except Exception as e:
self.log.info("getLSTMMultivariate method error. Error msg: "+str(e))
return 'Error',modelName.upper(),self.scoreParam.lower(),'NA',None,selectedColumns,'','{}',pd.DataFrame(),lag_order,None
return 'Success',modelName.upper(),self.scoreParam.lower(),scoringparam_v,model,selectedColumns,error_matrix,scoredetails,df_predicted,lag_order,scaler
def getUniVarientLSTMModel(self):
try:
self.log.info('Status:- |... TimeSeries Algorithm applied: LSTM')
modelName='lstm'
lstmconfig = self.modelconfig['lstm']
df = self.data
try:
# df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
##For bug 13513 - If the datetime needs UTC timestamp process, except part will handle.
try:
#for non utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
except:
#for utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce',utc=True)
df = df.dropna()
except:
pass
tdata = df.drop([self.dateTimeFeature], axis=1)
tdata.index = df[self.dateTimeFeature]
tdata = pd.DataFrame(tdata[self.targetFeature])
selectedColumns = self.targetFeature+','+self.dateTimeFeature
selectedColumns = selectedColumns.split(",")
aion_dlts_obj = timeseriesDLUnivariate(lstmconfig,self.trainPercentage,self.targetFeature,self.dateTimeFeature,modelName)
status,lstm_mse,lstm_rmse,r2,mae,lstm_model,df_predicted_lstm,lag_order,scaler = aion_dlts_obj.ts_lstm(tdata)
if status.lower() == 'success':
## Added for additional scoring params
if (self.scoreParam.lower() == "r2"):
scoringparam_v=r2
self.log.info("LSTM Univariant User selected scoring parameter is r2. r2 value: "+str(r2))
elif (self.scoreParam.lower() == "rmse"):
scoringparam_v=lstm_rmse
self.log.info("LSTM Univariant User selected scoring parameter is RMSE. Rmse value: "+str(lstm_rmse))
elif (self.scoreParam.lower() == "mse"):
scoringparam_v=lstm_mse
self.log.info("LSTM Univariant User selected scoring parameter is MSE. Mse value: "+str(lstm_mse))
elif (self.scoreParam.lower() == "mae"):
scoringparam_v=mae
self.log.info("LSTM Univariant User selected scoring parameter is MAE. Mae value: "+str(mae))
else:
scoringparam_v=lstm_rmse
error_matrix = '"RMSE":"'+str(round(lstm_rmse,2))+'","MSE":"'+str(round(lstm_mse,2))+'"'
error_matrix=error_matrix+',"R2":"'+str(round(r2,2))+'","MAE":"'+str(round(mae,2))+'"'
self.log.info("LSTM Univariant, all scoring param results: "+str(error_matrix))
self.log.info('Status:- |... Score '+self.scoreParam.capitalize()+': '+str(round(scoringparam_v,2))) #task 11997 displaying user selected scoring parameter in status logs
scoredetails = '{"Model":"LSTM Univariant","Score":'+str(scoringparam_v)+',"Scoring Param": "'+str(self.scoreParam.lower())+'"}'
return 'Success',modelName.upper(),self.scoreParam.lower(),scoringparam_v,lstm_model,selectedColumns,error_matrix,scoredetails,df_predicted_lstm,lag_order,scaler
else:
return 'Error',modelName.upper(),self.scoreParam.lower(),0,None,selectedColumns,'','{}',pd.DataFrame(),0,None
except Exception as inst:
self.log.info('<!------------- LSTM Error ---------------> '+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return 'Error',modelName.upper(),self.scoreParam.lower(),0,None,selectedColumns,'','{}',pd.DataFrame(),0,None
def getUniVarientMLPModel(self):
try:
self.log.info('Status:- |... TimeSeries Algorithm applied: MLP')
modelName='mlp'
lstmconfig = self.modelconfig['mlp']
df = self.data
try:
# df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
##For bug 13513 - If the datetime needs UTC timestamp process, except part will handle.
try:
#for non utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce')
except:
#for utc timestamp
df[self.dateTimeFeature] = pd.to_datetime(df[self.dateTimeFeature],errors='coerce',utc=True)
df = df.dropna()
except:
pass
tdata = df.drop([self.dateTimeFeature], axis=1)
tdata.index = df[self.dateTimeFeature]
tdata = pd.DataFrame(tdata[self.targetFeature])
selectedColumns = self.targetFeature+','+self.dateTimeFeature
selectedColumns = selectedColumns.split(",")
aion_dlts_obj = timeseriesDLUnivariate(lstmconfig,self.trainPercentage,self.targetFeature,self.dateTimeFeature,modelName)
mlp_mse,mlp_rmse,r2,mae,mlp_model,df_predicted_mlp,look_back,scaler = aion_dlts_obj.mlpDL(tdata)
## Added for additional scoring params
if (self.scoreParam.lower() == "r2"):
scoringparam_v=r2
self.log.info("MLP Univariant User selected scoring parameter is R2. R2 value: "+str(r2))
elif (self.scoreParam.lower() == "rmse"):
scoringparam_v=mlp_rmse
self.log.info("MLP Univariant User selected scoring parameter is RMSE. Rmse value: "+str(mlp_rmse))
elif (self.scoreParam.lower() == "mse"):
scoringparam_v=mlp_mse
self.log.info("MLP Univariant User selected scoring parameter is MSE. Mse value: "+str(mlp_mse))
elif (self.scoreParam.lower() == "mae"):
scoringparam_v=mae
self.log.info("MLP Univariant User selected scoring parameter is MAE. Mae value: "+str(mae))
else:
scoringparam_v=mlp_rmse
error_matrix = '"RMSE":"'+str(round(mlp_rmse,2))+'","MSE":"'+str(round(mlp_mse,2))+'"'
error_matrix=error_matrix+',"R2":"'+str(round(r2,2))+'","MAE":"'+str(round(mae,2))+'"'
self.log.info("MLP Univariant, all scoring param results: "+str(error_matrix))
self.log.info('Status:- |... Score '+self.scoreParam.capitalize()+': '+str(round(scoringparam_v,2))) #task 11997 displaying user selected scoring parameter in status logs
scoredetails = '{"Model":"MLP","Score":'+str(scoringparam_v)+',"Scoring Param": "'+str(self.scoreParam.lower())+'"}'
return 'Success',modelName.upper(),self.scoreParam.lower(),scoringparam_v,mlp_model,selectedColumns,error_matrix,scoredetails,df_predicted_mlp,look_back,scaler
except Exception as inst:
import traceback
self.log.info("MLP Error in timeseries module: \n"+str(traceback.print_exc()))
self.log.info('<!------------- MLP Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return 'Error',modelName.upper(),self.scoreParam.lower(),0,None,selectedColumns,'','{}',pd.DataFrame(),0,None
def getARIMAmodel(self,predicted_data_file):
try:
modelName='arima'
status,eion_arima_obj = self.get_arima_values()
self.log.info('Status:- |... TimeSeries Algorithm applied: ARIMA')
selected_feature_list = self.data[self.targetFeature].values
selected_feature_list = selected_feature_list.astype('int32')
self.log.info('-------> Target Feature First 5 Rows: ')
self.log.info(self.data[self.targetFeature].head(5))
X_Train = pd.DataFrame(self.data[self.targetFeature])
try:
# self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce')
##For bug 13513 - If the datetime needs UTC timestamp process, except part will handle.
try:
#for non utc timestamp
self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce')
except:
#for utc timestamp
self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce',utc=True)
self.data = self.data.dropna()
except:
pass
if status.lower() == 'success':
self.model_fit,mae,rmse_arima,mse,r2,aic_score,mape,valid,pred = eion_arima_obj.eion_arima(X_Train)
## Added for additional scoring params
if (self.scoreParam.lower() == "r2"):
scoringparam_v=r2
self.log.info("ARIMA Univariant User selected scoring parameter is r2. r2 value: "+str(r2))
elif (self.scoreParam.lower() == "rmse"):
scoringparam_v=rmse_arima
self.log.info("ARIMA Univariant User selected scoring parameter is RMSE. RMSE value: "+str(rmse_arima))
elif (self.scoreParam.lower() == "mse"):
scoringparam_v=mse
self.log.info("ARIMA Univariant User selected scoring parameter is MSE. MSE value: "+str(mse))
elif (self.scoreParam.lower() == "mae"):
scoringparam_v=mae
self.log.info("ARIMA Univariant User selected scoring parameter is MAE. RMSE value: "+str(rmse_arima))
else:
scoringparam_v=rmse_arima
error_matrix = '"RMSE":"'+str(round(rmse_arima,2))+'","MSE":"'+str(round(mse,2))+'"'
error_matrix=error_matrix+',"R2":"'+str(round(r2,2))+'","MAE":"'+str(round(mae,2))+'"'
self.log.info('Status:- |... Score '+self.scoreParam.capitalize()+': '+str(round(scoringparam_v,2))) #task 11997 displaying user selected scoring parameter in status logs
self.log.info("ARIMA all scoring param results: "+str(error_matrix))
scoredetails = '{"Model":"ARIMA","Score":'+str(scoringparam_v)+',"Scoring Param": "'+str(self.scoreParam.lower())+'"}'
df_pred = pd.DataFrame()
df_pred[self.targetFeature+'_actual'] = valid[self.targetFeature]
df_pred[self.targetFeature+'_pred'] = pred
self.log.info('Status:- |... Score '+self.scoreParam.capitalize()+': '+str(round(scoringparam_v,2))) #task 11997 displaying user selected scoring parameter in status logs
self.log.info("ARIMA AIC Score: "+str(round(aic_score,2)))
selectedColumns = self.targetFeature+','+self.dateTimeFeature
selectedColumns = selectedColumns.split(",")
self.log.info('------------- End ARIMA Model -------------\n')
return('Success',modelName.upper(),self.scoreParam.lower(),scoringparam_v,self.model_fit,selectedColumns,error_matrix,scoredetails,self.dictDiffCount,self.pred_freq,self.additional_regressors,rmse_arima,df_pred)
else:
return 'Error',modelName.upper(),self.scoreParam.lower(),0,None,selectedColumns,'','{}',self.dictDiffCount,self.pred_freq,self.additional_regressors,0,pd.DataFrame()
except Exception as inst:
self.log.info('<!------------- Arima Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return 'Error',modelName.upper(),self.scoreParam.lower(),0,None,selectedColumns,'','{}',self.dictDiffCount,self.pred_freq,self.additional_regressors,0,pd.DataFrame()
## VAR model fn
def getVARmodel(self):
from sklearn.metrics import r2_score
modelName="var"
if modelName.lower()=='var':
try:
self.log.info('-------> Top 5 Rows: ')
self.log.info(self.data.head(5))
if(self.targetFeature != ''):
self.selectedColumns = self.targetFeature
self.selectedColumns = self.selectedColumns+','+self.dateTimeFeature
self.selectedColumns = self.selectedColumns.split(",")
self.data = self.data[self.selectedColumns]
try:
# self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce')
##For bug 13513 - If the datetime needs UTC timestamp process, except part will handle.
try:
#for non utc timestamp
self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce')
except:
#for utc timestamp
self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature],errors='coerce',utc=True)
self.data = self.data.dropna()
except:
pass
self.data = self.data.groupby(self.dateTimeFeature).mean()
self.data = self.data.reset_index()
tdata = self.data.drop([self.dateTimeFeature], axis=1)
tdata.index = self.data[self.dateTimeFeature]
cols = tdata.columns
self.log.info('------------- Start VAR Model -------------')
size = int(len(tdata) * (self.trainPercentage/100))
train = tdata.iloc[0:(int(size))]
valid = tdata.iloc[(int(size)):(int(len(tdata))-1)]
start = time.time()
model = VAR(endog=train)
self.log.info('Status:- |... TimeSeries Algorithm applied: VAR')
self.log.info("\n-------------Selecting max lag order -----------")
modelFitted = model.fit()
lag_order = modelFitted.k_ar
self.log.info('------->lag_order: '+str(lag_order))
executionTime=time.time() - start
self.log.info('-------> Time: '+str(executionTime)+'\n')
prediction = modelFitted.forecast(train.values[-lag_order:], steps=len(valid))
pred = pd.DataFrame(index=range(0,len(prediction)),columns=tdata.columns)
for j in range(0,(len(tdata.columns))):
for i in range(0, len(prediction)):
pred.iloc[i][j] = prediction[i][j]
self.log.info("\n--------------- Modal Validation Start ---------------")
df_pred = pd.DataFrame()
#check rmse
error_matrix = '"FeaturesMatrix":['
valid = valid.reset_index()
var_rmse = 0
mse_dict={}
rmse_dict={}
mae_dict={}
r2_dict={}
for i in tdata.columns:
if(error_matrix != '"FeaturesMatrix":['):
error_matrix = error_matrix+','
df_pred[i+'_actual'] = valid[i]
df_pred[i+'_pred'] = pred[i]
rmse_var = sqrt(mean_squared_error(valid[i],pred[i]))
mse = mean_squared_error(valid[i],pred[i])
mae = mean_absolute_error(valid[i],pred[i])
mape=self.mean_absolute_percentage_error(valid[i],pred[i])
r2 = r2_score(valid[i],pred[i])
mse_dict[i]=mse
rmse_dict[i]=rmse_var
r2_dict[i]=r2
mae_dict[i]=mae
var_rmse += round(rmse_var,2)
error_matrix += '{"Features":"'+i+'","MSE":"'+str(round(mse,2))+'","MAPE":"'+str(round(mape,2))+'","RMSE":"'+str(round(rmse_var,2))+'","MAE":"'+str(round(mae,2))+'"}'
self.log.info ("------->Feature: "+str(i))
self.log.info ("---------->MAE: "+str(mae))
self.log.info ("---------->MSE: "+str(mse))
self.log.info ("---------->RMSE :"+str(rmse_var))
self.log.info("------------->MAPE :"+str(mape))
self.log.info("------------->R2 :"+str(r2))
var_rmse = var_rmse/len(tdata.columns)
self.log.info("--------------- Modal Validation End ---------------\n")
pred=self.invertTransformation(train,self.targetFeature,pred,self.dictDiffCount)
self.log.info("-------> Predictions on Test Data")
self.log.info(pred.head(4))
model = VAR(endog=tdata)
self.model_fit = model.fit()
aic_score = self.model_fit.aic
lag_order = self.model_fit.k_ar
self.log.info('------->AIC Score: '+str(aic_score))
error_matrix += ']'
scoredetails = '{"Model":"VAR","Score":'+str(aic_score)+'}'
self.log.info('Status:- |... Score AIC: '+str(round(aic_score,2)))
tsModelTestObj=timeseriesModelTests(self.data1,self.targetFeature,self.dateTimeFeature,0)
lenTargetFeature=len(self.targetFeature)
r2=list(r2_dict.values())[-1]
rmse=list(rmse_dict.values())[-1]
mse=list(mse_dict.values())[-1]
mae=list(mae_dict.values())[-1]
## Added for additional scoring params
if (self.scoreParam.lower() == "r2"):
scoringparam_v=r2
self.log.info("VAR Univariant User selected scoring parameter is r2. r2 value: "+str(r2))
self.log.info('Status:- |... Score R2(Avg): '+str(round(r2,2)))
elif (self.scoreParam.lower() == "rmse"):
scoringparam_v=rmse
self.log.info("VAR Univariant User selected scoring parameter is RMSE. RMSE value: "+str(rmse))
self.log.info('Status:- |... Score RMSE(Avg): '+str(round(rmse,2)))
elif (self.scoreParam.lower() == "mse"):
scoringparam_v=mse
self.log.info("VAR Univariant User selected scoring parameter is MSE. MSE value: "+str(mse))
self.log.info('Status:- |... Score MSE(Avg): '+str(round(mse,2)))
elif (self.scoreParam.lower() == "mae"):
scoringparam_v=mae
self.log.info("VAR Univariant User selected scoring parameter is MAE. MAE value: "+str(mae))
self.log.info('Status:- |... Score MAE(Avg): '+str(round(mae,2)))
else:
scoringparam_v=rmse
self.log.info('Status:- |... Score RMSE(Avg): '+str(round(rmse,2)))
error_matrix = '"RMSE":"'+str(round(rmse,2))+'","MSE":"'+str(round(mse,2))+'"'
error_matrix=error_matrix+',"R2":"'+str(round(r2,2))+'","MAE":"'+str(round(mae,2))+'"'
self.log.info("VAR, all scoring param results: "+str(error_matrix))
scoredetails = '{"Model":"VAR","Score":'+str(round(scoringparam_v,2))+',"Scoring Param": "'+str(self.scoreParam)+'"}'
self.log.info('------------- Start VAR Model End-------------\n')
#countDependantFeature=0
out = durbin_watson(tdata)
for col, val in zip(tdata.columns, out):
self.log.info(col +':'+str(round(val, 2)))
if val>0.0 and val <1.5:
self.log.info("There is a positive correlation")
elif val>=1.5 and val <=2.5:
self.log.info("Relationship is Normal ")
else :
self.log.info("There is a negative correlation")
return('Success',modelName.upper(),self.scoreParam.lower(),round(scoringparam_v,2),self.model_fit,self.selectedColumns,error_matrix,scoredetails,df_pred,self.dictDiffCount,self.pred_freq,self.additional_regressors,lag_order)
except Exception as inst:
self.log.info('<!------------- Var model Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return 'Error',modelName.upper(),self.scoreParam.lower(),0,None,self.selectedColumns,'','{}',pd.DataFrame(),self.dictDiffCount,self.pred_freq,self.additional_regressors,0
## Get best time series algorithm among selected algs
def getbestmodel(self,rmse_prophet,rmse_arima,rmse_lstm,rmse_mlp,rmse_var):
best_model=''
## For R2 score
if (self.scoreParam.lower() == 'r2'):
modelScore = []
modelScore.append(rmse_prophet)
modelScore.append(rmse_arima)
modelScore.append(rmse_lstm)
modelScore.append(rmse_mlp)
modelScore.append(rmse_var)
if (max(modelScore) == rmse_arima):
best_model='arima'
self.log.info('Status:- |... TimeSeries Best Algorithm: ARIMA')
return best_model
elif (max(modelScore) == rmse_prophet):
best_model='fbprophet'
self.log.info('Status:- |... TimeSeries Best Algorithm: FBPROPHET')
return best_model
elif (max(modelScore) == rmse_lstm):
best_model='lstm'
self.log.info('Status:- |... TimeSeries Best Algorithm: LSTM')
return best_model
elif (max(modelScore) == rmse_mlp):
best_model='mlp'
self.log.info('Status:- |... TimeSeries Best Algorithm: MLP')
return best_model
elif (max(modelScore) == rmse_var):
best_model='var'
self.log.info('Status:- |... TimeSeries Best Algorithm: VAR')
return best_model
else:
#'Both arima and fbprophet rmse are equal, so both models are performing equal.
## So, selecting arima as best one.
best_model='arima'
return best_model
else:
modelScore = []
modelScore.append(rmse_prophet)
modelScore.append(rmse_arima)
modelScore.append(rmse_lstm)
modelScore.append(rmse_mlp)
modelScore.append(rmse_var)
if (min(modelScore) == rmse_arima and rmse_arima != 0xFFFF):
best_model='arima'
self.log.info('Status:- |... TimeSeries Best Algorithm: ARIMA')
return best_model
elif (min(modelScore) == rmse_prophet and rmse_prophet != 0xFFFF):
best_model='fbprophet'
self.log.info('Status:- |... TimeSeries Best Algorithm: FBPROPHET')
return best_model
elif (min(modelScore) == rmse_lstm and rmse_lstm != 0xFFFF):
best_model='lstm'
self.log.info('Status:- |... TimeSeries Best Algorithm: LSTM')
return best_model
elif (min(modelScore) == rmse_mlp and rmse_mlp != 0xFFFF):
best_model='mlp'
self.log.info('Status:- |... TimeSeries Best Algorithm: MLP')
return best_model
elif (min(modelScore) == rmse_var and rmse_var != 0xFFFF):
best_model='var'
self.log.info('Status:- |... TimeSeries Best Algorithm: VAR')
return best_model
else:
#'Both arima and fbprophet rmse are equal, so both models are performing equal.
## So, selecting arima as best one.
best_model='arima'
return best_model
## Selecting best model algorithm
def bestmodelProcess(self,modelNames,nfeatures,trained_data_file,tFeature,predicted_data_file,dataFolderLocation):
try:
best_model=''
lag_order = 1
predict_var=None
predict_arima=None
predict_lstm=None
predict_mlp=None
predict_fbprophet=None
modelNames = modelNames
modelNames=[x.lower() for x in modelNames]
inputFeature_len=nfeatures
status = 'Success'
if 'fbprophet' in modelNames:
status,modelName_prophet,fbprophet,rmse_prophet,fp_model_fit,selectedColumns_prophet,error_matrix_prophet,scoredetails_prophet,dictDiffCount_prophet,pred_freq_prophet,additional_regressors_prophet,predict_fbprophet = self.getfbprophetmodel(predicted_data_file,dataFolderLocation,tFeature)
if status.lower() == 'error':
self.log.info('-------------> FBPROPHET RMSE Score: Error')
if (self.scoreParam.lower() == 'r2'):
rmse_prophet = -0xFFFF
else:
rmse_prophet = 0xFFFF
else:
self.log.info("-------------> FBPROPHET RMSE Score:\t"+str(round(rmse_prophet,2)))
else:
if (self.scoreParam.lower() == 'r2'):
rmse_prophet = -0xFFFF
else:
rmse_prophet = 0xFFFF
if 'arima' in modelNames:
status,modelName,aic,rmse_arima,ar_model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,rmse_arima_act,predict_arima = self.getARIMAmodel(predicted_data_file)
if status.lower() == 'error':
self.log.info('-------------> ARIMA RMSE Score: Error')
if (self.scoreParam.lower() == 'r2'):
rmse_arima = -0xFFFF
else:
rmse_arima = 0xFFFF
else:
self.log.info('-------------> ARIMA RMSE Score:\t'+str(round(rmse_arima,2)))
else:
if (self.scoreParam.lower() == 'r2'):
rmse_arima = -0xFFFF ## -65535
else:
rmse_arima = 0xFFFF
if 'lstm' in modelNames:
if inputFeature_len == 1:
status,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,predict_lstm,lag_order,lstm_scaler = self.getUniVarientLSTMModel()
else:
status,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,predict_lstm,lag_order,lstm_scaler = self.getLSTMMultivariate()
if status.lower() == 'error':
self.log.info('-------------> LSTM RMSE Score: Error')
if (self.scoreParam.lower() == 'r2'):
rmse_lstm = -0xFFFF
else:
rmse_lstm = 0xFFFF
else:
self.log.info('-------------> LSTM RMSE Score:\t'+str(round(rmse_lstm,2)))
else:
if (self.scoreParam.lower() == 'r2'):
rmse_lstm = -0xFFFF
else:
rmse_lstm = 0xFFFF
if 'mlp' in modelNames:
status,modelName_mlp,score_type,rmse_mlp,mlp_model_fit,mlp_selectedColumns,error_matrix_mlp,scoredetails_mlp,predict_mlp,lag_order,mlp_scaler = self.getUniVarientMLPModel()
if status.lower() == 'error':
self.log.info('-------------> MLP Score: Error')
if (self.scoreParam.lower() == 'r2'):
rmse_mlp = -0xFFFF
else:
rmse_mlp = 0xFFFF
else:
self.log.info('-------------> MLP RMSE Score:\t'+str(round(rmse_mlp,2)))
else:
if (self.scoreParam.lower() == 'r2'):
rmse_mlp = -0xFFFF
else:
rmse_mlp = 0xFFFF
if 'var' in modelNames:
status,modelName_var,score_var_type,rmse_var,var_model,var_selectedColumns,error_matrix_var,scoredetails_var,predict_var,dictDiffCount,pred_freq,additional_regressors,lag_order = self.getVARmodel()
if status.lower() == 'error':
self.log.info('-------------> VAR Score: Error')
if (self.scoreParam.lower() == 'r2'):
rmse_var = -0xFFFF
else:
rmse_var = 0xFFFF
else:
if (self.scoreParam.lower() == 'r2'):
rmse_var = -0xFFFF
else:
rmse_var = 0xFFFF
best_model = self.getbestmodel(rmse_prophet,rmse_arima,rmse_lstm,rmse_mlp,rmse_var)
if (best_model.lower() == 'arima'):
self.log.info('Best model is ARIMA based on metric '+str(self.scoreParam.lower()))
predict_arima.to_csv(predicted_data_file)
filename,saved_model = self.save_model(ar_model_fit)
return best_model,modelName,aic,rmse_arima,ar_model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,filename,saved_model,lag_order,'NA'
elif (best_model.lower() == 'fbprophet'):
self.log.info('Best model is fbprophet based on metric '+str(self.scoreParam.lower()))
predict_fbprophet.to_csv(predicted_data_file)
filename,saved_model = self.save_model(fp_model_fit)
return best_model,modelName_prophet,fbprophet,rmse_prophet,fp_model_fit,selectedColumns_prophet,error_matrix_prophet,scoredetails_prophet,dictDiffCount_prophet,pred_freq_prophet,additional_regressors_prophet,filename,saved_model,lag_order,'NA'
elif (best_model.lower() == 'var'):
self.log.info('Best model is VAR based on metric '+str(self.scoreParam.lower()))
self.data.to_csv(trained_data_file)
predict_var.to_csv(predicted_data_file)
filename,saved_model = self.save_model(var_model)
return best_model,modelName_var,score_var_type,rmse_var,var_model,var_selectedColumns,error_matrix_var,scoredetails_var,dictDiffCount,pred_freq,additional_regressors,filename,saved_model,lag_order,'NA'
elif (best_model.lower() == 'lstm'):
self.log.info('Best model is LSTM based on metric '+str(self.scoreParam.lower()))
predict_lstm.to_csv(predicted_data_file)
filename,saved_model,scaler_model = self.save_dl_model(lstm_model_fit,lstm_scaler)
return best_model,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,self.dictDiffCount,self.pred_freq,self.additional_regressors,filename,saved_model,lag_order,scaler_model
elif (best_model.lower() == 'mlp'):
self.log.info('Best model is MLP based on metric '+str(self.scoreParam.lower()))
predict_mlp.to_csv(predicted_data_file)
filename,saved_model,scaler_model = self.save_dl_model(mlp_model_fit,mlp_scaler)
return best_model,modelName_mlp,score_type,rmse_mlp,mlp_model_fit,mlp_selectedColumns,error_matrix_mlp,scoredetails_mlp,self.dictDiffCount,self.pred_freq,self.additional_regressors,filename,saved_model,lag_order,scaler_model
else:
pass
except Exception as e:
self.log.info('Issue in running multi time series algorithm selection process..Please check the config params')
self.log.info('error: '+str(e))
#Method to determine seasonality and stationrity in the input data features. (Task:12622,12623)
def seasonality_stationarity_test(self):
##The below part is to test stationarity and sessonality in the given time series data based on statsmodels lib.
#self.data,self.targetFeature,self.dateTimeFeature
self.log.info("<-------------- Time series stationarity and seasonality test Started...---------------->\n")
ts_sstest=tsStationarySeasonalityTest(self.data,self.deployLocation)
## Time series Stationary check
## Currently stationarity check method set as Augmented dickey fuller, but kpss method also implemented.
stationary_method='adfuller'
if (isinstance(self.targetFeature,list)):
target=self.targetFeature
pass
elif (isinstance(self.targetFeature,str)):
target=list(self.targetFeature.split(','))
stats_model,n_lags,p_value,stationary_result,stationary_combined_res=ts_sstest.stationary_check(target,self.dateTimeFeature,stationary_method)
## Time series Seasonality check
##Seasonal model default set as additive
seasonal_model="additive"
df,decompose_result_mult,seasonality_result,seasonality_combined_res=ts_sstest.seasonal_check(target,self.dateTimeFeature,seasonal_model)
self.log.info("<-------------- Time series stationarity and seasonality test completed.---------------->\n")
return stationary_result,seasonality_result
#Main timeseries function.
def timeseries_learning(self,trained_data_file,predicted_data_file,dataFolderLocation):
dataFolderLocation=dataFolderLocation
lag_order = 1
# ##The below part is to test stationarity and sessonality in the given time series data based on statsmodels lib.
stationary_result,seasonality_result=self.seasonality_stationarity_test()
try :
tFeature = self.targetFeature.split(',')
lentFeature=len(tFeature)
try:
if lentFeature > 1:
if any('timeseriesforecasting' in x.lower() for x in self.modelName): #task 11997
self.modelName.remove('timeseriesforecasting')
if 'arima' in self.modelName:
self.log.info('Status:- |... TimeSeries algorithm ARIMA not supported for multiple features')
self.modelName.remove('arima')
if 'fbprophet' in self.modelName:
self.log.info('Status:- |... TimeSeries algorithm FBPROPHET not supported for multiple features')
self.modelName.remove('fbprophet')
if 'mlp' in self.modelName:
self.log.info('Status:- |... TimeSeries algorithm MLP not supported for multiple features')
self.modelName.remove('mlp')
if len(self.modelName) == 0:
self.log.info('--------> Default Set to VAR')
self.modelName.append('var')
if lentFeature == 1:
if any('timeseriesforecasting' in x.lower() for x in self.modelName): #task 11997
self.modelName.remove('timeseriesforecasting')
if 'var' in self.modelName:
self.log.info('Status:- |... TimeSeries algorithm VAR not supported for single feature')
self.modelName.remove('var')
if len(self.modelName) == 0:
self.log.info('--------> Default Set to ARIMA,FBProphet')
self.modelName.append('arima')
except Exception as e:
self.log.info('input model name error: '+ str(e))
self.log.info("error in user selected model, may be wrong configuration, please check.")
if (len(self.modelName) > 1):
try:
self.log.info('User selected models: '+str(self.modelName))
best_model,modelName,score_type,score,model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,filename,saved_model,lag_order,scaler_transformation = self.bestmodelProcess(self.modelName,lentFeature,trained_data_file,tFeature,predicted_data_file,dataFolderLocation)
return best_model,modelName,score_type,score,model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,filename,saved_model,lag_order,scaler_transformation
except Exception as e:
self.log.info('multi model timeseries processing error '+str(e))
else:
self.modelName = self.modelName[0]
## Normal arima ,var or fbprophet model call (user selects only one model at a time)
if self.modelName.lower() == 'fbprophet':
try:
model_name='fbprophet'
status,modelName,fbprophet,rmse_prophet,fp_model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,predict_output = self.getfbprophetmodel(predicted_data_file,dataFolderLocation,tFeature)
if status.lower() == 'success':
predict_output.to_csv(predicted_data_file)
filename,saved_model = self.save_model(fp_model_fit)
return 'self.modelName',modelName,fbprophet,rmse_prophet,fp_model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,filename,saved_model,lag_order,'NA'
else:
raise Exception('Exception during model training')
except Exception as e:
self.log.info('fbprophet error....')
self.log.info(e)
elif self.modelName.lower() == 'encoder_decoder_lstm_mvi_uvo':
try:
status,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,predict_lstm,lag_order,lstm_scaler = self.getEncDecLSTMMultVrtInUniVrtOut()
if status.lower() == 'success':
predict_lstm.to_csv(predicted_data_file)
filename,saved_model,scaler_model = self.save_dl_model(lstm_model_fit,lstm_scaler)
return self.modelName,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,self.dictDiffCount,self.pred_freq,self.additional_regressors,filename,saved_model,lag_order,scaler_model
else:
raise Exception('Exception during model training')
except Exception as inst:
self.log.info('<!------------- LSTM Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
elif self.modelName.lower() == 'lstm':
try:
if lentFeature == 1:
status,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,predict_lstm,lag_order,lstm_scaler = self.getUniVarientLSTMModel()
else:
status,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,predict_lstm,lag_order,lstm_scaler = self.getLSTMMultivariate()
if status.lower() == 'success':
predict_lstm.to_csv(predicted_data_file)
filename,saved_model,scaler_model = self.save_dl_model(lstm_model_fit,lstm_scaler)
return self.modelName,modelName_lstm,score_type,rmse_lstm,lstm_model_fit,lstm_selectedColumns,error_matrix_lstm,scoredetails_lstm,self.dictDiffCount,self.pred_freq,self.additional_regressors,filename,saved_model,lag_order,scaler_model
else:
raise Exception('Exception during model training')
except Exception as inst:
self.log.info('<!------------- LSTM Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
elif self.modelName.lower() == 'mlp':
try:
status,modelName_mlp,score_type,rmse_mlp,mlp_model_fit,mlp_selectedColumns,error_matrix_mlp,scoredetails_mlp,predict_mlp,lag_order,mlp_scaler = self.getUniVarientMLPModel()
if status.lower() == 'success':
predict_mlp.to_csv(predicted_data_file)
filename,saved_model,scaler_model = self.save_dl_model(mlp_model_fit,mlp_scaler)
return self.modelName,modelName_mlp,score_type,rmse_mlp,mlp_model_fit,mlp_selectedColumns,error_matrix_mlp,scoredetails_mlp,self.dictDiffCount,self.pred_freq,self.additional_regressors,filename,saved_model,lag_order,scaler_model
else:
raise Exception('Exception during model training')
except Exception as inst:
self.log.info('<!------------- MLP Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
else:
#task 12627 time series profiler removed
if lentFeature>1:
self.modelName='var'
self.data.to_csv(trained_data_file)
else:
self.modelName='arima'
if self.modelName.lower()=='var':
tsModelTestObj=timeseriesModelTests(self.data,self.targetFeature,self.dateTimeFeature,0)
self.data,self.dictDiffCount=tsModelTestObj.StatinaryChecks(self.dictDiffCount)
#self.log.info('Status:- |... Stationary Check Done.')
gtestResults,countVariables=tsModelTestObj.grangersCausationMatrix(self.data,tFeature)
if countVariables >= (lentFeature*lentFeature)-(lentFeature) or ((lentFeature*lentFeature)-(lentFeature))/2 :
coIntegrationVectors=tsModelTestObj.coIntegrationTest(self.data)
if coIntegrationVectors<=lentFeature:
self.log.info("There are statistically significant relationship in data ")
self.log.info('Status:- |... Statistically Check Done. Statistically significant relations')
else:
self.log.info("There are no statistically significant relationship in data")
self.log.info('Status:- |... Statistically Check Done. No statistically significant relations')
else:
self.modelName='arima'
if self.modelName.lower()=='var':
try:
self.log.info('ARIMA, FBProphet cannot apply, Input data contains more than one feature, only VAR algorithm can apply, applying VAR by AION \n')
status,modelName,aic,aic_score,model_fit,selectedColumns,error_matrix,scoredetails,predict_var,dictDiffCount,pred_freq,additional_regressors,lag_order = self.getVARmodel()
if status.lower() == 'success':
filename,saved_model = self.save_model(model_fit)
predict_var.to_csv(predicted_data_file)
return self.modelName,modelName,aic,aic_score,model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,filename,saved_model,lag_order,'NA'
else:
raise Exception('Exception during VAR model training')
except Exception as inst:
self.log.info('<!------------- Var model Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
if self.modelName.lower() == 'arima':
try:
status,modelName,aic,scoringparam_v,model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,rmse_arima_act,predict_output = self.getARIMAmodel(predicted_data_file)
if status.lower() == 'success':
predict_output.to_csv(predicted_data_file)
filename,saved_model = self.save_model(model_fit)
lag_order=0
return self.modelName,modelName,aic,scoringparam_v,model_fit,selectedColumns,error_matrix,scoredetails,dictDiffCount,pred_freq,additional_regressors,filename,saved_model,lag_order,'NA'
else:
raise Exception('Exception during model training')
except Exception as inst:
self.log.info('<!------------- Arima Error ---------------> '+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
except Exception as inst:
self.log.info('<!------------- TimeSeries Learning Error ---------------> '+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def invertTransformation(self,Xtrain,targetFeature, preddf,dictDiffCount):
try:
dfforecast = preddf.copy()
self.log.info(dfforecast.head(5))
columns =targetFeature.split(",")
self.log.info(columns)
self.log.info(dictDiffCount)
for col in columns:
if col in dictDiffCount:
if dictDiffCount[col]==2:
dfforecast[col] = (Xtrain[col].iloc[-1]-Xtrain[col].iloc[-2]) + dfforecast[col].cumsum()
dfforecast[col] = Xtrain[col].iloc[-1] + dfforecast[col].cumsum()
# Roll back 1st Diff
return dfforecast
except Exception as inst:
self.log.info('<!------------- invertTransformation Error ---------------> '+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno)) |
ts_arima_eion.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
# For timeseries pyramid pdaarima module
from pmdarima.arima import auto_arima
import pmdarima as pm
import json
#Python sklearn & std libraries
import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.feature_selection import VarianceThreshold
from sklearn.metrics import mean_absolute_error
from sklearn.metrics import mean_squared_error
#from sklearn.metrics import mean_absolute_percentage_error
from sklearn.linear_model import LinearRegression
from math import sqrt
import warnings
# For serialization.
#from sklearn.externals import joblib
import pickle
import os,sys
# For ploting (mathlab)
import matplotlib.pyplot as plt
#Import eion config manager module
import logging
from sklearn import metrics
from sklearn.metrics import accuracy_score
import time
import os
import sys
# Eion arima module
class eion_arima ():
#Constructor
def __init__(self,configfile,testpercentage,sesonalityChecks,stationaryChecks): # eaobj - eion arima class object
try:
tsarima_params = configfile
self.testpercentage = testpercentage
self.start_p= int(tsarima_params['start_p'])
self.start_q= int(tsarima_params['start_q'])
self.max_p= int(tsarima_params['max_p'])
self.max_q= int(tsarima_params['max_q'])
self.max_d= int(tsarima_params['max_d'])
self.max_order= int(tsarima_params['max_order'])
self.start_Q= int(tsarima_params['start_Q'])
self.max_P= int(tsarima_params['max_P'])
self.max_D= int(tsarima_params['max_D'])
self.max_Q= int(tsarima_params['max_Q'])
self.m= int(tsarima_params['m'])
self.start_P= int(tsarima_params['start_P'])
self.seasonal= tsarima_params['seasonal']
#self.seasonal= sesonalityChecks
self.stationary=stationaryChecks
#print("self.seasonal: \n",self.seasonal)
#print("self.stationary: \n",self.stationary)
if self.seasonal and not self.seasonal.isspace():
if (self.seasonal.lower() == 'true'):
self.seasonal=True
elif (self.seasonal.lower() == 'false'):
self.seasonal=False
else:
self.seasonal=True
else:
self.seasonal=True
self.d= int(tsarima_params['d'])
self.D= int(tsarima_params['D'])
#self.trace= tsarima_params['trace']
self.error_action= tsarima_params['error_action']
self.suppress_warnings= tsarima_params['suppress_warnings']
self.stepwise= tsarima_params['stepwise']
#self.random= tsarima_params['random']
self.log = logging.getLogger('eion')
except Exception as inst:
self.log.info('<!------------- Arima INIT Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def mean_absolute_percentage_error(self,y_true, y_pred):
try:
y_true, y_pred=np.array(y_true), np.array(y_pred)
return np.mean(np.abs((y_true - y_pred) / y_true+sys.float_info.epsilon)) * 100
except Exception as inst:
self.log.info('<------------- mean_absolute_percentage_error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def eion_arima(self,train_data):
try:
start = time.time()
auto_arima_stepwise_fit = pm.auto_arima(train_data, start_p=self.start_p, start_q=self.start_q,max_p=self.max_p, max_q=self.max_q,max_d=self.max_d,max_P=self.max_P,max_D=self.max_D,max_Q=self.max_Q,max_order=self.max_order, m=self.m,start_P=self.start_P,start_Q=self.start_Q, seasonal=self.seasonal,stationary=self.stationary,d=self.d, D=self.D,error_action=self.error_action,suppress_warnings=self.suppress_warnings,stepwise=self.stepwise)
#auto_arima_stepwise_fit = pm.auto_arima(train_data, start_p=self.start_p, start_q=self.start_q,max_p=self.max_p, max_q=self.max_q,max_d=self.max_d,max_P=self.max_P,max_D=self.max_D,max_Q=self.max_Q,max_order=self.max_order, m=self.m,start_P=self.start_P,start_Q=self.start_Q, seasonal=True,stationary=True,d=self.d, D=self.D,error_action=self.error_action,suppress_warnings=self.suppress_warnings,random_state=20,stepwise=True)
aic_score = auto_arima_stepwise_fit.aic()
self.log.info('------->AIC Score: '+str(aic_score))
self.log.info('\n--------- Fit Summary --------------')
self.log.info (auto_arima_stepwise_fit.summary())
self.log.info('--------- Fit Summary End--------------\n')
self.log.info("\n--------------- Modal Validation Start ---------------")
size = int(len(train_data) * (100 - self.testpercentage)/100)
train = train_data.loc[0:size]
valid = train_data.loc[size:len(train_data)]
# valid_perc=((100-self.testpercentage)/100)
# valid_perc=round(valid_perc, 1)
# print("valid_perc: \n", valid_perc)
self.log.info("------->Train Data Shape: "+str(train.shape))
self.log.info("------->Valid Data Shape"+str(valid.shape))
start1=len(train)
end1=len(train_data)
modelfit = auto_arima_stepwise_fit.fit(train)
a_prediction = auto_arima_stepwise_fit.predict(valid.shape[0])
#a_prediction = auto_arima_stepwise_fit.predict(n_periods=len(valid))
#a_prediction = auto_arima_stepwise_fit.predict(start=start1,end=end1)
#print("a_prediction: \n",a_prediction)
#self.log.info(a_prediction)
mae = metrics.mean_absolute_error(valid, a_prediction)
self.log.info ("------->MAE: "+str(mae))
mape = self.mean_absolute_percentage_error(valid, a_prediction)
#mape=np.mean(np.abs((valid - a_prediction) / valid)) * 100
self.log.info ("------->MAPE :"+str(mape))
#RMSE
rmse = sqrt(mean_squared_error(valid,a_prediction))
mse = mean_squared_error(valid,a_prediction)
self.log.info ("------->RMSE :"+str(rmse))
self.log.info ("------->MSE :"+str(mse))
from sklearn.metrics import r2_score
r2 = r2_score(valid,a_prediction)
########### End ####################
# now we have the model
auto_arima_stepwise_fit.fit(train_data)
self.log.info("------------- Validate Model End----------------\n")
executionTime=time.time() - start
self.log.info('-------> Time: '+str(executionTime)+'\n')
return auto_arima_stepwise_fit,mae,rmse,mse,r2,aic_score,mape,valid,a_prediction
except Exception as inst:
self.log.info('<!------------- Arima Execute Error ---------------> '+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
|
timeseriesDLMultivariate.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import pandas as pd
import os
import numpy as np
import numpy
import pandas
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from tensorflow.keras.layers import LSTM
from sklearn.preprocessing import MinMaxScaler
import logging
import tensorflow as tf
from tensorflow.keras.layers import Dropout
import math
import tensorflow as tf
import keras_tuner
#from keras_tuner.engine.hyperparameters import HyperParameters
from keras_tuner.tuners import RandomSearch,BayesianOptimization ,Hyperband
from sklearn.model_selection import train_test_split
from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator
import warnings
warnings.simplefilter("ignore", UserWarning)
class timeseriesDLMultivariate:
def __init__(self,configfile,testpercentage,targetFeature,dateTimeFeature):
self.look_back=None
# self.df=df
self.epochs=None
self.batch_size=None
self.hidden_layers=None
self.optimizer=None
self.activation_fn="relu"
self.loss_fn=None
self.first_layer=None
self.dropout=None
self.model_name=None
self.dl_params = configfile
# self.data=data
self.targetFeature=targetFeature
self.dateTimeFeature=dateTimeFeature
self.testpercentage = float(testpercentage)
self.log = logging.getLogger('eion')
##Added for ts hpt (TFSTask:7033)
self.tuner_algorithm=""
self.num_features=0
##Get deep learning model hyperparameter from advanced config
def getdlparams(self):
val=self.dl_params
self.log.info('-------> The given mlp/lstm timeseries algorithm parameters:>>')
self.log.info(" "+str(val))
for k,v in val.items():
try:
if (k == "tuner_algorithm"):
self.tuner_algorithm=str(v)
elif (k == "activation"):
self.activation_fn=str(v)
elif (k == "optimizer"):
self.optimizer=str(v)
elif (k == "loss"):
self.loss_fn=str(v)
elif (k == "first_layer"):
if not isinstance(k,list):
self.first_layer=str(v).split(',')
else:
self.first_layer=k
elif (k == "lag_order"):
if isinstance(k,list):
k = ''.join(v)
k=int(float(str(v)))
else:
self.look_back=int(float(str(v)))
elif (k == "hidden_layers"):
self.hidden_layers=int(v)
elif (k == "dropout"):
if not isinstance(k,list):
self.dropout=str(v).split(',')
else:
self.dropout=k
elif (k == "batch_size"):
self.batch_size=int(v)
elif (k == "epochs"):
self.epochs=int(v)
elif (k == "model_name"):
self.model_name=str(v)
except Exception as e:
self.log.info('Exception occured in deeep learn param reading, setting up default params.')
self.activation_fn="relu"
self.optimizer="adam"
self.loss_fn="mean_squared_error"
self.first_layer=[8,512]
self.hidden_layers=1
self.look_back=int(2)
self.dropout=[0.1,0.5]
self.batch_size=2
self.epochs=50
self.model_name="lstmmodel.h5"
continue
# Reshape the data to the required input shape of the LSTM model
def create_dataset(self,X, y, n_steps):
Xs, ys = [], []
for i in range(len(X) - n_steps):
v = X.iloc[i:(i + n_steps)].values
Xs.append(v)
ys.append(y.iloc[i + n_steps])
return np.array(Xs), np.array(ys)
## Added function for hyperparam tuning (TFSTask:7033)
def build_model(self,hp):
n_features = len(self.targetFeature)
try:
loss=self.loss_fn
optimizer=self.optimizer
# self.getdlparams()
try:
if optimizer.lower() == "adam":
optimizer=tensorflow.keras.optimizers.Adam
elif(optimizer.lower() == "adadelta"):
optimizer=tensorflow.keras.optimizers.experimental.Adadelta
elif(optimizer.lower() == "nadam"):
optimizer=tensorflow.keras.optimizers.experimental.Nadam
elif(optimizer.lower() == "adagrad"):
optimizer=tensorflow.keras.optimizers.experimental.Adagrad
elif(optimizer.lower() == "adamax"):
optimizer=tensorflow.keras.optimizers.experimental.Adamax
elif(optimizer.lower() == "rmsprop"):
optimizer=tensorflow.keras.optimizers.experimental.RMSprop
elif(optimizer.lower() == "sgd"):
optimizer=tensorflow.keras.optimizers.experimental.SGD
else:
optimizer=tensorflow.keras.optimizers.Adam
except:
optimizer=tf.keras.optimizers.Adam
pass
# look_back_min=int(self.look_back[0])
# look_back_max=int(self.look_back[1])
first_layer_min=round(int(self.first_layer[0]))
first_layer_max=round(int(self.first_layer[1]))
dropout_min=float(self.dropout[0])
dropout_max=float(self.dropout[1])
model=tf.keras.Sequential()
try:
model.add(LSTM(units=hp.Int('units',min_value=first_layer_min,max_value=first_layer_max,step=16),input_shape=(self.look_back,self.num_features)))
except Exception as e:
import traceback
self.log.info("lstm build traceback: \n"+str(traceback.print_exc()))
return model
model.add(Dropout(hp.Float('Dropout_rate',min_value=dropout_min,max_value=dropout_max,step=0.1)))
model.add(Dense(units=n_features))
model.compile(optimizer=optimizer(hp.Choice('learning_rate',values=[1e-1,1e-2,1e-3,1e-4])),loss=loss,metrics=[self.loss_fn])
except Exception as e:
self.log.info(",Hyperparam tuning build_model err msg: \n"+ str(e))
return model
##Multivariate lstm prediction function (lstm model, train, prediction, metrics)
def lstm_multivariate(self,df):
try:
self.getdlparams()
n_features = len(self.targetFeature)
self.num_features=n_features
try:
if (type(self.targetFeature) is list):
pass
else:
self.targetFeature = list(self.targetFeature.split(","))
except:
pass
df_new = df[df.columns[df.columns.isin(self.targetFeature)]]
scaler=MinMaxScaler()
df_transformed=scaler.fit_transform(df_new)
## For hyperparam tuning below part is added.only for getting best model and best hyperparameters
train_size = int(len(df) * 0.80)
train_data, test_data = train_test_split(df, test_size=0.2, shuffle=False)
self.hpt_train=train_data
time_steps=self.look_back ## Just for initialization before hyperparameter tuning.
tuner_alg=self.tuner_algorithm
#The below create_dataset only for getting best model and best hyperparameters
X_train, y_train = self.create_dataset(train_data, train_data, time_steps)
X_test, y_test = self.create_dataset(test_data, test_data, time_steps)
self.log.info("Hyperparameter tuning algorithm is given by user (AION->Advanced configuration -> timeSeriesForecasting->LSTM): \n"+str(tuner_alg))
try:
## Remove untitled_project dir in AION root folder created by previous tuner search run
import shutil
shutil.rmtree(r".\untitled_project")
except:
pass
try:
if (tuner_alg.lower()=="randomsearch"):
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
elif (tuner_alg.lower()=="bayesianoptimization"):
tuner=BayesianOptimization(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
elif (tuner_alg.lower()=="hyperband"):
tuner=Hyperband(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_epochs=50,factor=3)
else:
self.log.info("The given alg is not implemented. Using default hyperparam tuning algorithm: RandomSearch.\n")
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
from keras.callbacks import EarlyStopping
stop_early = EarlyStopping(monitor='val_loss', patience=5)
except Exception as e:
self.log.info("The given alg have some issue, Using default hyperparam tuning algorithm: RandomSearch.\n")
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
self.log.info("tuner errmsg:\n"+str(e))
#hpt search for best params
try:
tuner.search(X_train, y_train,validation_data=(X_test, y_test),callbacks=[stop_early])
except:
tuner.search(x=X_train,y=y_train,validation_split=0.2,callbacks=[stop_early])
# best_model = tuner.get_best_models(num_models=1)[0]
# self.log.info("best_model.summary(): \n"+str(best_model.summary()))
best_hps=tuner.get_best_hyperparameters(num_trials=1)[0]
self.log.info("TS Multivariate LSTM best hyperparameter values:\n"+str(best_hps.values))
self.log.info("Activation fn:\n"+str(self.activation_fn))
# time_steps_best=best_hps.get('time_steps')
n_input=self.look_back
best_hmodel=tuner.hypermodel.build(best_hps)
optimizer=self.optimizer
self.first_layer=best_hps.get('units')
self.dropout=best_hps.get('Dropout_rate')
learning_rate=float(best_hps.get('learning_rate'))
try:
##TFSTask:7033, Added below try block for time series hyperparam tuning, here, for any optimizer, best learning_rate is provided from best_hps.
try:
if optimizer.lower() == "adam":
optimizer=tensorflow.keras.optimizers.Adam(learning_rate=learning_rate)
elif(optimizer.lower() == "adadelta"):
optimizer=tensorflow.keras.optimizers.experimental.Adadelta(learning_rate=learning_rate)
elif(optimizer.lower() == "nadam"):
optimizer=tensorflow.keras.optimizers.experimental.Nadam(learning_rate=learning_rate)
elif(optimizer.lower() == "adagrad"):
optimizer=tensorflow.keras.optimizers.experimental.Adagrad(learning_rate=learning_rate)
elif(optimizer.lower() == "adamax"):
optimizer=tensorflow.keras.optimizers.experimental.Adamax(learning_rate=learning_rate)
elif(optimizer.lower() == "rmsprop"):
optimizer=tensorflow.keras.optimizers.experimental.RMSprop(learning_rate=learning_rate)
elif(optimizer.lower() == "sgd"):
optimizer=tensorflow.keras.optimizers.experimental.SGD(learning_rate=learning_rate)
else:
optimizer=tensorflow.keras.optimizers.Adam(learning_rate=learning_rate)
except:
optimizer=tf.keras.optimizers.Adam(learning_rate=learning_rate)
pass
##From best hyperparameter values, now creating multivariate time series model using time generator.
t_lb=1
test_size=t_lb+1
train,test = train_test_split(df_transformed,test_size=0.2,shuffle=False)
generatorTrain=TimeseriesGenerator(df_transformed,df_transformed,length=n_input,batch_size=self.batch_size)
# generatorTest=TimeseriesGenerator(test,test,length=n_input,batch_size=self.batch_size)
batch_0=generatorTrain[0]
x,y=batch_0
epochs=int(self.epochs)
##Multivariate LSTM model
try:
from tensorflow.keras.layers import Dropout
model=Sequential()
model.add(LSTM(self.first_layer,activation=self.activation_fn,input_shape=(n_input,n_features)))
model.add(Dropout(self.dropout))
model.add(Dense(n_features))
model.compile(optimizer=self.optimizer,loss=self.loss_fn)
#model.fit(generatorTrain,epochs=epochs,batch_size=self.batch_size,shuffle=False)
model.fit_generator(generatorTrain, epochs=epochs,shuffle=False, verbose=0)
# lstm_mv_testScore_mse = model.evaluate(x, y, verbose=0)
except Exception as e:
self.log.info("multivariate model build error: error msg:: \n"+str(e))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
#predictions = model.predict_generator(generatorTest)
except Exception as e:
self.log.info("multivariate model build error: error msg:: \n"+str(e))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
try:
predictions=[]
future_pred_len=n_input
#To get values for prediction,taking look_back steps of rows
first_batch=test[-future_pred_len:]
c_batch = first_batch.reshape((1,future_pred_len,n_features))
current_pred=None
for i in range(len(test)):
#get pred for firstbatch
current_pred=model.predict_generator(c_batch)[0]
predictions.append(current_pred)
#remove first val
c_batch_rmv_first=c_batch[:,1:,:]
#update
c_batch=np.append(c_batch_rmv_first,[[current_pred]],axis=1)
prediction_actual=scaler.inverse_transform(predictions)
test_data_actual=scaler.inverse_transform(test)
mse=None
rmse=None
## Creating dataframe for actual,predictions
try:
pred_cols=list()
for i in range(len(self.targetFeature)):
pred_cols.append(self.targetFeature[i]+'_pred')
predictions = pd.DataFrame(prediction_actual, columns=pred_cols)
actual = pd.DataFrame(test_data_actual, columns=self.targetFeature)
actual.columns = [str(col) + '_actual' for col in df.columns]
df_predicted=pd.concat([actual,predictions],axis=1)
self.log.info("LSTM Multivariate prediction dataframe: \n"+str(df_predicted))
from math import sqrt
from sklearn.metrics import mean_squared_error
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error
target=self.targetFeature
mse_dict={}
rmse_dict={}
mae_dict={}
r2_dict={}
lstm_var = 0
for name in target:
index = df.columns.get_loc(name)
mse = mean_squared_error(test_data_actual[:,index],prediction_actual[:,index])
mse_dict[name]=mse
rmse=sqrt(mse)
rmse_dict[name]=rmse
lstm_var = lstm_var+rmse
self.log.info("Name of the target feature: "+str(name))
self.log.info("RMSE of the target feature: "+str(rmse))
r2 = r2_score(test_data_actual[:,index],prediction_actual[:,index])
r2_dict[name]=r2
mae = mean_absolute_error(test_data_actual[:,index],prediction_actual[:,index])
mae_dict[name]=mae
## For VAR comparison, send last target mse and rmse from above dict
lstm_var = lstm_var/len(target)
select_msekey=list(mse_dict.keys())[-1]
l_mse=list(mse_dict.values())[-1]
select_rmsekey=list(rmse_dict.keys())[-1]
l_rmse=list(rmse_dict.values())[-1]
select_r2key=list(r2_dict.keys())[-1]
l_r2=list(r2_dict.values())[-1]
select_maekey=list(mae_dict.keys())[-1]
l_mae=list(mae_dict.values())[-1]
self.log.info("Selected target feature of LSTM for best model selection: "+str(select_rmsekey))
self.log.info("lstm rmse: "+str(l_rmse))
self.log.info("lstm mse: "+str(l_mse))
self.log.info("lstm r2: "+str(l_r2))
self.log.info("lstm mae: "+str(l_mae))
except Exception as e:
import traceback
print(" traceback error:\n",traceback.print_exc())
self.log.info("prediction error traceback: \n"+str(traceback.print_exc()))
except Exception as e:
self.log.info("dataframe creation error. err.msg: "+str(e))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
return 'Success',round(l_mse,2),round(l_rmse,2),round(l_r2,2),round(l_mae,2),model,df_predicted,n_input,scaler
# import os
#predicted_file_name='lstm_prediction_df.csv'
#predicted_file_path=os.path.join(self.dataFolderLocation,predicted_file_name)
#df_predicted.to_csv(predicted_file_path)
##save model
#model_path = os.path.join(self.dataFolderLocation,self.model_name)
#self.log.info("mlp model saved at: "+str(model_path))
#model.save(model_path)
except Exception as e:
## Just use below traceback print to get detailed error information.
# import traceback
# print(" traceback error 7:\n",traceback.print_exc())
## Enable traceback for debugging
self.log.info("dataframe creation error. err.msg: "+str(e))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
|
tsDLMultiVrtInUniVrtOut.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import os
import sys
import pandas as pd
import numpy as np
import numpy
import pandas
import math
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Input, Dense, TimeDistributed, LSTM, Dropout, RepeatVector
from sklearn.preprocessing import MinMaxScaler
import logging
import tensorflow as tf
import keras_tuner
#from keras_tuner.engine.hyperparameters import HyperParameters
from keras_tuner.tuners import RandomSearch,BayesianOptimization ,Hyperband
from sklearn.model_selection import train_test_split
from tensorflow.keras.preprocessing.sequence import TimeseriesGenerator
import warnings
warnings.simplefilter("ignore", UserWarning)
from sklearn.metrics import mean_absolute_percentage_error
class tsDLMultiVrtInUniVrtOut:
def __init__(self,configfile,testpercentage,targetFeature,dateTimeFeature):
self.look_back=None
self.look_forward=None
# self.df=df
self.epochs=None
self.batch_size=None
self.hidden_layers=None
self.optimizer=None
self.activation_fn="relu"
self.loss_fn=None
self.first_layer=None
self.dropout=None
self.model_name=None
self.dl_params = configfile
# self.data=data
self.targetFeature=targetFeature
self.dateTimeFeature=dateTimeFeature
self.testpercentage = float(testpercentage)
self.log = logging.getLogger('eion')
##Added for ts hpt (TFSTask:7033)
self.tuner_algorithm=""
self.num_features=0
##Get deep learning model hyperparameter from advanced config
def getdlparams(self):
val=self.dl_params
self.log.info('-------> The given mlp/lstm timeseries algorithm parameters:>>')
self.log.info(" "+str(val))
for k,v in val.items():
try:
if (k == "tuner_algorithm"):
self.tuner_algorithm=str(v)
elif (k == "activation"):
if not isinstance(k,list):
self.activation_fn=str(v).split(',')
else:
self.activation_fn=k
elif (k == "optimizer"):
self.optimizer=str(v)
elif (k == "loss"):
self.loss_fn=str(v)
elif (k == "first_layer"):
if not isinstance(k,list):
self.first_layer=str(v).split(',')
else:
self.first_layer=k
elif (k == "lag_order"):
if isinstance(k,list):
k = ''.join(v)
k=int(float(str(v)))
else:
self.look_back=int(float(str(v)))
elif (k == "forward_order"):
if isinstance(k,list):
k = ''.join(v)
k=int(float(str(v)))
else:
self.look_forward=int(float(str(v)))
elif (k == "hidden_layers"):
self.hidden_layers=int(v)
elif (k == "dropout"):
if not isinstance(k,list):
self.dropout=str(v).split(',')
else:
self.dropout=k
elif (k == "batch_size"):
self.batch_size=int(v)
elif (k == "epochs"):
self.epochs=int(v)
elif (k == "model_name"):
self.model_name=str(v)
except Exception as e:
self.log.info('Exception occured in deeep learn param reading, setting up default params.')
self.activation_fn="relu"
self.optimizer="adam"
self.loss_fn="mean_squared_error"
self.first_layer=[8,512]
self.hidden_layers=1
self.look_back=int(2)
self.dropout=[0.0,0.1,0.01]
self.batch_size=2
self.epochs=50
self.model_name="lstmmodel.h5"
continue
# Reshape the data to the required input shape of the LSTM model
def create_dataset(self,series, n_past, n_future, targetcolindx):
X, y = list(), list()
for window_start in range(len(series)):
past_end = window_start + n_past
future_end = past_end + n_future
if future_end > len(series):
break
# slicing the past and future parts of the window
past, future = series[window_start:past_end, :], series[past_end:future_end, targetcolindx]
X.append(past)
y.append(future)
return np.array(X), np.array(y)
#return X, y
## Added function for hyperparam tuning (TFSTask:7033)
def build_model(self,hp):
n_features = self.num_features
try:
loss=self.loss_fn
optimizer=self.optimizer
# self.getdlparams()
try:
if optimizer.lower() == "adam":
optimizer=tensorflow.keras.optimizers.Adam
elif(optimizer.lower() == "adadelta"):
optimizer=tensorflow.keras.optimizers.experimental.Adadelta
elif(optimizer.lower() == "nadam"):
optimizer=tensorflow.keras.optimizers.experimental.Nadam
elif(optimizer.lower() == "adagrad"):
optimizer=tensorflow.keras.optimizers.experimental.Adagrad
elif(optimizer.lower() == "adamax"):
optimizer=tensorflow.keras.optimizers.experimental.Adamax
elif(optimizer.lower() == "rmsprop"):
optimizer=tensorflow.keras.optimizers.experimental.RMSprop
elif(optimizer.lower() == "sgd"):
optimizer=tensorflow.keras.optimizers.experimental.SGD
else:
optimizer=tensorflow.keras.optimizers.Adam
except:
optimizer=tf.keras.optimizers.Adam
pass
# look_back_min=int(self.look_back[0])
# look_back_max=int(self.look_back[1])
first_layer_min=round(int(self.first_layer[0]))
first_layer_max=round(int(self.first_layer[1]))
dropout_min=float(self.dropout[0])
dropout_max=float(self.dropout[1])
dropout_step=float(self.dropout[2])
#import pdb; pdb.set_trace()
n_past= self.look_back
n_future = self.look_back
encoder_l = {}
encoder_outputs = {}
encoder_states = {}
decoder_l = {}
decoder_outputs = {}
encoder_inputs = Input(shape=(n_past, n_features))
try:
if(self.hidden_layers > 0):
encoder_l[0] = LSTM(units=hp.Int('enc_input_unit',min_value=first_layer_min,max_value=first_layer_max,step=32), activation = hp.Choice(f'enc_input_activation', values = self.activation_fn), return_sequences = True, return_state=True)
else:
encoder_l[0] = LSTM(units=hp.Int('enc_input_unit',min_value=first_layer_min,max_value=first_layer_max,step=32), activation = hp.Choice(f'enc_input_activation', values = self.activation_fn), return_state=True)
except Exception as e:
import traceback
self.log.info("lstm build traceback: \n"+str(traceback.print_exc()))
model=tf.keras.Sequential()
return model
encoder_outputs[0] = encoder_l[0](encoder_inputs)
encoder_states[0] = encoder_outputs[0][1:]
if(self.hidden_layers > 0):
for indx in range(self.hidden_layers):
lindx = indx + 1
if lindx == self.hidden_layers:
encoder_l[lindx] = LSTM(units=hp.Int(f'enc_lstm_units_{lindx}',min_value=first_layer_min,max_value=first_layer_max,step=32), dropout=hp.Float(f'enc_lstm_dropout_{lindx}',min_value=dropout_min,max_value=dropout_max,step=dropout_step), activation = hp.Choice(f'enc_lstm_activation_{lindx}', values = self.activation_fn), return_state=True)
else:
encoder_l[lindx] = LSTM(units=hp.Int(f'enc_lstm_units_{lindx}',min_value=first_layer_min,max_value=first_layer_max,step=32), dropout=hp.Float(f'enc_lstm_dropout_{lindx}',min_value=dropout_min,max_value=dropout_max,step=dropout_step), activation = hp.Choice(f'enc_lstm_activation_{lindx}', values = self.activation_fn), return_sequences = True, return_state=True)
encoder_outputs[lindx] = encoder_l[lindx](encoder_outputs[indx][0])
encoder_states[lindx] = encoder_outputs[lindx][1:]
decoder_inputs = RepeatVector(n_future)(encoder_outputs[self.hidden_layers][0])
else:
decoder_inputs = RepeatVector(n_future)(encoder_outputs[0][0])
#
if(self.hidden_layers > 0):
decoder_l[0] = LSTM(encoder_states[0][0].get_shape()[1], activation = hp.Choice(f'dec_input_activation', values = self.activation_fn), return_sequences=True)(decoder_inputs,initial_state = encoder_states[0])
else:
decoder_l[0] = LSTM(encoder_states[0][0].get_shape()[1], activation = hp.Choice(f'dec_input_activation', values = self.activation_fn), return_sequences=True)(decoder_inputs,initial_state = encoder_states[0])
if(self.hidden_layers > 0):
for indx in range(self.hidden_layers):
lindx = indx + 1
decoder_l[lindx] = LSTM(encoder_states[lindx][0].get_shape()[1], activation = hp.Choice(f'dec_lstm_activation_{lindx}', values = self.activation_fn), return_sequences=True)(decoder_l[indx],initial_state = encoder_states[lindx])
decoder_outputs[0] = TimeDistributed(tf.keras.layers.Dense(decoder_l[self.hidden_layers][0].get_shape()[1], activation = hp.Choice(f'dec_output_activation_1', values = self.activation_fn)))(decoder_l[self.hidden_layers])
decoder_outputs[1] = TimeDistributed(tf.keras.layers.Dense(1))(decoder_outputs[0])
else:
# decoder_outputs[0] = TimeDistributed(tf.keras.layers.Dense(decoder_l[0][0].get_shape()[1]))(decoder_l[0])
# decoder_outputs[1] = LSTM(200, return_sequences=True)(decoder_outputs[0])
# decoder_outputs[2] = tf.keras.layers.Flatten()(decoder_outputs[1])
# decoder_outputs[3] = tf.keras.layers.Dense(1)(decoder_outputs[2])
decoder_outputs[0] = TimeDistributed(tf.keras.layers.Dense(decoder_l[0][0].get_shape()[1], activation = hp.Choice(f'dec_output_activation_1', values = self.activation_fn)))(decoder_l[0])
decoder_outputs[1] = TimeDistributed(tf.keras.layers.Dense(1))(decoder_outputs[0])
#
model = tf.keras.models.Model(encoder_inputs,decoder_outputs[1])
self.log.info(model.summary())
model.compile(optimizer=optimizer(hp.Choice('learning_rate',values=[1e-1,1e-2,1e-3,1e-4])),loss=loss,metrics=[self.loss_fn])
except Exception as e:
import traceback
self.log.info(",Hyperparam tuning build_model err msg: \n"+ str(e))
self.log.info("Hyperparam tuning build_model err traceback: \n"+str(traceback.print_exc()))
return model
##LSTM ecncoder decoder with multivariate input and univarite output prediction function (lstm model, train, prediction, metrics)
def lstm_encdec_mvin_uvout(self,df):
try:
loss=self.loss_fn
self.getdlparams()
n_features = len(df.columns)
self.num_features=n_features
n_past= self.look_back
n_future = self.look_back
try:
if (type(self.targetFeature) is list):
pass
else:
self.targetFeature = list(self.targetFeature.split(","))
except:
pass
targetColIndx = []
for target in self.targetFeature:
targetColIndx.append(df.columns.get_loc(target))
#if user doesnt applies any transformation, this will get applied
scaler=MinMaxScaler()
df_trnsf=scaler.fit_transform(df)
train_data, test_data = train_test_split(df_trnsf, test_size=0.2, shuffle=False)
tuner_alg=self.tuner_algorithm
#The below create_dataset only for getting best model and best hyperparameters
X_train, y_train = self.create_dataset(train_data, n_past, n_future, targetColIndx)
X_test, y_test = self.create_dataset(test_data, n_past, n_future, targetColIndx)
# X_train = X_train.reshape((X_train.shape[0], X_train.shape[1],n_features))
# y_train = y_train.reshape((y_train.shape[0], y_train.shape[1], 1))
self.log.info("Hyperparameter tuning algorithm is given by user (AION->Advanced configuration -> timeSeriesForecasting->LSTM): \n"+str(tuner_alg))
try:
## Remove untitled_project dir in AION root folder created by previous tuner search run
import shutil
shutil.rmtree(r".\untitled_project")
except:
pass
try:
if (tuner_alg.lower()=="randomsearch"):
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=1,executions_per_trial=3)
elif (tuner_alg.lower()=="bayesianoptimization"):
tuner=BayesianOptimization(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
elif (tuner_alg.lower()=="hyperband"):
tuner=Hyperband(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_epochs=50,factor=3)
else:
self.log.info("The given alg is not implemented. Using default hyperparam tuning algorithm: RandomSearch.\n")
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
from keras.callbacks import EarlyStopping
stop_early = EarlyStopping(monitor='val_loss', patience=5)
except Exception as e:
import traceback
self.log.info("The given alg have some issue, Using default hyperparam tuning algorithm: RandomSearch.\n"+str(e))
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=1,executions_per_trial=3)
self.log.info("Started Exception default Random Search")
#hpt search for best params
try:
self.log.info("First try: Tuner search started")
tuner.search(X_train, y_train,validation_data=(X_test, y_test), callbacks=[stop_early])
self.log.info("First try: Tuner search ends")
except Exception as e:
self.log.info("Second try: Tuner search starts.\n"+str(e))
tuner.search(x=X_train,y=y_train,validation_split=0.2, callbacks=[stop_early])
self.log.info("Second try: Tuner search ends")
# best_model = tuner.get_best_models(num_models=1)[0]
#self.log.info("best_model.summary(): \n"+str(best_model.summary()))
best_hps=tuner.get_best_hyperparameters(num_trials=1)[0]
self.log.info("TS Multivariate LSTM best hyperparameter values:\n"+str(best_hps.values))
self.log.info("Activation fn:\n"+str(self.activation_fn))
n_input=self.look_back
best_hmodel=tuner.hypermodel.build(best_hps)
optimizer=self.optimizer
learning_rate=float(best_hps.get('learning_rate'))
try:
##TFSTask:7033, Added below try block for time series hyperparam tuning, here, for any optimizer, best learning_rate is provided from best_hps.
try:
if optimizer.lower() == "adam":
optimizer=tensorflow.keras.optimizers.Adam(learning_rate=learning_rate)
elif(optimizer.lower() == "adadelta"):
optimizer=tensorflow.keras.optimizers.experimental.Adadelta(learning_rate=learning_rate)
elif(optimizer.lower() == "nadam"):
optimizer=tensorflow.keras.optimizers.experimental.Nadam(learning_rate=learning_rate)
elif(optimizer.lower() == "adagrad"):
optimizer=tensorflow.keras.optimizers.experimental.Adagrad(learning_rate=learning_rate)
elif(optimizer.lower() == "adamax"):
optimizer=tensorflow.keras.optimizers.experimental.Adamax(learning_rate=learning_rate)
elif(optimizer.lower() == "rmsprop"):
optimizer=tensorflow.keras.optimizers.experimental.RMSprop(learning_rate=learning_rate)
elif(optimizer.lower() == "sgd"):
optimizer=tensorflow.keras.optimizers.experimental.SGD(learning_rate=learning_rate)
else:
optimizer=tensorflow.keras.optimizers.Adam(learning_rate=learning_rate)
except:
optimizer=tf.keras.optimizers.Adam(learning_rate=learning_rate)
pass
##From best hyperparameter values, now creating multivariate time series model using time generator.
generatorTrain=TimeseriesGenerator(X_train, y_train, length=n_past, batch_size=self.batch_size)
# generatorTest=TimeseriesGenerator(test,test,length=n_input,batch_size=self.batch_size)
batch_0=generatorTrain[0]
x,y=batch_0
epochs=int(self.epochs)
##Multivariate LSTM model
try:
encoder_l = {}
encoder_outputs = {}
encoder_states = {}
decoder_l = {}
decoder_outputs = {}
enc_lstm_dropout = {}
enc_input_unit = best_hps.get('enc_input_unit')
enc_input_activation = best_hps.get('enc_input_activation')
dec_input_activation = best_hps.get('dec_input_activation')
dec_output_activation_1 = best_hps.get('dec_output_activation_1')
enc_lstm_units = {}
enc_lstm_activation = {}
dec_lstm_activation = {}
for indx in range(self.hidden_layers):
lindx = indx + 1
enc_lstm_units[lindx] = best_hps.get('enc_lstm_units_'+str(lindx))
enc_lstm_activation[lindx] = best_hps.get('enc_lstm_activation_'+str(lindx))
dec_lstm_activation[lindx] = best_hps.get('dec_lstm_activation_'+str(lindx))
enc_lstm_dropout[lindx] = best_hps.get('enc_lstm_dropout_'+str(lindx))
encoder_inputs = Input(shape=(n_past, n_features))
if(self.hidden_layers > 0):
encoder_l[0] = LSTM(enc_input_unit, activation = enc_input_activation, return_sequences = True, return_state=True)
else:
encoder_l[0] = LSTM(enc_input_unit, activation = enc_input_activation, return_state=True)
encoder_outputs[0] = encoder_l[0](encoder_inputs)
encoder_states[0] = encoder_outputs[0][1:]
if(self.hidden_layers > 0):
for indx in range(self.hidden_layers):
lindx = indx + 1
if lindx == self.hidden_layers:
encoder_l[lindx] = LSTM(enc_lstm_units[lindx], dropout = enc_lstm_dropout[lindx], activation = enc_lstm_activation[lindx], return_state=True)
else:
encoder_l[lindx] = LSTM(enc_lstm_units[lindx], dropout = enc_lstm_dropout[lindx], activation = enc_lstm_activation[lindx], return_sequences = True, return_state=True)
encoder_outputs[lindx] = encoder_l[lindx](encoder_outputs[indx][0])
encoder_states[lindx] = encoder_outputs[lindx][1:]
decoder_inputs = RepeatVector(n_future)(encoder_outputs[self.hidden_layers][0])
else:
decoder_inputs = RepeatVector(n_future)(encoder_outputs[0][0])
#
if(self.hidden_layers > 0):
decoder_l[0] = LSTM(encoder_states[0][0].get_shape()[1], activation = dec_input_activation, return_sequences=True)(decoder_inputs,initial_state = encoder_states[0])
else:
decoder_l[0] = LSTM(encoder_states[0][0].get_shape()[1], activation = dec_input_activation, return_sequences=True)(decoder_inputs,initial_state = encoder_states[0])
if(self.hidden_layers > 0):
for indx in range(self.hidden_layers):
lindx = indx + 1
decoder_l[lindx] = LSTM(encoder_states[lindx][0].get_shape()[1], activation = dec_lstm_activation[lindx], return_sequences=True)(decoder_l[indx],initial_state = encoder_states[lindx])
decoder_outputs[0] = TimeDistributed(tf.keras.layers.Dense(decoder_l[self.hidden_layers][0].get_shape()[1], activation = dec_output_activation_1))(decoder_l[self.hidden_layers])
decoder_outputs[1] = TimeDistributed(tf.keras.layers.Dense(1))(decoder_outputs[0])
else:
decoder_outputs[0] = TimeDistributed(tf.keras.layers.Dense(decoder_l[0][0].get_shape()[1], activation = dec_output_activation_1))(decoder_l[0])
decoder_outputs[1] = TimeDistributed(tf.keras.layers.Dense(1))(decoder_outputs[0])
#
model = tf.keras.models.Model(encoder_inputs,decoder_outputs[1])
self.log.info(model.summary())
self.log.info("loss="+self.loss_fn)
model.compile(optimizer=optimizer,loss=self.loss_fn,metrics=[self.loss_fn])
#model.fit_generator(generatorTrain, epochs=epochs,shuffle=False, verbose=0)
model.fit(X_train, y_train, batch_size=self.batch_size, epochs=epochs,shuffle=False, verbose=2)
except Exception as e:
import traceback
self.log.info("multivariate model build error: error msg:: \n"+str(e))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
#predictions = model.predict_generator(generatorTest)
except Exception as e:
import traceback
self.log.info("optimizer and timesereis generator build error: error msg:: \n"+str(e))
self.log.info("optimizer and timesereis generator build error traceback: \n"+str(traceback.print_exc()))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
try:
predictions=[]
X_test, y_test = self.create_dataset(test_data, n_past, n_future, targetColIndx)
predictions = model.predict(X_test)
self.log.info(predictions)
#convert the x test(includes target) to 2d as inverse transform accepts only 2d values
xtestlen = len(X_test)
xtest_2d = X_test.ravel().reshape(xtestlen * n_past, n_features)
#inverse tranform of actual value
xtest_2d = scaler.inverse_transform(xtest_2d)
actual = xtest_2d[:, targetColIndx]
#inverse tranform of predicted value
prediction_1d = predictions.ravel()
prediction_1d = prediction_1d.reshape(len(prediction_1d),1)
self.log.info(prediction_1d)
xtest_2d[:, targetColIndx] = prediction_1d
xtest_2d = scaler.inverse_transform(xtest_2d)
predictions = xtest_2d[:, targetColIndx]
mse=None
rmse=None
## Creating dataframe for actual,predictions
try:
pred_cols=list()
actual_cols=list()
for i in range(len(self.targetFeature)):
pred_cols.append(self.targetFeature[i]+'_pred')
actual_cols.append(self.targetFeature[i]+'_actual')
predictions = pd.DataFrame(predictions.ravel(), columns=pred_cols)
actual = pd.DataFrame(actual.ravel(), columns=actual_cols)
df_predicted=pd.concat([actual,predictions],axis=1)
self.log.info("LSTM Multivariate prediction dataframe: \n"+str(df_predicted))
from math import sqrt
from sklearn.metrics import mean_squared_error
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error
target=self.targetFeature
mse_dict={}
rmse_dict={}
mae_dict={}
mape_dict={}
r2_dict={}
lstm_var = 0
self.log.info(actual.shape)
self.log.info(actual)
self.log.info(predictions.shape)
self.log.info(predictions)
mse = mean_squared_error(actual,predictions)
mse_dict[self.targetFeature[0]]=mse
rmse=sqrt(mse)
rmse_dict[self.targetFeature[0]]=rmse
lstm_var = lstm_var+rmse
self.log.info("Name of the target feature: "+str(self.targetFeature))
self.log.info("RMSE of the target feature: "+str(rmse))
r2 = r2_score(actual,predictions)
r2_dict[self.targetFeature[0]]=r2
mae = mean_absolute_error(actual,predictions)
mae_dict[self.targetFeature[0]]=mae
mape = mean_absolute_percentage_error(actual,predictions)
mape_dict[self.targetFeature[0]]=mape
## For VAR comparison, send last target mse and rmse from above dict
lstm_var = lstm_var/len(target)
select_msekey=list(mse_dict.keys())[-1]
l_mse=list(mse_dict.values())[-1]
select_rmsekey=list(rmse_dict.keys())[-1]
l_rmse=list(rmse_dict.values())[-1]
select_r2key=list(r2_dict.keys())[-1]
l_r2=list(r2_dict.values())[-1]
select_maekey=list(mae_dict.keys())[-1]
l_mae=list(mae_dict.values())[-1]
l_mape=list(mape_dict.values())[-1]
self.log.info("Selected target feature of LSTM for best model selection: "+str(select_rmsekey))
self.log.info("lstm rmse: "+str(l_rmse))
self.log.info("lstm mse: "+str(l_mse))
self.log.info("lstm r2: "+str(l_r2))
self.log.info("lstm mae: "+str(l_mae))
self.log.info("lstm mape: "+str(l_mape))
except Exception as e:
import traceback
self.log.info("prediction error traceback: \n"+str(traceback.print_exc()))
except Exception as e:
import traceback
self.log.info("dataframe creation error. err.msg: "+str(e))
self.log.info("dataframe creation error traceback: \n"+str(traceback.print_exc()))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
return 'Success',round(l_mse,2),round(l_rmse,2),round(l_r2,2),round(l_mae,2),model,df_predicted,n_input,scaler
# import os
#predicted_file_name='lstm_prediction_df.csv'
#predicted_file_path=os.path.join(self.dataFolderLocation,predicted_file_name)
#df_predicted.to_csv(predicted_file_path)
##save model
#model_path = os.path.join(self.dataFolderLocation,self.model_name)
#self.log.info("mlp model saved at: "+str(model_path))
#model.save(model_path)
except Exception as e:
import traceback
## Just use below traceback print to get detailed error information.
# import traceback
# print(" traceback error 7:\n",traceback.print_exc())
## Enable traceback for debugging
self.log.info("dataframe creation error. err.msg: "+str(e))
self.log.info("Final exception traceback: \n"+str(traceback.print_exc()))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
|
ts_modelvalidation.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import json
#Python sklearn & std libraries
import numpy as np
import pandas as pd
from time_series.ts_arima_eion import eion_arima
from statsmodels.tsa.vector_ar.vecm import coint_johansen
from statsmodels.tsa.vector_ar.var_model import VAR
from math import *
from sklearn.metrics import mean_squared_error
from sklearn.metrics import mean_absolute_error
from math import sqrt
import logging
import os
import sys
import time
from statsmodels.tsa.arima_model import ARIMA
from sklearn.metrics import mean_squared_error
from pandas import read_csv
from statsmodels.tsa.stattools import adfuller
import pmdarima as pm
from statsmodels.tsa.stattools import grangercausalitytests
from statsmodels.stats.stattools import durbin_watson
from sklearn.utils import check_array
class timeseriesModelTests():
def __init__(self,data,targetFeature,datetimeFeature,count):
#self.tsConfig = tsConfig
#self.modelconfig = modelconfig
#self.modelList = modelList
self.data = data
self.targetFeature = targetFeature
self.dateTimeFeature = datetimeFeature
self.count=count
self.log = logging.getLogger('eion')
def StatinaryChecks(self,dictDiffCount):
self.log.info("\n---------------Start Stationary Checks-----------")
tFeature = self.targetFeature.split(',')
tFeature.append(self.dateTimeFeature)
self.data=self.data[tFeature]
tFeature.remove(self.dateTimeFeature)
lengthtFeature=len(tFeature)
diffCount=0
try :
for features in (tFeature):
XSt = self.data[features]
XSt=XSt.values
resultSt = adfuller(XSt,autolag='AIC')
stationaryFlag = False
#print(resultSt)
self.log.info('-------> Features: '+str(features))
self.log.info('----------> ADF Statistic: '+str(resultSt[0]))
self.log.info('----------> p-value: %f' % resultSt[1])
if resultSt[1]<= 0.05:
self.log.info("-------------> Converted As Stationary Data")
stationaryFlag = True
else:
self.log.info("-------------> Stationary Conversion Required")
stationaryFlag = False
self.log.info('----------> Critical Values')
for key, value in resultSt[4].items():
self.log.info('----------> '+str(key)+': '+str(value))
if stationaryFlag == False:
self.data[features]=self.data[features].diff()
self.data=self.data.dropna()
dictDiffCount[features]=1
XStt = self.data[features]
XStt=XStt.values
resultStt = adfuller(XStt)
if resultStt[1] > 0.05:
self.data[features]=self.data[features].diff()
self.data=self.data.dropna()
dictDiffCount[features]=2
XSttt = self.data[features]
XSttt=XSttt.values
resultSttt = adfuller(XSttt)
if resultSttt[1]<= 0.05:
stationaryFlag = True
else:
stationaryFlag = True
self.log.info("------------->"+str(dictDiffCount))
if stationaryFlag == True:
self.log.info("----------> Equals to Stationary Data")
else:
self.log.info("----------> Not Equal To Stationary Data")
self.log.info("-------> Stationary data diff()")
self.log.info(dictDiffCount)
self.log.info("---------------Start Stationary Checks Ends-----------\n")
return self.data,dictDiffCount
except Exception as inst:
self.log.info('<!------------- Time Series Stationary Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def varTimeseriesModelTests(self,data):
try :
tFeature = self.targetFeature.split(',')
self.log.info("\n--------- Start Granger Causality Test Results ------------")
gtest=grangercausalitytests(data[tFeature], maxlag=15, addconst=True, verbose=True)
self.log.info("-------> GrangerCausalitytest Results "+str(gtest.values()))
self.log.info("--------- End Granger Causality Test Results ------------\n")
return gtest
except Exception as inst:
self.log.info('<!------------- Time Series Granger Causality testTest Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def grangersCausationMatrix(self,data, variables, test='ssr_chi2test', verbose=False):
try :
countVariables=0
self.log.info(len(variables))
self.log.info("\n--------------Start GrangersCausationMatrix---------------")
df = pd.DataFrame(np.zeros((len(variables), len(variables))), columns=variables, index=variables)
for c in df.columns:
for r in df.index:
test_result = grangercausalitytests(data[[r, c]], maxlag=12, verbose=False)
p_values = [round(test_result[i+1][0][test][1],4) for i in range(12)]
if verbose: print(f'Y = {r}, X = {c}, P Values = {p_values}')
min_p_value = np.min(p_values)
df.loc[r, c] = min_p_value
df.columns = [var + '_x' for var in variables]
df.index = [var + '_y' for var in variables]
self.log.info(df)
for i in range(len(variables)):
for j in range(len(variables)):
if i!=j and df.iloc[i][j]<0.05 and df.iloc[i][j]<0.05:
countVariables=countVariables+1
self.log.info("--------------End GrangersCausationMatrix---------------\n")
return df,countVariables
except Exception as inst:
self.log.info('<!------------- Time Series grangersCausationMatrix Test Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return df,countVariables
def coIntegrationTest(self,data):
try :
tdata = data.drop([self.dateTimeFeature], axis=1)
tdata.index = data[self.dateTimeFeature]
cols = tdata.columns
self.log.info("\n-------------- Start of the Co Integration test ---------------")
lenTargetFeature=len(self.targetFeature)
countIntegrationFeature=0
N, l = tdata.shape
jres = coint_johansen(tdata, 0, 1)
trstat = jres.lr1
tsignf = jres.cvt
for i in range(l):
if trstat[i] > tsignf[i, 1]:
r = i + 1
jres.r = r
jres.evecr = jres.evec[:, :r]
jres.r = r
countIntegrationFeature=jres.r
jres.evecr = jres.evec[:, :r]
self.log.info('------->coint_johansen trace statistics: '+str(trstat))
self.log.info('------->coint_johansen critical values:')
self.log.info(tsignf)
self.log.info("------->There are "+str(countIntegrationFeature)+" Co-Integration vectors")
self.log.info("-------------- End of the Co Integration test ---------------\n")
return countIntegrationFeature
except Exception as inst:
self.log.info('<!------------- Time Series Co-Integration Test Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
tsStationarySeasonalityTest.py | import pandas as pd
import numpy as np
from statsmodels.tsa.stattools import adfuller
from statsmodels.tsa.stattools import kpss
from statsmodels.tsa.seasonal import seasonal_decompose
import logging
import os
import warnings
warnings.filterwarnings('ignore')
## Main class to find out seassonality and stationary in timeseries data.
class tsStationarySeasonalityTest:
def __init__(self,df,deployLocation):
self.df=df
self.deployLocation=deployLocation
self.log = logging.getLogger('eion')
## to get the timeseries data stationary information
def stationary_model(self,df,target_features,stationary_check_method):
self.log.info("<------ Time Series stationary test started.....------------->\n")
self.log.info("<------ Feature used:------------->\t"+str(target_features))
stationary_status=None
if (stationary_check_method.lower()=='adfuller'):
stats_model=adfuller(df[target_features])
# p_val=adf_result[1]
statistic, p_value, n_lags, num_bservations,critical_values,info_criterion_best=stats_model[0],stats_model[1],stats_model[2],stats_model[3],stats_model[4],stats_model[5]
##Uncomment below logs when required.
self.log.info("Adfuller test (time series stationary test) p_value: \t"+str(p_value))
# self.log.info("Adfuller test (time series stationary test) statistics: \t"+str(statistic))
# self.log.info("Adfuller test (time series stationary test) number of lags (time steps): \t"+str(n_lags))
# self.log.info("Adfuller test (time series stationary test) Critical values: \n")
##To display critical values
# for key, value in stats_model[4].items():
# self.log.info(" \t"+str(key)+"\t"+str(value))
if (p_value>0.05):
stationary_status="feature is non-stationary"
self.log.info('Status:- |... '+str(target_features)+' is non stationary')
elif(p_value<0.05):
stationary_status="feature is stationary"
self.log.info('Status:- |... '+str(target_features)+' is stationary')
##kpss is opposite to ADF in considering null hypothesis. In KPSS, if null hypothesis,then it is stationary as oppose to ADF.
elif (stationary_check_method.lower()=='kpss'):
from statsmodels.tsa.stattools import kpss
stats_model = kpss(df[target_features])
statistic, p_value, n_lags, critical_values=stats_model[0],stats_model[1],stats_model[2],stats_model[3]
self.log.info("kpss test (time series stationary test) p_value: \t"+str(p_value))
self.log.info("kpss test (time series stationary test) statistics: \t"+str(statistic))
self.log.info("kpss test (time series stationary test) number of lags (time steps): \t"+str(n_lags))
self.log.info("kpss test (time series stationary test) Critical values: \n")
for key, value in stats_model[3].items():
self.log.info(" \t"+str(key)+"\t"+str(value))
##In kpss, the stationary condition is opposite to Adafuller.
if (p_value>0.05):
self.log.info('Status:- |... '+str(target_features)+' is stationary')
else:
self.log.info('Status:- |... '+str(target_features)+' is non stationary')
return stats_model,n_lags,p_value,stationary_status
## Get stationary details
def stationary_check(self,target_features,time_col,method):
df=self.df
try:
df[time_col]=pd.to_datetime(df[time_col])
except Exception as e:
self.log.info("issue in datetime conversion...\n"+str(e))
df=df.set_index(time_col)
try:
stationary_check_method=method
except:
stationary_check_method='adfuller'
if (len(target_features) == 1):
try:
if isinstance(target_features,list):
target_features=''.join(target_features)
elif isinstance(target_features,int):
target_features=str(target_features)
elif isinstance(target_features,str):
pass
except Exception as e:
self.log.info("stationary check target feature error: \t"+str(e))
stationary_result={}
stats_model,n_lags,p_value,stationary_status=self.stationary_model(df,target_features,stationary_check_method)
stationary_result[target_features]=stationary_status
elif(len(target_features) > 1):
stationary_result={}
for col in df.columns:
# self.log.info("Multivariate feature for Stationary check:\t"+str(col))
stats_model,n_lags,p_value,stationary_status=self.stationary_model(df,col,stationary_check_method)
stationary_result[col]=stationary_status
else:
self.log.info("TS Stationarity Test: Error in target feature, pls check.\n.")
# self.log.info("Feature based stationarity_result:\n"+str(stationary_result))
# ## Stationary component for whole dataset
stationary_combined_res=dict()
# stats_model,n_lags,p_value,stationary_status=self.stationary_all_features(time_col,'adfuller')
c_dict=[k for k,v in stationary_result.items() if 'non-stationary' in v]
if (len(c_dict)>=1):
stationary_combined_res['dataframe_stationarity']='Non-Stationary'
self.log.info('Status:- |... Data is non stationarity')
else:
stationary_combined_res['dataframe_stationarity']='Stationary'
# self.log.info("Stationarity information for whole dataset:\n"+str(stationary_combined_res))
self.log.info("Time series Stationarity test completed.\n")
return stats_model,n_lags,p_value,stationary_result,stationary_combined_res
#Get seasonality by using seasonal_decompose lib.
def seasonality_model(self,target_features,df):
self.log.info("<------ Time Series Seasonality test started.....------------->\n")
self.log.info("<------ Feature used:------------->\n"+str(target_features))
seasonality_status=None
try:
try:
stats_model = kpss(df[target_features])
statistic, p_value, n_lags, critical_values=stats_model[0],stats_model[1],stats_model[2],stats_model[3]
except:
n_lags=1
pass
try:
df_target=self.df[target_features]
decompose_result_mult = seasonal_decompose(df_target,model='additive', extrapolate_trend='freq', period=n_lags)
except Exception as e:
self.log.info("Logging seasonality_model decompose_result_mult: \t"+str(e))
##If additive model (type of seasonal component) failed, use multiplicative
decompose_result_mult = seasonal_decompose(df_target,model='multiplicative', extrapolate_trend='freq', period=1)
trend = decompose_result_mult.trend
observed=decompose_result_mult.observed
seasonal = decompose_result_mult.seasonal
residual = decompose_result_mult.resid
try:
if isinstance(df_target, pd.Series):
auto_correlation = df_target.autocorr(lag=n_lags)
# self.log.info("seasonality test: auto_correlation value:\n"+str(auto_correlation))
elif isinstance(df_target, pd.DataFrame):
df_target = df_target.squeeze()
auto_correlation = df_target.autocorr(lag=n_lags)
# self.log.info("seasonality test: auto_correlation value:\n"+str(auto_correlation))
except:
pass
self.log.info("<------------------ Time series Seasonality test result:------------------>")
if (seasonal.sum()==0):
seasonality_status="feature don't have seasonality (non seasonality)."
self.log.info('Status:- |... '+str(target_features)+' does not have seasonality')
self.log.info("<----- The model feature: "+str(target_features)+" does not have significant seasonality.----->\n")
else:
seasonality_status="feature has seasonality."
self.log.info('Status:- |... '+str(target_features)+' have seasonality')
##Please use the below plot for GUI show (seasonality components)
# decompose_result_mult.plot()
df['observed'] = decompose_result_mult.observed
df['residual'] = decompose_result_mult.resid
df['seasonal'] = decompose_result_mult.seasonal
df['trend'] = decompose_result_mult.trend
df_name='timeseries_seasonality_check_'+f"{target_features}"+'.csv'
dir_n = os.path.join(self.deployLocation,'data','seasonality')
if not os.path.exists(dir_n):
os.makedirs(dir_n)
model_path=os.path.join(dir_n,df_name)
self.log.info("Seasonality information saved as dataframe at:\t "+str(model_path))
## Seasonal component for whole dataset
df.to_csv(model_path)
except Exception as e:
self.log.info("Seasonality function exception: \t"+str(e))
return df,decompose_result_mult,seasonality_status
##Main function to check seasonlity in data
def seasonal_check(self,target_features,time_col,seasonal_model):
df=self.df
# self.log.info("seasonal check started... \n")
try:
df[time_col]=pd.to_datetime(df[time_col])
except Exception as e:
self.log.info("Issue in datetime conversion...\n"+str(e))
df=df.set_index(time_col)
if (len(target_features)==1):
try:
if isinstance(target_features,list):
target_features=''.join(target_features)
elif isinstance(target_features,int):
target_features=str(target_features)
elif isinstance(target_features,str):
pass
except Exception as e:
self.log.info("stationary check target feature error: \t"+str(e))
## Seasonal component for individual feature based.
seasonality_result=dict()
df,decompose_result_mult,seasonality_status = self.seasonality_model(target_features,df)
seasonality_result[target_features]=seasonality_status
elif(len(target_features) > 1):
seasonality_result=dict()
self.log.info("TS seasonality Test: The problem type is time series Multivariate.")
for col in df.columns:
df,decompose_result_mult,seasonality_status = self.seasonality_model(col,df)
seasonality_result[col]=seasonality_status
else:
self.log.info("TS seasonality Test: Error in target feature, pls check.\n.")
# self.log.info("Feature based seasonality_result:\n"+str(seasonality_result))
# ## Seasonal component for whole dataset
seasonality_combined_res=dict()
c_dict=[k for k,v in seasonality_result.items() if 'non seasonality' in v]
if (len(c_dict)>=1):
seasonality_combined_res['dataframe_seasonality']='No Seasonal elements'
else:
seasonality_combined_res['dataframe_seasonality']='contains seasonal elements.'
# self.log.info("Seasonality information for whole dataset:\n"+str(seasonality_combined_res))
self.log.info("Time series Seasonality test completed.\n")
return df,decompose_result_mult,seasonality_result,seasonality_combined_res
#Main fn for standalone test purpose
if __name__=='__main__':
print("Inside seasonality-stationary test main function...")
print("Below code used for standalone test purpose.")
# df=pd.read_csv(r"C:\AION_Works\Data\order_forecast_ts.csv")
# print("df info: \n",df.info())
# df=df.drop('index',axis=1)
# time_col="DateTime"
# target='order1'
# stationary_method='adfuller'
# seasonal_model="additive" ## two models are available: 1.multiplicative, 2.additive
# if (isinstance(target,list)):
# pass
# elif (isinstance(target,str)):
# target=list(target.split(','))
# cls_ins=aion_ts_stationary_seassonality_test(df)
# stats_model,n_lags,p_value=cls_ins.stationary_check(target,time_col,stationary_method)
# df,decompose_result_mult=cls_ins.seasonal_check(target,time_col,seasonal_model)
# print(" Time series stationary and seasonality check completed.") |
timeseriesDLUnivariate.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import pandas as pd
# import os
import tensorflow as tf
import numpy as np
from sklearn.preprocessing import MinMaxScaler
from sklearn.model_selection import train_test_split
import math
from sklearn.metrics import mean_squared_error
from tensorflow.keras.layers import Dense
from tensorflow.keras.layers import Dropout
from tensorflow.keras import Sequential
from tensorflow.keras.layers import LSTM
import logging
# import kerastuner
import keras_tuner
#from keras_tuner.engine.hyperparameters import HyperParameters
from keras_tuner.tuners import RandomSearch,BayesianOptimization ,Hyperband
import warnings
warnings.simplefilter("ignore", UserWarning)
# from keras.models import load_model
# from tensorflow.keras.optimizers import SGD
# from tensorflow.keras.utils import load_model
from tensorflow.keras.models import load_model
class timeseriesDLUnivariate:
def __init__(self,configfile,testpercentage,targetFeature,dateTimeFeature,modelName):
self.look_back=None
#Preprocessed dataframe
# self.df=df
self.savedmodelname=None
self.deploy_location=None
self.epochs=None
self.batch_size=None
self.hidden_layers=None
self.optimizer=None
self.activation_fn=None
self.loss_fn=None
self.first_layer=None
self.dropout=None
self.model_name=None
self.hpt_train=None
##Below is model type (MLP or lstm)
self.model_type=modelName
#self.dataFolderLocation=str(dataFolderLocation)
##Added for ts hpt
self.tuner_algorithm=""
self.dl_params = configfile
# self.data=data
self.targetFeature=targetFeature
self.dateTimeFeature=dateTimeFeature
self.testpercentage = testpercentage
self.log = logging.getLogger('eion')
#To extract dict key,values
def extract_params(self,dict):
self.dict=dict
for k,v in self.dict.items():
return k,v
##Get deep learning model hyperparameter from advanced config
def getdlparams(self):
val=self.dl_params
self.log.info('-------> The given mlp/lstm timeseries algorithm parameters:>>')
self.log.info(" "+str(val))
for k,v in val.items():
try:
if (k == "tuner_algorithm"):
self.tuner_algorithm=str(v)
elif (k == "activation"):
self.activation_fn=str(v)
elif (k == "optimizer"):
self.optimizer=str(v)
elif (k == "loss"):
self.loss_fn=str(v)
elif (k == "first_layer"):
if not isinstance(k,list):
self.first_layer=str(v).split(',')
else:
self.first_layer=k
elif (k == "lag_order"):
if isinstance(k,list):
k = ''.join(v)
k=int(float(str(v)))
else:
self.look_back=int(float(str(v)))
elif (k == "hidden_layers"):
self.hidden_layers=int(v)
elif (k == "dropout"):
if not isinstance(k,list):
self.dropout=str(v).split(',')
else:
self.dropout=k
elif (k == "batch_size"):
self.batch_size=int(v)
elif (k == "epochs"):
self.epochs=int(v)
elif (k == "model_name"):
self.model_name=str(v)
except Exception as e:
self.log.info('Exception occured in deeep learn param reading, setting up default params.')
self.activation_fn="relu"
self.optimizer="adam"
self.loss_fn="mean_squared_error"
self.first_layer=[8,512]
self.hidden_layers=1
self.look_back=int(2)
self.dropout=[0.1,0.5]
self.batch_size=2
self.epochs=50
self.model_name="lstmmodel.h5"
continue
## Just use this if user need to create dataframe from input data.
def createdf(self,df):
target=""
# splitting reframed to X and Y considering the first column to be out target featureX=reframed.drop(['var1(t)'],axis=1)
X=df.drop([target],axis=1)
Y=df[target]
X_values=X.values
Y_values=Y.values
n_predict=len(Y_values)
train_X,train_Y = X_values[:(X_values.shape[0]-n_predict),:],Y_values[:(X_values.shape[0]-n_predict)]
test_X,test_Y = X_values[(X_values.shape[0]-n_predict):,:],Y_values[(X_values.shape[0]-n_predict):]
#reshaping train and test to feed to LSTM
train_X = train_X.reshape((train_X.shape[0], 1, train_X.shape[1]))
test_X = test_X.reshape((test_X.shape[0], 1, test_X.shape[1]))
return train_X,train_Y,test_X,test_Y
# convert an array of values into a dataset matrix
def numpydf(self,dataset, look_back):
dataX, dataY = [], []
for i in range(len(dataset)-look_back-1):
a = dataset[i:(i+look_back), 0]
dataX.append(a)
dataY.append(dataset[i + look_back, 0])
# x,y=numpy.array(dataX), numpy.array(dataY)
return np.array(dataX), np.array(dataY)
def model_save(self,model):
import os.path
savedmodelname=self.model_name
path = os.path.join(self.deploy_location,savedmodelname)
model.save(path)
return (savedmodelname)
## MLP model buid
def mlpDL(self,df):
self.log.info("MLP timeseries learning starts.....")
try:
self.getdlparams()
# look_back = self.look_back
dataset = df.values
dataset = dataset.astype('float32')
##The below Kwiatkowski-Phillips-Schmidt-Shin (kpss) statsmodel lib used for stationary check as well getting number of lags.
##number of lag calculated just for reference ,not used now.
#Dont delete this, just use in future.
from statsmodels.tsa.stattools import kpss
statistic, p_value, n_lags, critical_values = kpss(df[self.targetFeature])
self.log.info("Based on kpss statsmodel, lag order (time steps to calculate next prediction) is: \t"+str(n_lags))
scaler = MinMaxScaler(feature_range=(0, 1))
dataset = scaler.fit_transform(dataset)
# split into train and test sets
train_size = int(len(dataset) * 0.80)
test_size = len(dataset) - train_size
train, test = dataset[0:train_size,:], dataset[train_size:len(dataset),:]
self.hpt_train=train
tuner_alg=self.tuner_algorithm
try:
## Remove untitled_project dir in AION root folder created by previous tuner search run
import shutil
shutil.rmtree(r".\untitled_project")
except:
pass
if (tuner_alg.lower()=="randomsearch"):
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
elif (tuner_alg.lower()=="bayesianoptimization"):
tuner=BayesianOptimization(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
elif (tuner_alg.lower()=="hyperband"):
tuner=Hyperband(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_epochs=50,factor=3)
# tuner.search(X[...,np.new_axis],y,epochs=2,validation_data=(y[...,np.newaxis]))
stop_early = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5)
try:
tuner.search(x=train,y=train,validation_data=(test,test),callbacks=[stop_early])
except:
tuner.search(x=train,y=train,validation_split=0.2,callbacks=[stop_early])
# best_model=tuner.get_best_models(num_models=1)[0]
best_hps=tuner.get_best_hyperparameters(num_trials=1)[0]
best_first_layer=best_hps.get('units')
best_dropout=best_hps.get('Dropout_rate')
best_learning_rate=float(best_hps.get('learning_rate'))
self.log.info("best hyperparameter values for mlp: \n"+str(best_hps.values))
look_back = 1 ## Because univariate problemtype
trainX, trainY = self.numpydf(train, look_back)
testX, testY = self.numpydf(test, look_back)
best_hmodel=tuner.hypermodel.build(best_hps)
##Added for mlp issue,because tuner build also need to compile.
try:
best_hmodel.compile(loss=self.loss_fn, optimizer=self.optimizer)
except:
pass
model_fit = best_hmodel.fit(trainX, trainY, epochs=self.epochs, batch_size=self.batch_size, verbose=2)
val_acc_per_epoch = model_fit.history['loss']
best_epoch = val_acc_per_epoch.index(min(val_acc_per_epoch)) + 1
self.log.info("MLP best epochs value:\n"+str(best_epoch))
trainScore = best_hmodel.evaluate(trainX, trainY, verbose=0)
testScore = best_hmodel.evaluate(testX, testY, verbose=0)
#Scoring values for the model
mse_eval=testScore
try:
#If mse_eval is list of values
min_v=min(mse_eval)
except:
#If mse_eval is single value
min_v=mse_eval
rmse_eval = math.sqrt(min_v)
# generate predictions for training
trainPredict = best_hmodel.predict(trainX)
#print(testX)
testPredict = best_hmodel.predict(testX)
#print(testPredict)
# invert predictions, because we used mimanmax scaler
trainY = scaler.inverse_transform([trainY])
trainPredict = scaler.inverse_transform(trainPredict)
## For test data
testY = scaler.inverse_transform([testY])
testPredict = scaler.inverse_transform(testPredict)
## Creating dataframe for actual,predictions
predictions = pd.DataFrame(testPredict, columns=[self.targetFeature+'_pred'])
actual = pd.DataFrame(testY.T, columns=[self.targetFeature+'_actual'])
df_predicted=pd.concat([actual,predictions],axis=1)
#print(df_predicted)
from math import sqrt
from sklearn.metrics import mean_squared_error
try:
mse_mlp = mean_squared_error(testY.T,testPredict)
rmse_mlp=sqrt(mse_mlp)
self.log.info('mse_mlp: '+str(mse_mlp))
self.log.info('rmse_mlp: '+str(rmse_mlp))
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error
r2 = r2_score(testY.T,testPredict)
mae = mean_absolute_error(testY.T,testPredict)
self.log.info('r2_mlp: '+str(r2))
self.log.info('mae_mlp: '+str(mae))
except Exception as e:
import traceback
self.log.info("MLP dataframe creation error traceback: \n"+str(traceback.print_exc()))
self.log.info(e)
# df_predicted.to_csv('mlp_prediction.csv')
except Exception as e:
self.log.info("MLP timeseries model traceback error msg e: "+str(e))
self.log.info("MLP training successfully completed.\n")
return mse_mlp,rmse_mlp,r2,mae,best_hmodel,df_predicted,look_back,scaler
## Added function for hyperparam tuning (TFSTask:7033)
def build_model(self,hp):
try:
loss=self.loss_fn
optimizer=self.optimizer
try:
if optimizer.lower() == "adam":
optimizer=tf.keras.optimizers.Adam
elif(optimizer.lower() == "adadelta"):
optimizer=tf.keras.optimizers.experimental.Adadelta
elif(optimizer.lower() == "nadam"):
optimizer=tf.keras.optimizers.experimental.Nadam
elif(optimizer.lower() == "adagrad"):
optimizer=tf.keras.optimizers.experimental.Adagrad
elif(optimizer.lower() == "adamax"):
optimizer=tf.keras.optimizers.experimental.Adamax
elif(optimizer.lower() == "rmsprop"):
optimizer=tf.keras.optimizers.experimental.RMSprop
elif(optimizer.lower() == "sgd"):
optimizer=tf.keras.optimizers.experimental.SGD
else:
optimizer=tf.keras.optimizers.Adam
except:
optimizer=tf.keras.optimizers.Adam
pass
first_layer_min=round(int(self.first_layer[0]))
first_layer_max=round(int(self.first_layer[1]))
dropout_min=float(self.dropout[0])
dropout_max=float(self.dropout[1])
model=tf.keras.Sequential()
if (self.model_type.lower() == 'lstm'):
model.add(LSTM(units=hp.Int('units',min_value=first_layer_min,max_value=first_layer_max,step=16),input_shape=(self.look_back,self.hpt_train.shape[1]),
activation=hp.Choice('dense_activation',values=['relu'])))
elif (self.model_type.lower() == 'mlp'):
# model.add(Dense(units=hp.Int('units',min_value=first_layer_min,max_value=first_layer_max,step=16),input_dim=(hp.Int('time_steps',min_value=look_back_min,max_value=look_back_max,step=1)),
# activation='relu'))
##input_dim is 1 because mlp is for univariate.
model.add(Dense(units=hp.Int('units',min_value=first_layer_min,max_value=first_layer_max,step=16),input_dim=(1),activation='relu'))
model.add(Dropout(hp.Float('Dropout_rate',min_value=dropout_min,max_value=dropout_max,step=0.1)))
model.add(Dense(units=1))
model.compile(optimizer=optimizer(hp.Choice('learning_rate',values=[1e-1,1e-2,1e-3,1e-4])),loss=loss,metrics=[loss])
except Exception as e:
import traceback
self.log.info("lstm errorbuild_model traceback: \n"+str(traceback.print_exc()))
return model
##LSTM timeseries function call
def ts_lstm(self,df):
self.log.info("lstm network model learning starts.....\n")
try:
self.getdlparams()
dataset = df.values
dataset = dataset.astype('float32')
##The below Kwiatkowski-Phillips-Schmidt-Shin (kpss) statsmodel lib used for stationary check as well getting number of lags.
##number of lag calculated just for reference ,not used now.
#Dont delete this, just use in future.
from statsmodels.tsa.stattools import kpss
statistic, p_value, n_lags, critical_values = kpss(df[self.targetFeature])
self.log.info("Based on kpss statsmodel, lag order (time steps to calculate next prediction) is: \t"+str(n_lags))
# normalize the dataset
scaler = MinMaxScaler(feature_range=(0, 1))
dataset = scaler.fit_transform(dataset)
# split into train and test sets
train_size = int(len(dataset) * 0.80)
test_size = len(dataset) - train_size
train, test = dataset[0:train_size,:], dataset[train_size:len(dataset),:]
self.hpt_train=train
tuner_alg=self.tuner_algorithm
try:
## Remove untitled_project dir in AION root folder created by previous tuner search run
import shutil
shutil.rmtree(r".\untitled_project")
except:
pass
if (tuner_alg.lower()=="randomsearch"):
tuner=RandomSearch(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
elif (tuner_alg.lower()=="bayesianoptimization"):
tuner=BayesianOptimization(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_trials=5,executions_per_trial=3)
elif (tuner_alg.lower()=="hyperband"):
tuner=Hyperband(self.build_model,keras_tuner.Objective("val_loss", direction="min"),max_epochs=50,factor=3)
# tuner.search(X[...,np.new_axis],y,epochs=2,validation_data=(y[...,np.newaxis]))
from keras.callbacks import EarlyStopping
stop_early = EarlyStopping(monitor='val_loss', patience=5)
##Need both x and y with same dimention.
tuner.search(x=train,y=train,validation_split=0.2,callbacks=[stop_early])
# tuner.search(x=train,y=test,validation_data=(test,test),callbacks=[stop_early])
best_hps=tuner.get_best_hyperparameters(num_trials=1)[0]
best_time_steps=self.look_back
self.log.info("best lag order or lookback (time_steps) for LSTM: \n"+str(best_time_steps))
self.log.info("best hyperparameter values for LSTM: \n"+str(best_hps.values))
look_back = best_time_steps
trainX, trainY = self.numpydf(train, look_back)
testX, testY = self.numpydf(test, look_back)
# reshape input to be [samples, time steps, features]
trainX = np.reshape(trainX, (trainX.shape[0], trainX.shape[1], 1))
testX = np.reshape(testX, (testX.shape[0], testX.shape[1], 1))
#create and fit the LSTM network
best_hmodel=tuner.hypermodel.build(best_hps)
try:
best_hmodel.compile(loss=self.loss_fn, optimizer=self.optimizer)
except:
pass
model_fit = best_hmodel.fit(trainX, trainY, validation_split=0.2, epochs=self.epochs, batch_size=self.batch_size, verbose=2)
val_acc_per_epoch = model_fit.history['loss']
best_epoch = val_acc_per_epoch.index(min(val_acc_per_epoch)) + 1
self.log.info("best epochs value:\n"+str(best_epoch))
# best_hmodel=tuner.hypermodel.build(best_hps)
# best_hmodel.fit(x=trainX,y=trainY,validation_split=0.2,epochs=best_epoch)
##Using model_evaluate,calculate mse
# mse_eval = model.evaluate(testX, testY, verbose=0)
mse_eval = best_hmodel.evaluate(testX, testY, verbose=0)
try:
#If mse_eval is list of values
min_v=min(mse_eval)
except:
#If mse_eval is single value
min_v=mse_eval
rmse_eval=math.sqrt(min_v)
# self.log.info('LSTM mse:'+str(mse_eval))
# self.log.info('LSTM rmse:'+str(rmse_eval))
# lstm time series predictions
trainPredict = best_hmodel.predict(trainX)
testPredict = best_hmodel.predict(testX)
# invert predictions, because we used mim=nmax scaler
trainY = scaler.inverse_transform([trainY])
trainPredict = scaler.inverse_transform(trainPredict)
testY = scaler.inverse_transform([testY])
testPredict = scaler.inverse_transform(testPredict)
## Creating dataframe for actual,predictions
predictions = pd.DataFrame(testPredict, columns=[self.targetFeature+'_pred'])
actual = pd.DataFrame(testY.T, columns=[self.targetFeature+'_actual'])
df_predicted=pd.concat([actual,predictions],axis=1)
from math import sqrt
from sklearn.metrics import mean_squared_error
try:
mse_lstm=None
mse_lstm = mean_squared_error(testY.T,testPredict)
rmse_lstm=sqrt(mse_lstm)
self.log.info("mse_lstm: "+str(mse_lstm))
self.log.info("rmse_lstm: "+str(rmse_lstm))
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error
r2 = r2_score(testY.T,testPredict)
mae = mean_absolute_error(testY.T,testPredict)
self.log.info('r2_lstm: '+str(r2))
self.log.info('mae_lstm: '+str(mae))
except Exception as e:
self.log.info("lstm error loss fns"+str(e))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
except Exception as e:
import traceback
self.log.info("lstm training error traceback: \n"+str(traceback.print_exc()))
return 'Error',0,0,0,0,None,pd.DataFrame(),0,None
return 'Success',mse_lstm,rmse_lstm,r2,mae,best_hmodel,df_predicted,look_back,scaler
if __name__ == '__main__':
print('Inside timeseriesDLUnivariate main....\n')
# tsdl_obj = timeseriesDLUnivariate()
## for testing purpose
'''
df1= pd.read_csv(r"C:\aiontest\testPrograms\Data\energydemand.csv",encoding='utf-8', engine='python')
dateTimeFeature = "utcTimeStamp"
targetFeature="temperature"
try:
df1[dateTimeFeature] = pd.to_datetime(df1[dateTimeFeature]) #, format = '%d/%m/%Y %H.%M')
except:
pass
tdata = df1.drop([dateTimeFeature], axis=1)
tdata.index = df1[dateTimeFeature]
tdata = pd.DataFrame(tdata[targetFeature])
cols = tdata.columns
mse,rmse,model = tsdl_obj.mlpDL(tdata)
lmse,lrmse,lstmmodel = tsdl_obj.ts_lstm(tdata)
print("mlp mse: \n",mse)
print("mlp rmse: \n",rmse)
print("lstm mse: \n",lmse)
print("lstm rmse: \n",lrmse)
savedmodelname=tsdl_obj.model_save(lstmmodel)
'''
|
aion_fbprophet.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
# For timeseries pyramid pdaarima module
import json
#Python sklearn & std libraries
import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.feature_selection import VarianceThreshold
from sklearn.metrics import mean_squared_error, r2_score, mean_absolute_error
#from sklearn.metrics import mean_absolute_percentage_error
from sklearn.linear_model import LinearRegression
from math import sqrt
import warnings
# For serialization.
#from sklearn.externals import joblib
import pickle
import os,sys
# For ploting (mathlab)
import matplotlib.pyplot as plt
import plotly
#Import eion config manager module
import logging
from sklearn import metrics
from sklearn.metrics import accuracy_score
import time
import random
import statsmodels.api as sm
# prophet by Facebook
# time series analysis
#from statsmodels.tsa.seasonal import seasonal_decompose
#from statsmodels.graphics.tsaplots import plot_acf, plot_pacf
from prophet.plot import plot_plotly,plot_components_plotly
#import seaborn as sns
from sklearn.model_selection import ParameterGrid
import holidays
#from prophet.diagnostics import performance_metrics
#from prophet.diagnostics import cross_validation
from sklearn.metrics import mean_squared_error, r2_score, mean_absolute_error
import logging,sys
from scipy.special import inv_boxcox
from prophet.diagnostics import cross_validation
#from sklearn.metrics import mean_absolute_percentage_error
warnings.filterwarnings("ignore")
# Aion Prophet module
class aion_fbprophet ():
#Constructor
def __init__(self,configfile,testpercentage,data,targetFeature,dateTimeFeature):
try:
self.tsprophet_params = configfile
self.data=data
self.targetFeature=targetFeature
self.dateTimeFeature=dateTimeFeature
self.testpercentage = testpercentage
self.log = logging.getLogger('eion')
except Exception as inst:
self.log.info('<!------------- Prophet INIT Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
#Find datetime column
def get_datetime_col(self,data):
df=data
dt_col=[]
categorical_features=[]
discrete_features=[]
# Here, I am checking each column type, whether it is object type or float or int. Then I am trying to convert the
# Object type to datetime format using python pd.to_datetime() function. If the column converts , it is datetime format, else it is some other format (categorical or discrete)
for col in df.columns:
if (df[col].dtype == 'object' or df[col].dtype == 'datetime64[ns]' ):
try:
df[col] = pd.to_datetime(df[col])
dt_col.append(col)
except ValueError:
categorical_features.append(col)
pass
elif (df[col].dtype == 'float64' or 'int64' or 'int' or 'float64' or 'float'): #('int' or 'float' or 'int64' or 'float64')):
#print("discrete features found..\n")
discrete_features.append(col)
else:
pass
#Uncomment to know the datetime, categorical and continuous cols
# print ("Date time colms: dt_col: \n",dt_col)
# print("categorical features: \n",categorical_features)
# print("continuous features: \n",discrete_features)
return dt_col
def get_predict_frequency(self,df,datetime_col_name):
#dt_col=pd.to_datetime(df[datetime_col_name], format='%m/%d/%Y %H:%M:%S')
dt_col=pd.to_datetime(df[datetime_col_name])
#df['tvalue'] = df[datetime_col_name]
df['time_diff'] = (df[datetime_col_name]-df[datetime_col_name].shift()).fillna(pd.Timedelta('0'))
mean_diff_dt=df['time_diff'].mean()
time_diff_secs=mean_diff_dt.total_seconds()
time_sec_2_hr=((time_diff_secs/60)/60)
pred_freq=""
time_sec_2_hr=round(time_sec_2_hr)
#For abbreviation ,refer https://pandas.pydata.org/pandas-docs/stable/user_guide/timeseries.html#timeseries-offset-aliases
if (time_sec_2_hr < 1):
pred_freq="min"
else:
if (time_sec_2_hr >= 24):
if (time_sec_2_hr > 168):
if(time_sec_2_hr > 696 or time_sec_2_hr < 744): # based on 29 days, to 31 days
if(time_sec_2_hr > 8760):
pred_freq="Y"
else:
pred_freq="M"
else:
pred_freq="W"
else:
pred_freq="D"
else:
pred_freq="H"
pass
return pred_freq
#To extract dict key,values
def extract_params(self,dict):
self.dict=dict
for k,v in self.dict.items():
return k,v
def mean_absolute_percentage_error(self,y_true, y_pred):
if (y_true.isin([0]).sum() > 0):
y_true=y_true.mask(y_true==0).fillna(y_true.mean())
try:
y_true, y_pred=np.array(y_true), np.array(y_pred)
#return np.mean(np.abs((y_true - y_pred) / y_true+sys.float_info.epsilon)) * 100
return np.mean(np.abs((y_true - y_pred) / y_true)) * 100
except Exception as inst:
self.log.info('<------------- mean_absolute_percentage_error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def regressor_list(self,regressorstr):
lst = regressorstr.split (",")
reg_list=[]
for i in lst:
reg_list.append(i)
#print(reg_list)
return reg_list
# def get_regressors(self,reg):
# print("get extra inputs for prophet...\n")
def aion_probhet(self,train_data,datetime_col_name,predicted_data_file,dataFolderLocation):
from prophet import Prophet
#Getting prophet params
#key,val = self.extract_params(self.tsprophet_params)
val=self.tsprophet_params
self.log.info('-------> The given prophet algorithm parameters:>>')
self.log.info(" "+str(val))
changepoint_prior_scale=[]
changepoint_range=[]
mcmc_samples=[]
interval_width=[]
holidays_prior_scale=[]
n_changepoints=[]
uncertainty_samples=[]
seasonality_prior_scale=[]
seasonality_mode=""
yearly_seasonality=None
weekly_seasonality=None
daily_seasonality=None
additional_regressors=""
holiday_country_name=""
holiday_years=[]
no_of_periods=0
pred_frequncy=""
for k,v in val.items():
try:
if (k == "seasonality_mode"):
seasonality_mode=v
elif (k == "changepoint_prior_scale"):
changepoint_prior_scale=[float(i) for i in v.split(',')]
elif (k == "changepoint_range"):
changepoint_range=[float(i) for i in v.split(',')]
elif (k == "yearly_seasonality"):
if v.lower() == 'true':
yearly_seasonality=True
elif v.lower() == 'false':
yearly_seasonality=False
elif v.lower() == 'auto':
yearly_seasonality=v
else:
yearly_seasonality=True
elif (k == "weekly_seasonality"):
if v.lower() == 'true':
weekly_seasonality=True
elif v.lower() == 'false':
weekly_seasonality=False
elif v.lower() == 'auto':
weekly_seasonality=v
else:
weekly_seasonality=False
#weekly_seasonality=v
elif (k == "daily_seasonality"):
if v.lower() == 'true':
daily_seasonality=True
elif v.lower() == 'false':
daily_seasonality=False
elif v.lower() == 'auto':
daily_seasonality=v
else:
daily_seasonality=False
elif (k == "mcmc_samples"):
mcmc_samples=[float(i) for i in v.split(',')]
elif (k == "interval_width"):
interval_width=[float(i) for i in v.split(',')]
elif (k == "holidays_prior_scale"):
#holidays_prior_scale=float(v)
holidays_prior_scale=[float(i) for i in v.split(',')]
elif (k == "n_changepoints"):
n_changepoints=[int(i) for i in v.split(',')]
elif (k == "uncertainty_samples"):
uncertainty_samples=[float(i) for i in v.split(',')]
elif (k == "seasonality_prior_scale"):
seasonality_prior_scale=[float(i) for i in v.split(',')]
elif (k == "additional_regressors"):
additional_regressors=str(v)
elif (k == "holiday_country_name"):
holiday_country_name=v
elif (k == "holiday_years"):
holiday_years=[int(i) for i in v.split(',')]
elif (k == "no_of_periods"):
no_of_periods=int(v)
elif (k == "pred_frequncy"):
pred_frequncy=v
else:
self.log.info("Invalid string.")
except Exception:
continue
try:
start = time.time()
datetime_col_name=str(datetime_col_name)
target_col=str(self.targetFeature)
#extra_regressors=additional_regressors
reg_list=self.regressor_list(additional_regressors)
get_dtcol=""
get_dtcol=self.get_datetime_col(self.data)[0]
#get predict frequency for user data
pred_freq= str(self.get_predict_frequency(self.data,datetime_col_name))
if (pred_frequncy):
pred_frequncy=pred_frequncy
else:
#If user not defined predict_freq in aion config or GUI, our algorithm will find automatically by get_predict_frequency() method
pred_frequncy=pred_freq
self.log.info("Auto Predict frequency period (Hour-H/Day-D/Week-W/Month-M/Year-Y): \n"+str(pred_frequncy))
#For proper datetime format check.
self.data[self.dateTimeFeature] = pd.to_datetime(self.data[self.dateTimeFeature])
filterd_df = self.data.filter([get_dtcol,target_col])
holiday = pd.DataFrame([])
holiday_specified=holidays.CountryHoliday(holiday_country_name,years=holiday_years)
for date, name in sorted(holiday_specified.items()):
holiday = holiday.append(pd.DataFrame({'ds': date, 'holiday': "Holidays"}, index=[0]), ignore_index=True)
holiday['ds'] = pd.to_datetime(holiday['ds'], format='%Y-%m-%d %H:%M:%S', errors='ignore')
filterd_df=filterd_df.rename(columns={self.dateTimeFeature:'ds',target_col:'y'})
#Set seasonality model
try:
if not seasonality_mode:
self.log.info('empty input for seasonality_mode parameter in aion configuration file.Please check. Setting default mode: additive. \n')
seasonality_mode=[]
seasonality_mode=['additive']
multiplicative_s="multiplicative"
additive_s="additive"
else:
seasonality_mode = seasonality_mode.split(',')
len_seasonality_mode=len(seasonality_mode)
except ValueError as e:
self.log.info(e)
params_grid = {'seasonality_mode':(seasonality_mode),
'changepoint_prior_scale':changepoint_prior_scale,
'changepoint_range': changepoint_range,
'yearly_seasonality': [yearly_seasonality],
'weekly_seasonality': [weekly_seasonality],
'daily_seasonality': [daily_seasonality],
'mcmc_samples': mcmc_samples,
'interval_width': interval_width,
'holidays_prior_scale':holidays_prior_scale,
'n_changepoints' : n_changepoints,
'uncertainty_samples': uncertainty_samples,
'seasonality_prior_scale': seasonality_prior_scale}
grid = ParameterGrid(params_grid)
p_cnt = 0
for p in grid:
p_cnt = p_cnt+1
self.log.info("--------------- Total Possible prophet iterations: --------------- \n")
self.log.info(p_cnt)
self.log.info("\n--------------- Modal Validation Start ---------------")
size = int(len(filterd_df) * (100 - self.testpercentage)/100)
train = filterd_df.loc[0:size]
valid = filterd_df.loc[size:len(filterd_df)]
self.log.info("------->Train Data Shape: "+str(train.shape))
self.log.info("------->Valid Data Shape"+str(valid.shape))
X_train = train
X_test = valid
len_test=len(X_test)
#For add_regressor,copy the add_regressor columns to use.
if (additional_regressors):
df1=pd.DataFrame()
df1[additional_regressors]=self.data[additional_regressors]
model_parameters_mape = pd.DataFrame(columns = ['MAPE','Parameters'])
model_parameters_rmse = pd.DataFrame(columns = ['rmse','Parameters'])
model_parameters_mse = pd.DataFrame(columns = ['mse','Parameters'])
model_parameters_mae = pd.DataFrame(columns = ['MAE','Parameters'])
model_parameters_r2 = pd.DataFrame(columns = ['r2','Parameters'])
for P in grid:
pred_forecast = pd.DataFrame()
random.seed(0)
train_model =Prophet(changepoint_prior_scale = P['changepoint_prior_scale'],
seasonality_mode=P['seasonality_mode'],
changepoint_range=P['changepoint_range'],
holidays_prior_scale = P['holidays_prior_scale'],
n_changepoints = P['n_changepoints'],
mcmc_samples=P['mcmc_samples'],
interval_width=P['interval_width'],
uncertainty_samples=P['uncertainty_samples'],
seasonality_prior_scale= P['seasonality_prior_scale'],
holidays=holiday,
weekly_seasonality=P['weekly_seasonality'],
daily_seasonality = P['daily_seasonality'],
yearly_seasonality = P['yearly_seasonality']
)
train_forecast=pd.DataFrame()
try:
train_model.fit(X_train)
train_forecast = train_model.make_future_dataframe(periods=len_test, freq=pred_frequncy,include_history = False)
train_forecast = train_model.predict(train_forecast)
except ValueError as e:
self.log.info(e)
self.log.info ("------->Check mcmc_samples value in aion confiuration, either 0 (default) or defined value,e.g.mcmc_samples:'300' to be set.If no idea on value, set to default.\n")
pred_forecast=train_forecast[['ds','yhat']]
Actual=X_test
len_act=len(Actual['y'])
len_pred=len(pred_forecast['yhat'])
MAPE = self.mean_absolute_percentage_error(Actual['y'],abs(pred_forecast['yhat']))
model_parameters_mape = model_parameters_mape.append({'MAPE':MAPE,'Parameters':p},ignore_index=True)
#MAE
MAE = mean_absolute_error(Actual['y'],abs(pred_forecast['yhat']))
rmse = sqrt(mean_squared_error(Actual['y'],abs(pred_forecast['yhat'])))
mse = mean_squared_error(Actual['y'],abs(pred_forecast['yhat']))
r2 = r2_score(Actual['y'],abs(pred_forecast['yhat']))
# self.log.info ("------->Prophet RMSE :"+str(rmse))
# self.log.info ("------->Prophet MSE :"+str(mse))
# self.log.info ("------->Prophet MAE :"+str(MAE))
# self.log.info ("------->Prophet R2 :"+str(r2))
model_parameters_mape = model_parameters_mape.append({'MAPE':MAPE,'Parameters':p},ignore_index=True)
model_parameters_rmse = model_parameters_rmse.append({'rmse':rmse,'Parameters':p},ignore_index=True)
model_parameters_mse = model_parameters_mse.append({'mse':mse,'Parameters':p},ignore_index=True)
model_parameters_mae = model_parameters_mae.append({'MAE':MAE,'Parameters':p},ignore_index=True)
model_parameters_r2 = model_parameters_r2.append({'r2':r2,'Parameters':p},ignore_index=True)
#end of for loop
parameters_mape = model_parameters_mape.sort_values(by=['MAPE'])
parameters_mape = parameters_mape.reset_index(drop=True)
best_params_mape=parameters_mape['Parameters'][0]
# print("Best Parameters on which the model has the least MAPE is: \n",best_params_mape)
best_mape_score=parameters_mape['MAPE'].iloc[0]
#self.log.info('------->Mean absolute percent error log: \n ')
#self.log.info('------->best_mape_score: \n '+str(best_mape_score))
parameters_rmse = model_parameters_rmse.sort_values(by=['rmse'])
parameters_rmse = parameters_rmse.reset_index(drop=True)
best_params_rmse=parameters_rmse['Parameters'][0]
best_rmse_score=parameters_rmse['rmse'].iloc[0]
#self.log.info('------->Root Man Squared Error log (Prophet timeseries): \n ')
#self.log.info('------->best_rmse_score ((Prophet timeseries)): \n '+str(best_rmse_score))
#mse
parameters_mse = model_parameters_mse.sort_values(by=['mse'])
parameters_mse = parameters_mse.reset_index(drop=True)
best_params_mse = parameters_mse['Parameters'][0]
best_mse_score=parameters_mse['mse'].iloc[0]
#MAE
parameters_mae = model_parameters_mae.sort_values(by=['MAE'])
parameters_mae = parameters_mae.reset_index(drop=True)
best_params_mae = parameters_mae['Parameters'][0]
best_mae_score=parameters_mae['MAE'].iloc[0]
# R2 score
parameters_r2 = model_parameters_r2.sort_values(by=['r2'])
parameters_r2 = parameters_r2.reset_index(drop=False)
best_params_r2 = parameters_r2['Parameters'][0]
best_r2_score=parameters_r2['r2'].iloc[0]
#Final best prophet mse,rmse,mape scores
# self.log.info ("------->Prophet RMSE :"+str(best_rmse_score))
# self.log.info ("------->Prophet MSE :"+str(best_mse_score))
# self.log.info ("------->Prophet MAE :"+str(best_mae_score))
# self.log.info ("------->Prophet R2 :"+str(best_r2_score))
#Extracting best model parameters
for k,v in best_params_mape.items():
try:
if (k == "changepoint_prior_scale"):
changepoint_prior_scale=float(v)
elif (k == "changepoint_range"):
changepoint_range=float(v)
elif (k == "daily_seasonality"):
daily_seasonality=v
elif (k == "holidays_prior_scale"):
holidays_prior_scale=float(v)
elif (k == "interval_width"):
interval_width=float(v)
elif (k == "mcmc_samples"):
mcmc_samples=float(v)
elif (k == "n_changepoints"):
n_changepoints=int(v)
elif (k == "seasonality_mode"):
seasonality_mode=str(v)
elif (k == "seasonality_prior_scale"):
seasonality_prior_scale=int(v)
elif (k == "uncertainty_samples"):
uncertainty_samples=float(v)
elif (k == "weekly_seasonality"):
weekly_seasonality=v
elif (k == "yearly_seasonality"):
yearly_seasonality=v
else:
pass
except Exception as e:
self.log.info("\n prophet time series config param parsing error"+str(e))
#continue
self.log.info("\n Best prophet model accuracy parameters.\n ")
#Prophet model based on mape best params.
best_prophet_model = Prophet(holidays=holiday,
changepoint_prior_scale= changepoint_prior_scale,
holidays_prior_scale = holidays_prior_scale,
n_changepoints = n_changepoints,
seasonality_mode = seasonality_mode,
weekly_seasonality= weekly_seasonality,
daily_seasonality = daily_seasonality,
yearly_seasonality = yearly_seasonality,
interval_width=interval_width,
mcmc_samples=mcmc_samples,
changepoint_range=changepoint_range)
# If holiday not set using prophet model,we can add as below.
# best_prophet_model.add_country_holidays(country_name=holiday_country_name)
#prophet add_regressor ,adding additional influencer (regressor) features, but it different from multivariant model.
if (additional_regressors):
filterd_df[additional_regressors] = df1[additional_regressors]
filterd_df.reset_index(drop=True)
for v in reg_list:
best_prophet_model=best_prophet_model.add_regressor(v)
#best_prophet_model.fit(X_train)
else:
pass
#Model prophet fit, it should be done before make_future_dataframe
best_prophet_model.fit(filterd_df)
future = best_prophet_model.make_future_dataframe(periods=no_of_periods, freq=pred_frequncy,include_history = False)
if (additional_regressors):
future[additional_regressors] = filterd_df[additional_regressors]
future.reset_index(drop=True)
future=future.dropna()
else:
pass
#Final prediction
forecast = best_prophet_model.predict(future)
# forecast_df=forecast[['ds', 'yhat', 'yhat_lower', 'yhat_upper']]
# #Save forecast as csv file
# forecast_df.to_csv(r"prophet_realtime_user_steps.csv",index = False, header=True)
#Plot the predition and save in file
forecast_plot = best_prophet_model.plot(forecast)
imagefilename = os.path.join(dataFolderLocation,'log','img','prophet_fig.png')
forecast_plot.savefig(imagefilename)
#The below part is used to compare prophet predicted with actual value
#For train data
#Prophet model with train and test data, based on mape best params.
best_prophet_model_new = Prophet(holidays=holiday,
changepoint_prior_scale= changepoint_prior_scale,
holidays_prior_scale = holidays_prior_scale,
n_changepoints = n_changepoints,
seasonality_mode = seasonality_mode,
weekly_seasonality= weekly_seasonality,
daily_seasonality = daily_seasonality,
yearly_seasonality = yearly_seasonality,
interval_width=interval_width,
mcmc_samples=mcmc_samples,
changepoint_range=changepoint_range)
fp_forecast=pd.DataFrame()
try:
best_prophet_model_new.fit(X_train)
fp_forecast = best_prophet_model_new.make_future_dataframe(periods=len_test, freq=pred_frequncy,include_history = False)
fp_forecast = best_prophet_model_new.predict(fp_forecast)
except ValueError as e:
self.log.info(e)
self.log.info ("------->Check mcmc_samples value in aion confiuration, either 0 (default) or defined value,e.g.mcmc_samples:'300' to be set.If no idea on value, set to default.\n")
pred_forecast=fp_forecast[['ds','yhat']]
pred_forecast['ds']=Actual['ds'].to_numpy()
Actual.ds = pd.to_datetime(Actual.ds)
pred_forecast.ds = pd.to_datetime(pred_forecast.ds)
MAE = mean_absolute_error(Actual['y'],abs(pred_forecast['yhat']))
rmse = sqrt(mean_squared_error(Actual['y'],abs(pred_forecast['yhat'])))
mse = mean_squared_error(Actual['y'],abs(pred_forecast['yhat']))
r2 = r2_score(Actual['y'],abs(pred_forecast['yhat']))
MAPE = self.mean_absolute_percentage_error(Actual['y'],abs(pred_forecast['yhat']))
#Final best prophet mse,rmse,mape scores
self.log.info ("------->Prophet RMSE : "+str(rmse))
self.log.info ("------->Prophet MSE : "+str(mse))
self.log.info ("------->Prophet MAE : "+str(MAE))
self.log.info ("------->Prophet R2 : "+str(r2))
self.log.info("------->Prophet MAPE: "+str(MAPE))
#self.log.info(MAPE)
#self.log.info('------->best_mape_score: \n '+str(best_mape_score))
prophet_df = pd.merge(Actual,pred_forecast, on=['ds'], how='left')
cols = ['ds','y','yhat']
prophet_df_new = prophet_df[cols]
prophet_df_new.dropna(inplace=True)
actualfeature = target_col+'_actual'
predictfeature = target_col+'_pred'
prophet_df_new=prophet_df_new.rename(columns={'ds': 'datetime', 'y': actualfeature,'yhat': predictfeature})
#prophet_df_new.to_csv(predicted_data_file)
#cv_results = cross_validation( model = best_prophet_model, initial = pd.to_timedelta(no_of_periods,unit=pred_frequncy), horizon = pd.to_timedelta(no_of_periods,unit=pred_frequncy))
#forecast_df=forecast[['ds', 'yhat', 'yhat_lower', 'yhat_upper']]
#Save forecast as csv file
#forecast_df.to_csv(r"prophet_realtime_Output.csv",index = False, header=True)
# self.log.info('------->Prophet time series forecast (last 7 prediction for user view): \n ')
# self.log.info(forecast[['ds', 'yhat', 'yhat_lower', 'yhat_upper']].tail(7))
plot_prd=plot_plotly(best_prophet_model, forecast)
imagefilename = os.path.join(dataFolderLocation,'log','img','1_ppm_plot')
plotly.offline.plot(plot_prd, filename=imagefilename,auto_open = False)
plot_prd_components=plot_components_plotly(best_prophet_model, forecast)
imagefilename = os.path.join(dataFolderLocation,'log','img','2_ppm_plot')
plotly.offline.plot(plot_prd_components, filename=imagefilename,auto_open = False)
executionTime=(time.time() - start)
self.log.info('-------> Time: '+str(executionTime))
return best_prophet_model,best_mae_score,best_rmse_score,best_mse_score,best_mape_score,best_r2_score,pred_frequncy,additional_regressors,prophet_df_new
except Exception as inst:
#print("********** aion_fbprophet exception ************* \n")
self.log.info('<!------------- Prophet Execute Error ---------------> '+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
|
generic_feature_statistics_generator_test.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
from facets_overview.generic_feature_statistics_generator import GenericFeatureStatisticsGenerator
import numpy as np
import pandas as pd
from tensorflow.python.platform import googletest
class GenericFeatureStatisticsGeneratorTest(googletest.TestCase):
def setUp(self):
self.gfsg = GenericFeatureStatisticsGenerator()
def testProtoFromDataFrames(self):
data = [[1, 'hi'], [2, 'hello'], [3, 'hi']]
df = pd.DataFrame(data, columns=['testFeatureInt', 'testFeatureString'])
dataframes = [{'table': df, 'name': 'testDataset'}]
p = self.gfsg.ProtoFromDataFrames(dataframes)
self.assertEqual(1, len(p.datasets))
test_data = p.datasets[0]
self.assertEqual('testDataset', test_data.name)
self.assertEqual(3, test_data.num_examples)
self.assertEqual(2, len(test_data.features))
if test_data.features[0].name == 'testFeatureInt':
numfeat = test_data.features[0]
stringfeat = test_data.features[1]
else:
numfeat = test_data.features[1]
stringfeat = test_data.features[0]
self.assertEqual('testFeatureInt', numfeat.name)
self.assertEqual(self.gfsg.fs_proto.INT, numfeat.type)
self.assertEqual(1, numfeat.num_stats.min)
self.assertEqual(3, numfeat.num_stats.max)
self.assertEqual('testFeatureString', stringfeat.name)
self.assertEqual(self.gfsg.fs_proto.STRING, stringfeat.type)
self.assertEqual(2, stringfeat.string_stats.unique)
def testNdarrayToEntry(self):
arr = np.array([1.0, 2.0, None, float('nan'), 3.0], dtype=float)
entry = self.gfsg.NdarrayToEntry(arr)
self.assertEqual(2, entry['missing'])
arr = np.array(['a', 'b', float('nan'), 'c'], dtype=str)
entry = self.gfsg.NdarrayToEntry(arr)
self.assertEqual(1, entry['missing'])
def testNdarrayToEntryTimeTypes(self):
arr = np.array(
[np.datetime64('2005-02-25'),
np.datetime64('2006-02-25')],
dtype=np.datetime64)
entry = self.gfsg.NdarrayToEntry(arr)
self.assertEqual([1109289600000000000, 1140825600000000000], entry['vals'])
arr = np.array(
[np.datetime64('2009-01-01') - np.datetime64('2008-01-01')],
dtype=np.timedelta64)
entry = self.gfsg.NdarrayToEntry(arr)
self.assertEqual([31622400000000000], entry['vals'])
def testDTypeToType(self):
self.assertEqual(self.gfsg.fs_proto.INT,
self.gfsg.DtypeToType(np.dtype(np.int32)))
# Boolean and time types treated as int
self.assertEqual(self.gfsg.fs_proto.INT,
self.gfsg.DtypeToType(np.dtype(np.bool)))
self.assertEqual(self.gfsg.fs_proto.INT,
self.gfsg.DtypeToType(np.dtype(np.datetime64)))
self.assertEqual(self.gfsg.fs_proto.INT,
self.gfsg.DtypeToType(np.dtype(np.timedelta64)))
self.assertEqual(self.gfsg.fs_proto.FLOAT,
self.gfsg.DtypeToType(np.dtype(np.float32)))
self.assertEqual(self.gfsg.fs_proto.STRING,
self.gfsg.DtypeToType(np.dtype(np.str)))
# Unsupported types treated as string for now
self.assertEqual(self.gfsg.fs_proto.STRING,
self.gfsg.DtypeToType(np.dtype(np.void)))
def testGetDatasetsProtoFromEntriesLists(self):
entries = {}
entries['testFeature'] = {
'vals': [1, 2, 3],
'counts': [1, 1, 1],
'missing': 0,
'type': self.gfsg.fs_proto.INT
}
datasets = [{'entries': entries, 'size': 3, 'name': 'testDataset'}]
p = self.gfsg.GetDatasetsProto(datasets)
self.assertEqual(1, len(p.datasets))
test_data = p.datasets[0]
self.assertEqual('testDataset', test_data.name)
self.assertEqual(3, test_data.num_examples)
self.assertEqual(1, len(test_data.features))
numfeat = test_data.features[0]
self.assertEqual('testFeature', numfeat.name)
self.assertEqual(self.gfsg.fs_proto.INT, numfeat.type)
self.assertEqual(1, numfeat.num_stats.min)
self.assertEqual(3, numfeat.num_stats.max)
hist = numfeat.num_stats.common_stats.num_values_histogram
buckets = hist.buckets
self.assertEqual(self.gfsg.histogram_proto.QUANTILES, hist.type)
self.assertEqual(10, len(buckets))
self.assertEqual(1, buckets[0].low_value)
self.assertEqual(1, buckets[0].high_value)
self.assertEqual(.3, buckets[0].sample_count)
self.assertEqual(1, buckets[9].low_value)
self.assertEqual(1, buckets[9].high_value)
self.assertEqual(.3, buckets[9].sample_count)
def testGetDatasetsProtoSequenceExampleHistogram(self):
entries = {}
entries['testFeature'] = {
'vals': [1, 2, 2, 3],
'counts': [1, 2, 1],
'feat_lens': [1, 2, 1],
'missing': 0,
'type': self.gfsg.fs_proto.INT
}
datasets = [{'entries': entries, 'size': 3, 'name': 'testDataset'}]
p = self.gfsg.GetDatasetsProto(datasets)
hist = p.datasets[0].features[
0].num_stats.common_stats.feature_list_length_histogram
buckets = hist.buckets
self.assertEqual(self.gfsg.histogram_proto.QUANTILES, hist.type)
self.assertEqual(10, len(buckets))
self.assertEqual(1, buckets[0].low_value)
self.assertEqual(1, buckets[0].high_value)
self.assertEqual(.3, buckets[0].sample_count)
self.assertEqual(1.8, buckets[9].low_value)
self.assertEqual(2, buckets[9].high_value)
self.assertEqual(.3, buckets[9].sample_count)
def testGetDatasetsProtoWithWhitelist(self):
entries = {}
entries['testFeature'] = {
'vals': [1, 2, 3],
'counts': [1, 1, 1],
'missing': 0,
'type': self.gfsg.fs_proto.INT
}
entries['ignoreFeature'] = {
'vals': [5, 6],
'counts': [1, 1],
'missing': 1,
'type': self.gfsg.fs_proto.INT
}
datasets = [{'entries': entries, 'size': 3, 'name': 'testDataset'}]
p = self.gfsg.GetDatasetsProto(datasets, features=['testFeature'])
self.assertEqual(1, len(p.datasets))
test_data = p.datasets[0]
self.assertEqual('testDataset', test_data.name)
self.assertEqual(3, test_data.num_examples)
self.assertEqual(1, len(test_data.features))
numfeat = test_data.features[0]
self.assertEqual('testFeature', numfeat.name)
self.assertEqual(1, numfeat.num_stats.min)
def testGetDatasetsProtoWithMaxHistigramLevelsCount(self):
# Selected entries' lengths make it easy to compute average length
data = [['hi'], ['good'], ['hi'], ['hi'], ['a'], ['a']]
df = pd.DataFrame(data, columns=['testFeatureString'])
dataframes = [{'table': df, 'name': 'testDataset'}]
# Getting proto from ProtoFromDataFrames instead of GetDatasetsProto
# directly to avoid any hand written values ex: size of dataset.
p = self.gfsg.ProtoFromDataFrames(dataframes,
histogram_categorical_levels_count=2)
self.assertEqual(1, len(p.datasets))
test_data = p.datasets[0]
self.assertEqual('testDataset', test_data.name)
self.assertEqual(6, test_data.num_examples)
self.assertEqual(1, len(test_data.features))
numfeat = test_data.features[0]
self.assertEqual('testFeatureString', numfeat.name)
top_values = numfeat.string_stats.top_values
self.assertEqual(3, top_values[0].frequency)
self.assertEqual('hi', top_values[0].value)
self.assertEqual(3, numfeat.string_stats.unique)
self.assertEqual(2, numfeat.string_stats.avg_length)
rank_hist = numfeat.string_stats.rank_histogram
buckets = rank_hist.buckets
self.assertEqual(2, len(buckets))
self.assertEqual('hi', buckets[0].label)
self.assertEqual(3, buckets[0].sample_count)
self.assertEqual('a', buckets[1].label)
self.assertEqual(2, buckets[1].sample_count)
if __name__ == '__main__':
googletest.main()
|
generic_feature_statistics_generator.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Code for generating the feature_statistics proto from generic data.
The proto is used as input for the Overview visualization.
"""
from facets_overview.base_generic_feature_statistics_generator import BaseGenericFeatureStatisticsGenerator
import facets_overview.feature_statistics_pb2 as fs
class GenericFeatureStatisticsGenerator(BaseGenericFeatureStatisticsGenerator):
"""Generator of stats proto from generic data."""
def __init__(self):
BaseGenericFeatureStatisticsGenerator.__init__(
self, fs.FeatureNameStatistics, fs.DatasetFeatureStatisticsList,
fs.Histogram)
|
__init__.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
|
feature_statistics_generator_test.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
from facets_overview.feature_statistics_generator import FeatureStatisticsGenerator
import numpy as np
import tensorflow as tf
from tensorflow.python.platform import googletest
class FeatureStatisticsGeneratorTest(googletest.TestCase):
def setUp(self):
self.fs = FeatureStatisticsGenerator()
def testParseExampleInt(self):
# Tests parsing examples of integers
examples = []
for i in range(50):
example = tf.train.Example()
example.features.feature['num'].int64_list.value.append(i)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
self.assertEqual(1, len(entries))
self.assertIn('num', entries)
info = entries['num']
self.assertEqual(0, info['missing'])
self.assertEqual(self.fs.fs_proto.INT, info['type'])
for i in range(len(examples)):
self.assertEqual(1, info['counts'][i])
self.assertEqual(i, info['vals'][i])
def testParseExampleMissingValueList(self):
# Tests parsing examples of integers
examples = []
example = tf.train.Example()
# pylint: disable=pointless-statement
example.features.feature['str']
# pylint: enable=pointless-statement
examples.append(example)
example = tf.train.Example()
example.features.feature['str'].bytes_list.value.append(b'test')
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
self.assertEqual(1, len(entries))
self.assertIn('str', entries)
info = entries['str']
self.assertEqual(1, info['missing'])
self.assertEqual(self.fs.fs_proto.STRING, info['type'])
self.assertEqual(0, info['counts'][0])
self.assertEqual(1, info['counts'][1])
def _check_sequence_example_entries(self,
entries,
n_examples,
n_features,
feat_len=None):
self.assertIn('num', entries)
info = entries['num']
self.assertEqual(0, info['missing'])
self.assertEqual(self.fs.fs_proto.INT, info['type'])
for i in range(n_examples):
self.assertEqual(n_features, info['counts'][i])
if feat_len is not None:
self.assertEqual(feat_len, info['feat_lens'][i])
for i in range(n_examples * n_features):
self.assertEqual(i, info['vals'][i])
if feat_len is None:
self.assertEqual(0, len(info['feat_lens']))
def testParseExampleSequenceContext(self):
# Tests parsing examples of integers in context field
examples = []
for i in range(50):
example = tf.train.SequenceExample()
example.context.feature['num'].int64_list.value.append(i)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.context.feature,
example.feature_lists.feature_list, entries, i)
self._check_sequence_example_entries(entries, 50, 1)
self.assertEqual(1, len(entries))
def testParseExampleSequenceFeatureList(self):
examples = []
for i in range(50):
example = tf.train.SequenceExample()
feat = example.feature_lists.feature_list['num'].feature.add()
feat.int64_list.value.append(i)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.context.feature,
example.feature_lists.feature_list, entries, i)
self._check_sequence_example_entries(entries, 50, 1, 1)
def testParseExampleSequenceFeatureListMultipleEntriesInner(self):
examples = []
for i in range(2):
example = tf.train.SequenceExample()
feat = example.feature_lists.feature_list['num'].feature.add()
for j in range(25):
feat.int64_list.value.append(i * 25 + j)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.context.feature,
example.feature_lists.feature_list, entries, i)
self._check_sequence_example_entries(entries, 2, 25, 1)
def testParseExampleSequenceFeatureListMultipleEntriesOuter(self):
# Tests parsing examples of integers in context field
examples = []
for i in range(2):
example = tf.train.SequenceExample()
for j in range(25):
feat = example.feature_lists.feature_list['num'].feature.add()
feat.int64_list.value.append(i * 25 + j)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.context.feature,
example.feature_lists.feature_list, entries, i)
self._check_sequence_example_entries(entries, 2, 25, 25)
def testVaryingCountsAndMissing(self):
# Tests parsing examples of when some examples have missing features
examples = []
for i in range(5):
example = tf.train.Example()
example.features.feature['other'].int64_list.value.append(0)
for _ in range(i):
example.features.feature['num'].int64_list.value.append(i)
examples.append(example)
example = tf.train.Example()
example.features.feature['other'].int64_list.value.append(0)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
info = entries['num']
self.assertEqual(2, info['missing'])
self.assertEqual(4, len(info['counts']))
for i in range(4):
self.assertEqual(i + 1, info['counts'][i])
self.assertEqual(10, len(info['vals']))
def testParseExampleStringsAndFloats(self):
# Tests parsing examples of string and float features
examples = []
for i in range(50):
example = tf.train.Example()
example.features.feature['str'].bytes_list.value.append(b'hi')
example.features.feature['float'].float_list.value.append(i)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
self.assertEqual(2, len(entries))
self.assertEqual(self.fs.fs_proto.FLOAT, entries['float']['type'])
self.assertEqual(self.fs.fs_proto.STRING, entries['str']['type'])
for i in range(len(examples)):
self.assertEqual(1, entries['str']['counts'][i])
self.assertEqual(1, entries['float']['counts'][i])
self.assertEqual(i, entries['float']['vals'][i])
self.assertEqual('hi', entries['str']['vals'][i].decode(
'UTF-8', 'strict'))
def testParseExamplesTypeMismatch(self):
examples = []
example = tf.train.Example()
example.features.feature['feat'].int64_list.value.append(0)
examples.append(example)
example = tf.train.Example()
example.features.feature['feat'].bytes_list.value.append(b'str')
examples.append(example)
entries = {}
self.fs._ParseExample(examples[0].features.feature, [], entries, 0)
with self.assertRaises(TypeError):
self.fs._ParseExample(examples[1].features.feature, [], entries, 1)
def testGetDatasetsProtoFromEntriesLists(self):
entries = {}
entries['testFeature'] = {
'vals': [1, 2, 3],
'counts': [1, 1, 1],
'missing': 0,
'type': self.fs.fs_proto.INT
}
datasets = [{'entries': entries, 'size': 3, 'name': 'testDataset'}]
p = self.fs.GetDatasetsProto(datasets)
self.assertEqual(1, len(p.datasets))
test_data = p.datasets[0]
self.assertEqual('testDataset', test_data.name)
self.assertEqual(3, test_data.num_examples)
self.assertEqual(1, len(test_data.features))
numfeat = test_data.features[0]
self.assertEqual('testFeature', numfeat.name)
self.assertEqual(self.fs.fs_proto.INT, numfeat.type)
self.assertEqual(1, numfeat.num_stats.min)
self.assertEqual(3, numfeat.num_stats.max)
def testGetProtoNums(self):
# Tests converting int examples into the feature stats proto
examples = []
for i in range(50):
example = tf.train.Example()
example.features.feature['num'].int64_list.value.append(i)
examples.append(example)
example = tf.train.Example()
example.features.feature['other'].int64_list.value.append(0)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
datasets = [{'entries': entries, 'size': len(examples), 'name': 'test'}]
p = self.fs.GetDatasetsProto(datasets)
self.assertEqual(1, len(p.datasets))
test_data = p.datasets[0]
self.assertEqual('test', test_data.name)
self.assertEqual(51, test_data.num_examples)
numfeat = test_data.features[0] if (
test_data.features[0].name == 'num') else test_data.features[1]
self.assertEqual('num', numfeat.name)
self.assertEqual(self.fs.fs_proto.INT, numfeat.type)
self.assertEqual(0, numfeat.num_stats.min)
self.assertEqual(49, numfeat.num_stats.max)
self.assertEqual(24.5, numfeat.num_stats.mean)
self.assertEqual(24.5, numfeat.num_stats.median)
self.assertEqual(1, numfeat.num_stats.num_zeros)
self.assertAlmostEqual(14.430869689, numfeat.num_stats.std_dev, 4)
self.assertEqual(1, numfeat.num_stats.common_stats.num_missing)
self.assertEqual(50, numfeat.num_stats.common_stats.num_non_missing)
self.assertEqual(1, numfeat.num_stats.common_stats.min_num_values)
self.assertEqual(1, numfeat.num_stats.common_stats.max_num_values)
self.assertAlmostEqual(1, numfeat.num_stats.common_stats.avg_num_values, 4)
hist = numfeat.num_stats.common_stats.num_values_histogram
buckets = hist.buckets
self.assertEqual(self.fs.histogram_proto.QUANTILES, hist.type)
self.assertEqual(10, len(buckets))
self.assertEqual(1, buckets[0].low_value)
self.assertEqual(1, buckets[0].high_value)
self.assertEqual(5, buckets[0].sample_count)
self.assertEqual(1, buckets[9].low_value)
self.assertEqual(1, buckets[9].high_value)
self.assertEqual(5, buckets[9].sample_count)
self.assertEqual(2, len(numfeat.num_stats.histograms))
buckets = numfeat.num_stats.histograms[0].buckets
self.assertEqual(self.fs.histogram_proto.STANDARD,
numfeat.num_stats.histograms[0].type)
self.assertEqual(10, len(buckets))
self.assertEqual(0, buckets[0].low_value)
self.assertEqual(4.9, buckets[0].high_value)
self.assertEqual(5, buckets[0].sample_count)
self.assertAlmostEqual(44.1, buckets[9].low_value)
self.assertEqual(49, buckets[9].high_value)
self.assertEqual(5, buckets[9].sample_count)
buckets = numfeat.num_stats.histograms[1].buckets
self.assertEqual(self.fs.histogram_proto.QUANTILES,
numfeat.num_stats.histograms[1].type)
self.assertEqual(10, len(buckets))
self.assertEqual(0, buckets[0].low_value)
self.assertEqual(4.9, buckets[0].high_value)
self.assertEqual(5, buckets[0].sample_count)
self.assertAlmostEqual(44.1, buckets[9].low_value)
self.assertEqual(49, buckets[9].high_value)
self.assertEqual(5, buckets[9].sample_count)
def testQuantiles(self):
examples = []
for i in range(50):
example = tf.train.Example()
example.features.feature['num'].int64_list.value.append(i)
examples.append(example)
for i in range(50):
example = tf.train.Example()
example.features.feature['num'].int64_list.value.append(100)
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
datasets = [{'entries': entries, 'size': len(examples), 'name': 'test'}]
p = self.fs.GetDatasetsProto(datasets)
numfeat = p.datasets[0].features[0]
self.assertEqual(2, len(numfeat.num_stats.histograms))
self.assertEqual(self.fs.histogram_proto.QUANTILES,
numfeat.num_stats.histograms[1].type)
buckets = numfeat.num_stats.histograms[1].buckets
self.assertEqual(10, len(buckets))
self.assertEqual(0, buckets[0].low_value)
self.assertEqual(9.9, buckets[0].high_value)
self.assertEqual(10, buckets[0].sample_count)
self.assertEqual(100, buckets[9].low_value)
self.assertEqual(100, buckets[9].high_value)
self.assertEqual(10, buckets[9].sample_count)
def testInfinityAndNan(self):
examples = []
for i in range(50):
example = tf.train.Example()
example.features.feature['num'].float_list.value.append(i)
examples.append(example)
example = tf.train.Example()
example.features.feature['num'].float_list.value.append(float('inf'))
examples.append(example)
example = tf.train.Example()
example.features.feature['num'].float_list.value.append(float('-inf'))
examples.append(example)
example = tf.train.Example()
example.features.feature['num'].float_list.value.append(float('nan'))
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
datasets = [{'entries': entries, 'size': len(examples), 'name': 'test'}]
p = self.fs.GetDatasetsProto(datasets)
numfeat = p.datasets[0].features[0]
self.assertEqual('num', numfeat.name)
self.assertEqual(self.fs.fs_proto.FLOAT, numfeat.type)
self.assertTrue(np.isnan(numfeat.num_stats.min))
self.assertTrue(np.isnan(numfeat.num_stats.max))
self.assertTrue(np.isnan(numfeat.num_stats.mean))
self.assertTrue(np.isnan(numfeat.num_stats.median))
self.assertEqual(1, numfeat.num_stats.num_zeros)
self.assertTrue(np.isnan(numfeat.num_stats.std_dev))
self.assertEqual(53, numfeat.num_stats.common_stats.num_non_missing)
hist = buckets = numfeat.num_stats.histograms[0]
buckets = hist.buckets
self.assertEqual(self.fs.histogram_proto.STANDARD, hist.type)
self.assertEqual(1, hist.num_nan)
self.assertEqual(10, len(buckets))
self.assertEqual(float('-inf'), buckets[0].low_value)
self.assertEqual(4.9, buckets[0].high_value)
self.assertEqual(6, buckets[0].sample_count)
self.assertEqual(44.1, buckets[9].low_value)
self.assertEqual(float('inf'), buckets[9].high_value)
self.assertEqual(6, buckets[9].sample_count)
def testInfinitysOnly(self):
examples = []
example = tf.train.Example()
example.features.feature['num'].float_list.value.append(float('inf'))
examples.append(example)
example = tf.train.Example()
example.features.feature['num'].float_list.value.append(float('-inf'))
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
datasets = [{'entries': entries, 'size': len(examples), 'name': 'test'}]
p = self.fs.GetDatasetsProto(datasets)
numfeat = p.datasets[0].features[0]
hist = buckets = numfeat.num_stats.histograms[0]
buckets = hist.buckets
self.assertEqual(self.fs.histogram_proto.STANDARD, hist.type)
self.assertEqual(10, len(buckets))
self.assertEqual(float('-inf'), buckets[0].low_value)
self.assertEqual(0.1, buckets[0].high_value)
self.assertEqual(1, buckets[0].sample_count)
self.assertEqual(0.9, buckets[9].low_value)
self.assertEqual(float('inf'), buckets[9].high_value)
self.assertEqual(1, buckets[9].sample_count)
def testGetProtoStrings(self):
# Tests converting string examples into the feature stats proto
examples = []
for i in range(2):
example = tf.train.Example()
example.features.feature['str'].bytes_list.value.append(b'hello')
examples.append(example)
for i in range(3):
example = tf.train.Example()
example.features.feature['str'].bytes_list.value.append(b'hi')
examples.append(example)
example = tf.train.Example()
example.features.feature['str'].bytes_list.value.append(b'hey')
examples.append(example)
entries = {}
for i, example in enumerate(examples):
self.fs._ParseExample(example.features.feature, [], entries, i)
datasets = [{'entries': entries, 'size': len(examples), 'name': 'test'}]
p = self.fs.GetDatasetsProto(datasets)
self.assertEqual(1, len(p.datasets))
test_data = p.datasets[0]
self.assertEqual('test', test_data.name)
self.assertEqual(6, test_data.num_examples)
strfeat = test_data.features[0]
self.assertEqual('str', strfeat.name)
self.assertEqual(self.fs.fs_proto.STRING, strfeat.type)
self.assertEqual(3, strfeat.string_stats.unique)
self.assertAlmostEqual(19 / 6.0, strfeat.string_stats.avg_length, 4)
self.assertEqual(0, strfeat.string_stats.common_stats.num_missing)
self.assertEqual(6, strfeat.string_stats.common_stats.num_non_missing)
self.assertEqual(1, strfeat.string_stats.common_stats.min_num_values)
self.assertEqual(1, strfeat.string_stats.common_stats.max_num_values)
self.assertEqual(1, strfeat.string_stats.common_stats.avg_num_values)
hist = strfeat.string_stats.common_stats.num_values_histogram
buckets = hist.buckets
self.assertEqual(self.fs.histogram_proto.QUANTILES, hist.type)
self.assertEqual(10, len(buckets))
self.assertEqual(1, buckets[0].low_value)
self.assertEqual(1, buckets[0].high_value)
self.assertEqual(.6, buckets[0].sample_count)
self.assertEqual(1, buckets[9].low_value)
self.assertEqual(1, buckets[9].high_value)
self.assertEqual(.6, buckets[9].sample_count)
self.assertEqual(2, len(strfeat.string_stats.top_values))
self.assertEqual(3, strfeat.string_stats.top_values[0].frequency)
self.assertEqual('hi', strfeat.string_stats.top_values[0].value)
self.assertEqual(2, strfeat.string_stats.top_values[1].frequency)
self.assertEqual('hello', strfeat.string_stats.top_values[1].value)
buckets = strfeat.string_stats.rank_histogram.buckets
self.assertEqual(3, len(buckets))
self.assertEqual(0, buckets[0].low_rank)
self.assertEqual(0, buckets[0].high_rank)
self.assertEqual(3, buckets[0].sample_count)
self.assertEqual('hi', buckets[0].label)
self.assertEqual(2, buckets[2].low_rank)
self.assertEqual(2, buckets[2].high_rank)
self.assertEqual(1, buckets[2].sample_count)
self.assertEqual('hey', buckets[2].label)
def testGetProtoMultipleDatasets(self):
# Tests converting multiple datsets into the feature stats proto
# including ensuring feature order is consistent in the protos.
examples1 = []
for i in range(2):
example = tf.train.Example()
example.features.feature['str'].bytes_list.value.append(b'one')
example.features.feature['num'].int64_list.value.append(0)
examples1.append(example)
examples2 = []
example = tf.train.Example()
example.features.feature['num'].int64_list.value.append(1)
example.features.feature['str'].bytes_list.value.append(b'two')
examples2.append(example)
entries1 = {}
for i, example1 in enumerate(examples1):
self.fs._ParseExample(example1.features.feature, [], entries1, i)
entries2 = {}
for i, example2 in enumerate(examples2):
self.fs._ParseExample(example2.features.feature, [], entries2, i)
datasets = [{
'entries': entries1,
'size': len(examples1),
'name': 'test1'
}, {
'entries': entries2,
'size': len(examples2),
'name': 'test2'
}]
p = self.fs.GetDatasetsProto(datasets)
self.assertEqual(2, len(p.datasets))
test_data_1 = p.datasets[0]
self.assertEqual('test1', test_data_1.name)
self.assertEqual(2, test_data_1.num_examples)
num_feat_index = 0 if test_data_1.features[0].name == 'num' else 1
self.assertEqual(0, test_data_1.features[num_feat_index].num_stats.max)
test_data_2 = p.datasets[1]
self.assertEqual('test2', test_data_2.name)
self.assertEqual(1, test_data_2.num_examples)
self.assertEqual(1, test_data_2.features[num_feat_index].num_stats.max)
def testGetEntriesNoFiles(self):
features, num_examples = self.fs._GetEntries(['test'], 10,
lambda unused_path: [])
self.assertEqual(0, num_examples)
self.assertEqual({}, features)
@staticmethod
def get_example_iter():
def ex_iter(unused_filename):
examples = []
for i in range(50):
example = tf.train.Example()
example.features.feature['num'].int64_list.value.append(i)
examples.append(example.SerializeToString())
return examples
return ex_iter
def testGetEntries_one(self):
features, num_examples = self.fs._GetEntries(['test'], 1,
self.get_example_iter())
self.assertEqual(1, num_examples)
self.assertTrue('num' in features)
def testGetEntries_oneFile(self):
unused_features, num_examples = self.fs._GetEntries(['test'], 1000,
self.get_example_iter())
self.assertEqual(50, num_examples)
def testGetEntries_twoFiles(self):
unused_features, num_examples = self.fs._GetEntries(['test0', 'test1'],
1000,
self.get_example_iter())
self.assertEqual(100, num_examples)
def testGetEntries_stopInSecondFile(self):
unused_features, num_examples = self.fs._GetEntries([
'test@0', 'test@1', 'test@2', 'test@3', 'test@4', 'test@5', 'test@6',
'test@7', 'test@8', 'test@9'
], 75, self.get_example_iter())
self.assertEqual(75, num_examples)
if __name__ == '__main__':
googletest.main()
|
feature_statistics_generator.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Class for generating the feature_statistics proto.
The proto is used as input for the Overview visualization.
"""
from facets_overview.base_feature_statistics_generator import BaseFeatureStatisticsGenerator
import facets_overview.feature_statistics_pb2 as fs
class FeatureStatisticsGenerator(BaseFeatureStatisticsGenerator):
"""Generator of stats proto from TF data."""
def __init__(self):
BaseFeatureStatisticsGenerator.__init__(self, fs.FeatureNameStatistics,
fs.DatasetFeatureStatisticsList,
fs.Histogram)
|
base_generic_feature_statistics_generator.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Base class for generating the feature_statistics proto from generic data.
The proto is used as input for the Overview visualization.
"""
import numpy as np
import pandas as pd
import sys
class BaseGenericFeatureStatisticsGenerator(object):
"""Base class for generator of stats proto from generic data."""
def __init__(self, fs_proto, datasets_proto, histogram_proto):
self.fs_proto = fs_proto
self.datasets_proto = datasets_proto
self.histogram_proto = histogram_proto
def ProtoFromDataFrames(self, dataframes,
histogram_categorical_levels_count=None):
"""Creates a feature statistics proto from a set of pandas dataframes.
Args:
dataframes: A list of dicts describing tables for each dataset for the
proto. Each entry contains a 'table' field of the dataframe of the
data
and a 'name' field to identify the dataset in the proto.
histogram_categorical_levels_count: int, controls the maximum number of
levels to display in histograms for categorical features.
Useful to prevent codes/IDs features from bloating the stats object.
Defaults to None.
Returns:
The feature statistics proto for the provided tables.
"""
datasets = []
for dataframe in dataframes:
table = dataframe['table']
table_entries = {}
for col in table:
table_entries[col] = self.NdarrayToEntry(table[col])
datasets.append({
'entries': table_entries,
'size': len(table),
'name': dataframe['name']
})
return self.GetDatasetsProto(
datasets,
histogram_categorical_levels_count=histogram_categorical_levels_count)
def DtypeToType(self, dtype):
"""Converts a Numpy dtype to the FeatureNameStatistics.Type proto enum."""
if dtype.char in np.typecodes['AllFloat']:
return self.fs_proto.FLOAT
elif (dtype.char in np.typecodes['AllInteger'] or dtype == bool or
np.issubdtype(dtype, np.datetime64) or
np.issubdtype(dtype, np.timedelta64)):
return self.fs_proto.INT
else:
return self.fs_proto.STRING
def DtypeToNumberConverter(self, dtype):
"""Converts a Numpy dtype to a converter method if applicable.
The converter method takes in a numpy array of objects of the provided
dtype
and returns a numpy array of the numbers backing that object for
statistical
analysis. Returns None if no converter is necessary.
Args:
dtype: The numpy dtype to make a converter for.
Returns:
The converter method or None.
"""
if np.issubdtype(dtype, np.datetime64):
def DatetimesToNumbers(dt_list):
return np.array([pd.Timestamp(dt).value for dt in dt_list])
return DatetimesToNumbers
elif np.issubdtype(dtype, np.timedelta64):
def TimedetlasToNumbers(td_list):
return np.array([pd.Timedelta(td).value for td in td_list])
return TimedetlasToNumbers
else:
return None
def NdarrayToEntry(self, x):
"""Converts an ndarray to the Entry format."""
row_counts = []
for row in x:
try:
rc = np.count_nonzero(~np.isnan(row))
if rc != 0:
row_counts.append(rc)
except TypeError:
try:
row_counts.append(row.size)
except AttributeError:
row_counts.append(1)
data_type = self.DtypeToType(x.dtype)
converter = self.DtypeToNumberConverter(x.dtype)
flattened = x.ravel()
orig_size = len(flattened)
# Remove all None and nan values and count how many were removed.
flattened = flattened[flattened != np.array(None)]
if converter:
flattened = converter(flattened)
if data_type == self.fs_proto.STRING:
flattened_temp = []
for x in flattened:
try:
if str(x) != 'nan':
flattened_temp.append(x)
except UnicodeEncodeError:
if x.encode('utf-8') != 'nan':
flattened_temp.append(x)
flattened = flattened_temp
else:
flattened = flattened[~np.isnan(flattened)].tolist()
missing = orig_size - len(flattened)
return {
'vals': flattened,
'counts': row_counts,
'missing': missing,
'type': data_type
}
def GetDatasetsProto(self, datasets, features=None,
histogram_categorical_levels_count=None):
"""Generates the feature stats proto from dictionaries of feature values.
Args:
datasets: An array of dictionaries, one per dataset, each one containing:
- 'entries': The dictionary of features in the dataset from the parsed
examples.
- 'size': The number of examples parsed for the dataset.
- 'name': The name of the dataset.
features: A list of strings that is a whitelist of feature names to create
feature statistics for. If set to None then all features in the
dataset
are analyzed. Defaults to None.
histogram_categorical_levels_count: int, controls the maximum number of
levels to display in histograms for categorical features.
Useful to prevent codes/IDs features from bloating the stats object.
Defaults to None.
Returns:
The feature statistics proto for the provided datasets.
"""
features_seen = set()
whitelist_features = set(features) if features else None
all_datasets = self.datasets_proto()
# TODO(jwexler): Add ability to generate weighted feature stats
# if there is a specified weight feature in the dataset.
# Initialize each dataset
for dataset in datasets:
all_datasets.datasets.add(
name=dataset['name'], num_examples=dataset['size'])
# This outer loop ensures that for each feature seen in any of the provided
# datasets, we check the feature once against all datasets.
for outer_dataset in datasets:
for key, value in outer_dataset['entries'].items():
# If we have a feature whitelist and this feature is not in the
# whitelist then do not process it.
# If we have processed this feature already, no need to do it again.
if ((whitelist_features and key not in whitelist_features) or
key in features_seen):
continue
features_seen.add(key)
# Default to type int if no type is found, so that the fact that all
# values are missing from this feature can be displayed.
feature_type = value['type'] if 'type' in value else self.fs_proto.INT
# Process the found feature for each dataset.
for j, dataset in enumerate(datasets):
feat = all_datasets.datasets[j].features.add(
type=feature_type, name=key.encode('utf-8'))
value = dataset['entries'].get(key)
has_data = value is not None and (value['vals'].size != 0
if isinstance(
value['vals'], np.ndarray) else
value['vals'])
commonstats = None
# For numeric features, calculate numeric statistics.
if feat.type in (self.fs_proto.INT, self.fs_proto.FLOAT):
featstats = feat.num_stats
commonstats = featstats.common_stats
if has_data:
nums = value['vals']
featstats.std_dev = np.std(nums).item()
featstats.mean = np.mean(nums).item()
featstats.min = np.min(nums).item()
featstats.max = np.max(nums).item()
featstats.median = np.median(nums).item()
featstats.num_zeros = len(nums) - np.count_nonzero(nums)
nums = np.array(nums)
num_nan = len(nums[np.isnan(nums)])
num_posinf = len(nums[np.isposinf(nums)])
num_neginf = len(nums[np.isneginf(nums)])
# Remove all non-finite (including NaN) values from the numeric
# values in order to calculate histogram buckets/counts. The
# inf values will be added back to the first and last buckets.
nums = nums[np.isfinite(nums)]
counts, buckets = np.histogram(nums)
hist = featstats.histograms.add()
hist.type = self.histogram_proto.STANDARD
hist.num_nan = num_nan
for bucket_count in range(len(counts)):
bucket = hist.buckets.add(
low_value=buckets[bucket_count],
high_value=buckets[bucket_count + 1],
sample_count=counts[bucket_count].item())
# Add any negative or positive infinities to the first and last
# buckets in the histogram.
if bucket_count == 0 and num_neginf > 0:
bucket.low_value = float('-inf')
bucket.sample_count += num_neginf
elif bucket_count == len(counts) - 1 and num_posinf > 0:
bucket.high_value = float('inf')
bucket.sample_count += num_posinf
if not hist.buckets:
if num_neginf:
hist.buckets.add(
low_value=float('-inf'),
high_value=float('-inf'),
sample_count=num_neginf)
if num_posinf:
hist.buckets.add(
low_value=float('inf'),
high_value=float('inf'),
sample_count=num_posinf)
self._PopulateQuantilesHistogram(featstats.histograms.add(),nums.tolist())
elif feat.type == self.fs_proto.STRING:
featstats = feat.string_stats
commonstats = featstats.common_stats
if has_data:
strs = []
for item in value['vals']:
strs.append(item if hasattr(item, '__len__') else
item.encode('utf-8') if hasattr(item, 'encode') else str(
item))
featstats.avg_length = np.mean(np.vectorize(len)(strs))
vals, counts = np.unique(strs, return_counts=True)
featstats.unique = len(vals)
sorted_vals = sorted(zip(counts, vals), reverse=True)
sorted_vals = sorted_vals[:histogram_categorical_levels_count]
for val_index, val in enumerate(sorted_vals):
try:
if (sys.version_info.major < 3 or
isinstance(val[1], (bytes, bytearray))):
printable_val = val[1].decode('UTF-8', 'strict')
else:
printable_val = val[1]
except (UnicodeDecodeError, UnicodeEncodeError):
printable_val = '__BYTES_VALUE__'
bucket = featstats.rank_histogram.buckets.add(
low_rank=val_index,
high_rank=val_index,
sample_count=(val[0].item()),
label=printable_val)
if val_index < 2:
featstats.top_values.add(
value=bucket.label, frequency=bucket.sample_count)
# Add the common stats regardless of the feature type.
if has_data:
commonstats.num_missing = value['missing']
commonstats.num_non_missing = (all_datasets.datasets[j].num_examples
- featstats.common_stats.num_missing)
commonstats.min_num_values = int(np.min(value['counts']).astype(int))
commonstats.max_num_values = int(np.max(value['counts']).astype(int))
commonstats.avg_num_values = np.mean(value['counts'])
if 'feat_lens' in value and value['feat_lens']:
self._PopulateQuantilesHistogram(
commonstats.feature_list_length_histogram, value['feat_lens'])
self._PopulateQuantilesHistogram(commonstats.num_values_histogram,
value['counts'])
else:
commonstats.num_non_missing = 0
commonstats.num_missing = all_datasets.datasets[j].num_examples
return all_datasets
def _PopulateQuantilesHistogram(self, hist, nums):
"""Fills in the histogram with quantile information from the provided array.
Args:
hist: A Histogram proto message to fill in.
nums: A list of numbers to create a quantiles histogram from.
"""
if not nums:
return
num_quantile_buckets = 10
quantiles_to_get = [
x * 100 / num_quantile_buckets for x in range(num_quantile_buckets + 1)
]
try:
quantiles = np.percentile(nums, quantiles_to_get)
except:
quantiles = [0.0]
hist.type = self.histogram_proto.QUANTILES
quantiles_sample_count = float(len(nums)) / num_quantile_buckets
for low, high in zip(quantiles, quantiles[1:]):
hist.buckets.add(
low_value=low, high_value=high, sample_count=quantiles_sample_count)
|
base_feature_statistics_generator.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
"""Base class for generating the feature_statistics proto from TensorFlow data.
The proto is used as input for the Overview visualization.
"""
from functools import partial
from facets_overview.base_generic_feature_statistics_generator import BaseGenericFeatureStatisticsGenerator
import tensorflow as tf
# The feature name used to track sequence length when analyzing
# tf.SequenceExamples.
SEQUENCE_LENGTH_FEATURE_NAME = 'sequence length (derived feature)'
class BaseFeatureStatisticsGenerator(BaseGenericFeatureStatisticsGenerator):
"""Base class for generator of stats proto from TF data."""
def __init__(self, fs_proto, datasets_proto, histogram_proto):
BaseGenericFeatureStatisticsGenerator.__init__(
self, fs_proto, datasets_proto, histogram_proto)
def ProtoFromTfRecordFiles(self,
files,
max_entries=10000,
features=None,
is_sequence=False,
iterator_options=None,
histogram_categorical_levels_count=None):
"""Creates a feature statistics proto from a set of TFRecord files.
Args:
files: A list of dicts describing files for each dataset for the proto.
Each
entry contains a 'path' field with the path to the TFRecord file on
disk
and a 'name' field to identify the dataset in the proto.
max_entries: The maximum number of examples to load from each dataset
in order to create the proto. Defaults to 10000.
features: A list of strings that is a whitelist of feature names to create
feature statistics for. If set to None then all features in the
dataset
are analyzed. Defaults to None.
is_sequence: True if the input data from 'tables' are tf.SequenceExamples,
False if tf.Examples. Defaults to false.
iterator_options: Options to pass to the iterator that reads the examples.
Defaults to None.
histogram_categorical_levels_count: int, controls the maximum number of
levels to display in histograms for categorical features.
Useful to prevent codes/IDs features from bloating the stats object.
Defaults to None.
Returns:
The feature statistics proto for the provided files.
"""
datasets = []
for entry in files:
entries, size = self._GetTfRecordEntries(entry['path'], max_entries,
is_sequence, iterator_options)
datasets.append({'entries': entries, 'size': size, 'name': entry['name']})
return self.GetDatasetsProto(
datasets,
features,
histogram_categorical_levels_count)
def _ParseExample(self, example_features, example_feature_lists, entries,
index):
"""Parses data from an example, populating a dictionary of feature values.
Args:
example_features: A map of strings to tf.Features from the example.
example_feature_lists: A map of strings to tf.FeatureLists from the
example.
entries: A dictionary of all features parsed thus far and arrays of their
values. This is mutated by the function.
index: The index of the example to parse from a list of examples.
Raises:
TypeError: Raises an exception when a feature has inconsistent types
across
examples.
"""
features_seen = set()
for feature_list, is_feature in zip(
[example_features, example_feature_lists], [True, False]):
sequence_length = None
for feature_name in feature_list:
# If this feature has not been seen in previous examples, then
# initialize its entry into the entries dictionary.
if feature_name not in entries:
entries[feature_name] = {
'vals': [],
'counts': [],
'feat_lens': [],
'missing': index
}
feature_entry = entries[feature_name]
feature = feature_list[feature_name]
value_type = None
value_list = []
if is_feature:
# If parsing a tf.Feature, extract the type and values simply.
if feature.HasField('float_list'):
value_list = feature.float_list.value
value_type = self.fs_proto.FLOAT
elif feature.HasField('bytes_list'):
value_list = feature.bytes_list.value
value_type = self.fs_proto.STRING
elif feature.HasField('int64_list'):
value_list = feature.int64_list.value
value_type = self.fs_proto.INT
else:
# If parsing a tf.FeatureList, get the type and values by iterating
# over all Features in the FeatureList.
sequence_length = len(feature.feature)
if sequence_length != 0 and feature.feature[0].HasField('float_list'):
for feat in feature.feature:
for value in feat.float_list.value:
value_list.append(value)
value_type = self.fs_proto.FLOAT
elif sequence_length != 0 and feature.feature[0].HasField(
'bytes_list'):
for feat in feature.feature:
for value in feat.bytes_list.value:
value_list.append(value)
value_type = self.fs_proto.STRING
elif sequence_length != 0 and feature.feature[0].HasField(
'int64_list'):
for feat in feature.feature:
for value in feat.int64_list.value:
value_list.append(value)
value_type = self.fs_proto.INT
if value_type is not None:
if 'type' not in feature_entry:
feature_entry['type'] = value_type
elif feature_entry['type'] != value_type:
raise TypeError('type mismatch for feature ' + feature_name)
feature_entry['counts'].append(len(value_list))
feature_entry['vals'].extend(value_list)
if sequence_length is not None:
feature_entry['feat_lens'].append(sequence_length)
if value_list:
features_seen.add(feature_name)
# For all previously-seen features not found in this example, update the
# feature's missing value.
for f in entries:
fv = entries[f]
if f not in features_seen:
fv['missing'] += 1
def _GetEntries(self,
paths,
max_entries,
iterator_from_file,
is_sequence=False):
"""Extracts examples into a dictionary of feature values.
Args:
paths: A list of the paths to the files to parse.
max_entries: The maximum number of examples to load.
iterator_from_file: A method that takes a file path string and returns an
iterator to the examples in that file.
is_sequence: True if the input data from 'iterator_from_file' are
tf.SequenceExamples, False if tf.Examples. Defaults to false.
Returns:
A tuple with two elements:
- A dictionary of all features parsed thus far and arrays of their
values.
- The number of examples parsed.
"""
entries = {}
index = 0
for filepath in paths:
reader = iterator_from_file(filepath)
for record in reader:
if is_sequence:
sequence_example = tf.train.SequenceExample.FromString(record)
self._ParseExample(sequence_example.context.feature,
sequence_example.feature_lists.feature_list,
entries, index)
else:
self._ParseExample(
tf.train.Example.FromString(record).features.feature, [], entries,
index)
index += 1
if index == max_entries:
return entries, index
return entries, index
def _GetTfRecordEntries(self, path, max_entries, is_sequence,
iterator_options):
"""Extracts TFRecord examples into a dictionary of feature values.
Args:
path: The path to the TFRecord file(s).
max_entries: The maximum number of examples to load.
is_sequence: True if the input data from 'path' are tf.SequenceExamples,
False if tf.Examples. Defaults to false.
iterator_options: Options to pass to the iterator that reads the examples.
Defaults to None.
Returns:
A tuple with two elements:
- A dictionary of all features parsed thus far and arrays of their
values.
- The number of examples parsed.
"""
return self._GetEntries([path], max_entries,
partial(
tf.compat.v1.io.tf_record_iterator,
options=iterator_options), is_sequence)
|
feature_statistics_pb2.py | # Copyright 2017 Google Inc. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ==============================================================================
# Generated by the protocol buffer compiler. DO NOT EDIT!
# source: feature_statistics.proto
import sys
_b=sys.version_info[0]<3 and (lambda x:x) or (lambda x:x.encode('latin1'))
from google.protobuf import descriptor as _descriptor
from google.protobuf import message as _message
from google.protobuf import reflection as _reflection
from google.protobuf import symbol_database as _symbol_database
from google.protobuf import descriptor_pb2
# @@protoc_insertion_point(imports)
_sym_db = _symbol_database.Default()
DESCRIPTOR = _descriptor.FileDescriptor(
name='feature_statistics.proto',
package='featureStatistics',
syntax='proto3',
serialized_pb=_b('\n\x18\x66\x65\x61ture_statistics.proto\x12\x11\x66\x65\x61tureStatistics\"]\n\x1c\x44\x61tasetFeatureStatisticsList\x12=\n\x08\x64\x61tasets\x18\x01 \x03(\x0b\x32+.featureStatistics.DatasetFeatureStatistics\"\x99\x01\n\x18\x44\x61tasetFeatureStatistics\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\x14\n\x0cnum_examples\x18\x02 \x01(\x04\x12\x1d\n\x15weighted_num_examples\x18\x04 \x01(\x01\x12:\n\x08\x66\x65\x61tures\x18\x03 \x03(\x0b\x32(.featureStatistics.FeatureNameStatistics\"\x8b\x03\n\x15\x46\x65\x61tureNameStatistics\x12\x0c\n\x04name\x18\x01 \x01(\t\x12;\n\x04type\x18\x02 \x01(\x0e\x32-.featureStatistics.FeatureNameStatistics.Type\x12\x39\n\tnum_stats\x18\x03 \x01(\x0b\x32$.featureStatistics.NumericStatisticsH\x00\x12;\n\x0cstring_stats\x18\x04 \x01(\x0b\x32#.featureStatistics.StringStatisticsH\x00\x12\x39\n\x0b\x62ytes_stats\x18\x05 \x01(\x0b\x32\".featureStatistics.BytesStatisticsH\x00\x12\x38\n\x0c\x63ustom_stats\x18\x06 \x03(\x0b\x32\".featureStatistics.CustomStatistic\"1\n\x04Type\x12\x07\n\x03INT\x10\x00\x12\t\n\x05\x46LOAT\x10\x01\x12\n\n\x06STRING\x10\x02\x12\t\n\x05\x42YTES\x10\x03\x42\x07\n\x05stats\"x\n\x18WeightedCommonStatistics\x12\x17\n\x0fnum_non_missing\x18\x01 \x01(\x01\x12\x13\n\x0bnum_missing\x18\x02 \x01(\x01\x12\x16\n\x0e\x61vg_num_values\x18\x03 \x01(\x01\x12\x16\n\x0etot_num_values\x18\x04 \x01(\x01\"w\n\x0f\x43ustomStatistic\x12\x0c\n\x04name\x18\x01 \x01(\t\x12\r\n\x03num\x18\x02 \x01(\x01H\x00\x12\r\n\x03str\x18\x03 \x01(\tH\x00\x12\x31\n\thistogram\x18\x04 \x01(\x0b\x32\x1c.featureStatistics.HistogramH\x00\x42\x05\n\x03val\"\xaa\x02\n\x11NumericStatistics\x12\x39\n\x0c\x63ommon_stats\x18\x01 \x01(\x0b\x32#.featureStatistics.CommonStatistics\x12\x0c\n\x04mean\x18\x02 \x01(\x01\x12\x0f\n\x07std_dev\x18\x03 \x01(\x01\x12\x11\n\tnum_zeros\x18\x04 \x01(\x04\x12\x0b\n\x03min\x18\x05 \x01(\x01\x12\x0e\n\x06median\x18\x06 \x01(\x01\x12\x0b\n\x03max\x18\x07 \x01(\x01\x12\x30\n\nhistograms\x18\x08 \x03(\x0b\x32\x1c.featureStatistics.Histogram\x12L\n\x16weighted_numeric_stats\x18\t \x01(\x0b\x32,.featureStatistics.WeightedNumericStatistics\"\x8c\x03\n\x10StringStatistics\x12\x39\n\x0c\x63ommon_stats\x18\x01 \x01(\x0b\x32#.featureStatistics.CommonStatistics\x12\x0e\n\x06unique\x18\x02 \x01(\x04\x12\x44\n\ntop_values\x18\x03 \x03(\x0b\x32\x30.featureStatistics.StringStatistics.FreqAndValue\x12\x12\n\navg_length\x18\x04 \x01(\x02\x12\x38\n\x0erank_histogram\x18\x05 \x01(\x0b\x32 .featureStatistics.RankHistogram\x12J\n\x15weighted_string_stats\x18\x06 \x01(\x0b\x32+.featureStatistics.WeightedStringStatistics\x1aM\n\x0c\x46reqAndValue\x12\x1b\n\x0f\x64\x65precated_freq\x18\x01 \x01(\x04\x42\x02\x18\x01\x12\r\n\x05value\x18\x02 \x01(\t\x12\x11\n\tfrequency\x18\x03 \x01(\x01\"|\n\x19WeightedNumericStatistics\x12\x0c\n\x04mean\x18\x01 \x01(\x01\x12\x0f\n\x07std_dev\x18\x02 \x01(\x01\x12\x0e\n\x06median\x18\x03 \x01(\x01\x12\x30\n\nhistograms\x18\x04 \x03(\x0b\x32\x1c.featureStatistics.Histogram\"\x9a\x01\n\x18WeightedStringStatistics\x12\x44\n\ntop_values\x18\x01 \x03(\x0b\x32\x30.featureStatistics.StringStatistics.FreqAndValue\x12\x38\n\x0erank_histogram\x18\x02 \x01(\x0b\x32 .featureStatistics.RankHistogram\"\xa1\x01\n\x0f\x42ytesStatistics\x12\x39\n\x0c\x63ommon_stats\x18\x01 \x01(\x0b\x32#.featureStatistics.CommonStatistics\x12\x0e\n\x06unique\x18\x02 \x01(\x04\x12\x15\n\ravg_num_bytes\x18\x03 \x01(\x02\x12\x15\n\rmin_num_bytes\x18\x04 \x01(\x02\x12\x15\n\rmax_num_bytes\x18\x05 \x01(\x02\"\xed\x02\n\x10\x43ommonStatistics\x12\x17\n\x0fnum_non_missing\x18\x01 \x01(\x04\x12\x13\n\x0bnum_missing\x18\x02 \x01(\x04\x12\x16\n\x0emin_num_values\x18\x03 \x01(\x04\x12\x16\n\x0emax_num_values\x18\x04 \x01(\x04\x12\x16\n\x0e\x61vg_num_values\x18\x05 \x01(\x02\x12\x16\n\x0etot_num_values\x18\x08 \x01(\x04\x12:\n\x14num_values_histogram\x18\x06 \x01(\x0b\x32\x1c.featureStatistics.Histogram\x12J\n\x15weighted_common_stats\x18\x07 \x01(\x0b\x32+.featureStatistics.WeightedCommonStatistics\x12\x43\n\x1d\x66\x65\x61ture_list_length_histogram\x18\t \x01(\x0b\x32\x1c.featureStatistics.Histogram\"\xc4\x02\n\tHistogram\x12\x0f\n\x07num_nan\x18\x01 \x01(\x04\x12\x15\n\rnum_undefined\x18\x02 \x01(\x04\x12\x34\n\x07\x62uckets\x18\x03 \x03(\x0b\x32#.featureStatistics.Histogram.Bucket\x12\x38\n\x04type\x18\x04 \x01(\x0e\x32*.featureStatistics.Histogram.HistogramType\x12\x0c\n\x04name\x18\x05 \x01(\t\x1a\x63\n\x06\x42ucket\x12\x11\n\tlow_value\x18\x01 \x01(\x01\x12\x12\n\nhigh_value\x18\x02 \x01(\x01\x12\x1c\n\x10\x64\x65precated_count\x18\x03 \x01(\x04\x42\x02\x18\x01\x12\x14\n\x0csample_count\x18\x04 \x01(\x01\",\n\rHistogramType\x12\x0c\n\x08STANDARD\x10\x00\x12\r\n\tQUANTILES\x10\x01\"\xc9\x01\n\rRankHistogram\x12\x38\n\x07\x62uckets\x18\x01 \x03(\x0b\x32\'.featureStatistics.RankHistogram.Bucket\x12\x0c\n\x04name\x18\x02 \x01(\t\x1ap\n\x06\x42ucket\x12\x10\n\x08low_rank\x18\x01 \x01(\x04\x12\x11\n\thigh_rank\x18\x02 \x01(\x04\x12\x1c\n\x10\x64\x65precated_count\x18\x03 \x01(\x04\x42\x02\x18\x01\x12\r\n\x05label\x18\x04 \x01(\t\x12\x14\n\x0csample_count\x18\x05 \x01(\x01\x62\x06proto3')
)
_FEATURENAMESTATISTICS_TYPE = _descriptor.EnumDescriptor(
name='Type',
full_name='featureStatistics.FeatureNameStatistics.Type',
filename=None,
file=DESCRIPTOR,
values=[
_descriptor.EnumValueDescriptor(
name='INT', index=0, number=0,
options=None,
type=None),
_descriptor.EnumValueDescriptor(
name='FLOAT', index=1, number=1,
options=None,
type=None),
_descriptor.EnumValueDescriptor(
name='STRING', index=2, number=2,
options=None,
type=None),
_descriptor.EnumValueDescriptor(
name='BYTES', index=3, number=3,
options=None,
type=None),
],
containing_type=None,
options=None,
serialized_start=636,
serialized_end=685,
)
_sym_db.RegisterEnumDescriptor(_FEATURENAMESTATISTICS_TYPE)
_HISTOGRAM_HISTOGRAMTYPE = _descriptor.EnumDescriptor(
name='HistogramType',
full_name='featureStatistics.Histogram.HistogramType',
filename=None,
file=DESCRIPTOR,
values=[
_descriptor.EnumValueDescriptor(
name='STANDARD', index=0, number=0,
options=None,
type=None),
_descriptor.EnumValueDescriptor(
name='QUANTILES', index=1, number=1,
options=None,
type=None),
],
containing_type=None,
options=None,
serialized_start=2735,
serialized_end=2779,
)
_sym_db.RegisterEnumDescriptor(_HISTOGRAM_HISTOGRAMTYPE)
_DATASETFEATURESTATISTICSLIST = _descriptor.Descriptor(
name='DatasetFeatureStatisticsList',
full_name='featureStatistics.DatasetFeatureStatisticsList',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='datasets', full_name='featureStatistics.DatasetFeatureStatisticsList.datasets', index=0,
number=1, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=47,
serialized_end=140,
)
_DATASETFEATURESTATISTICS = _descriptor.Descriptor(
name='DatasetFeatureStatistics',
full_name='featureStatistics.DatasetFeatureStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='name', full_name='featureStatistics.DatasetFeatureStatistics.name', index=0,
number=1, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num_examples', full_name='featureStatistics.DatasetFeatureStatistics.num_examples', index=1,
number=2, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='weighted_num_examples', full_name='featureStatistics.DatasetFeatureStatistics.weighted_num_examples', index=2,
number=4, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='features', full_name='featureStatistics.DatasetFeatureStatistics.features', index=3,
number=3, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=143,
serialized_end=296,
)
_FEATURENAMESTATISTICS = _descriptor.Descriptor(
name='FeatureNameStatistics',
full_name='featureStatistics.FeatureNameStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='name', full_name='featureStatistics.FeatureNameStatistics.name', index=0,
number=1, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='type', full_name='featureStatistics.FeatureNameStatistics.type', index=1,
number=2, type=14, cpp_type=8, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num_stats', full_name='featureStatistics.FeatureNameStatistics.num_stats', index=2,
number=3, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='string_stats', full_name='featureStatistics.FeatureNameStatistics.string_stats', index=3,
number=4, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='bytes_stats', full_name='featureStatistics.FeatureNameStatistics.bytes_stats', index=4,
number=5, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='custom_stats', full_name='featureStatistics.FeatureNameStatistics.custom_stats', index=5,
number=6, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
_FEATURENAMESTATISTICS_TYPE,
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
_descriptor.OneofDescriptor(
name='stats', full_name='featureStatistics.FeatureNameStatistics.stats',
index=0, containing_type=None, fields=[]),
],
serialized_start=299,
serialized_end=694,
)
_WEIGHTEDCOMMONSTATISTICS = _descriptor.Descriptor(
name='WeightedCommonStatistics',
full_name='featureStatistics.WeightedCommonStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='num_non_missing', full_name='featureStatistics.WeightedCommonStatistics.num_non_missing', index=0,
number=1, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num_missing', full_name='featureStatistics.WeightedCommonStatistics.num_missing', index=1,
number=2, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='avg_num_values', full_name='featureStatistics.WeightedCommonStatistics.avg_num_values', index=2,
number=3, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='tot_num_values', full_name='featureStatistics.WeightedCommonStatistics.tot_num_values', index=3,
number=4, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=696,
serialized_end=816,
)
_CUSTOMSTATISTIC = _descriptor.Descriptor(
name='CustomStatistic',
full_name='featureStatistics.CustomStatistic',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='name', full_name='featureStatistics.CustomStatistic.name', index=0,
number=1, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num', full_name='featureStatistics.CustomStatistic.num', index=1,
number=2, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='str', full_name='featureStatistics.CustomStatistic.str', index=2,
number=3, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='histogram', full_name='featureStatistics.CustomStatistic.histogram', index=3,
number=4, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
_descriptor.OneofDescriptor(
name='val', full_name='featureStatistics.CustomStatistic.val',
index=0, containing_type=None, fields=[]),
],
serialized_start=818,
serialized_end=937,
)
_NUMERICSTATISTICS = _descriptor.Descriptor(
name='NumericStatistics',
full_name='featureStatistics.NumericStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='common_stats', full_name='featureStatistics.NumericStatistics.common_stats', index=0,
number=1, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='mean', full_name='featureStatistics.NumericStatistics.mean', index=1,
number=2, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='std_dev', full_name='featureStatistics.NumericStatistics.std_dev', index=2,
number=3, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num_zeros', full_name='featureStatistics.NumericStatistics.num_zeros', index=3,
number=4, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='min', full_name='featureStatistics.NumericStatistics.min', index=4,
number=5, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='median', full_name='featureStatistics.NumericStatistics.median', index=5,
number=6, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='max', full_name='featureStatistics.NumericStatistics.max', index=6,
number=7, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='histograms', full_name='featureStatistics.NumericStatistics.histograms', index=7,
number=8, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='weighted_numeric_stats', full_name='featureStatistics.NumericStatistics.weighted_numeric_stats', index=8,
number=9, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=940,
serialized_end=1238,
)
_STRINGSTATISTICS_FREQANDVALUE = _descriptor.Descriptor(
name='FreqAndValue',
full_name='featureStatistics.StringStatistics.FreqAndValue',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='deprecated_freq', full_name='featureStatistics.StringStatistics.FreqAndValue.deprecated_freq', index=0,
number=1, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=_descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b('\030\001'))),
_descriptor.FieldDescriptor(
name='value', full_name='featureStatistics.StringStatistics.FreqAndValue.value', index=1,
number=2, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='frequency', full_name='featureStatistics.StringStatistics.FreqAndValue.frequency', index=2,
number=3, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=1560,
serialized_end=1637,
)
_STRINGSTATISTICS = _descriptor.Descriptor(
name='StringStatistics',
full_name='featureStatistics.StringStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='common_stats', full_name='featureStatistics.StringStatistics.common_stats', index=0,
number=1, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='unique', full_name='featureStatistics.StringStatistics.unique', index=1,
number=2, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='top_values', full_name='featureStatistics.StringStatistics.top_values', index=2,
number=3, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='avg_length', full_name='featureStatistics.StringStatistics.avg_length', index=3,
number=4, type=2, cpp_type=6, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='rank_histogram', full_name='featureStatistics.StringStatistics.rank_histogram', index=4,
number=5, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='weighted_string_stats', full_name='featureStatistics.StringStatistics.weighted_string_stats', index=5,
number=6, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[_STRINGSTATISTICS_FREQANDVALUE, ],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=1241,
serialized_end=1637,
)
_WEIGHTEDNUMERICSTATISTICS = _descriptor.Descriptor(
name='WeightedNumericStatistics',
full_name='featureStatistics.WeightedNumericStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='mean', full_name='featureStatistics.WeightedNumericStatistics.mean', index=0,
number=1, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='std_dev', full_name='featureStatistics.WeightedNumericStatistics.std_dev', index=1,
number=2, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='median', full_name='featureStatistics.WeightedNumericStatistics.median', index=2,
number=3, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='histograms', full_name='featureStatistics.WeightedNumericStatistics.histograms', index=3,
number=4, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=1639,
serialized_end=1763,
)
_WEIGHTEDSTRINGSTATISTICS = _descriptor.Descriptor(
name='WeightedStringStatistics',
full_name='featureStatistics.WeightedStringStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='top_values', full_name='featureStatistics.WeightedStringStatistics.top_values', index=0,
number=1, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='rank_histogram', full_name='featureStatistics.WeightedStringStatistics.rank_histogram', index=1,
number=2, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=1766,
serialized_end=1920,
)
_BYTESSTATISTICS = _descriptor.Descriptor(
name='BytesStatistics',
full_name='featureStatistics.BytesStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='common_stats', full_name='featureStatistics.BytesStatistics.common_stats', index=0,
number=1, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='unique', full_name='featureStatistics.BytesStatistics.unique', index=1,
number=2, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='avg_num_bytes', full_name='featureStatistics.BytesStatistics.avg_num_bytes', index=2,
number=3, type=2, cpp_type=6, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='min_num_bytes', full_name='featureStatistics.BytesStatistics.min_num_bytes', index=3,
number=4, type=2, cpp_type=6, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='max_num_bytes', full_name='featureStatistics.BytesStatistics.max_num_bytes', index=4,
number=5, type=2, cpp_type=6, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=1923,
serialized_end=2084,
)
_COMMONSTATISTICS = _descriptor.Descriptor(
name='CommonStatistics',
full_name='featureStatistics.CommonStatistics',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='num_non_missing', full_name='featureStatistics.CommonStatistics.num_non_missing', index=0,
number=1, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num_missing', full_name='featureStatistics.CommonStatistics.num_missing', index=1,
number=2, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='min_num_values', full_name='featureStatistics.CommonStatistics.min_num_values', index=2,
number=3, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='max_num_values', full_name='featureStatistics.CommonStatistics.max_num_values', index=3,
number=4, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='avg_num_values', full_name='featureStatistics.CommonStatistics.avg_num_values', index=4,
number=5, type=2, cpp_type=6, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='tot_num_values', full_name='featureStatistics.CommonStatistics.tot_num_values', index=5,
number=8, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num_values_histogram', full_name='featureStatistics.CommonStatistics.num_values_histogram', index=6,
number=6, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='weighted_common_stats', full_name='featureStatistics.CommonStatistics.weighted_common_stats', index=7,
number=7, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='feature_list_length_histogram', full_name='featureStatistics.CommonStatistics.feature_list_length_histogram', index=8,
number=9, type=11, cpp_type=10, label=1,
has_default_value=False, default_value=None,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=2087,
serialized_end=2452,
)
_HISTOGRAM_BUCKET = _descriptor.Descriptor(
name='Bucket',
full_name='featureStatistics.Histogram.Bucket',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='low_value', full_name='featureStatistics.Histogram.Bucket.low_value', index=0,
number=1, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='high_value', full_name='featureStatistics.Histogram.Bucket.high_value', index=1,
number=2, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='deprecated_count', full_name='featureStatistics.Histogram.Bucket.deprecated_count', index=2,
number=3, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=_descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b('\030\001'))),
_descriptor.FieldDescriptor(
name='sample_count', full_name='featureStatistics.Histogram.Bucket.sample_count', index=3,
number=4, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=2634,
serialized_end=2733,
)
_HISTOGRAM = _descriptor.Descriptor(
name='Histogram',
full_name='featureStatistics.Histogram',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='num_nan', full_name='featureStatistics.Histogram.num_nan', index=0,
number=1, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='num_undefined', full_name='featureStatistics.Histogram.num_undefined', index=1,
number=2, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='buckets', full_name='featureStatistics.Histogram.buckets', index=2,
number=3, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='type', full_name='featureStatistics.Histogram.type', index=3,
number=4, type=14, cpp_type=8, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='name', full_name='featureStatistics.Histogram.name', index=4,
number=5, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[_HISTOGRAM_BUCKET, ],
enum_types=[
_HISTOGRAM_HISTOGRAMTYPE,
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=2455,
serialized_end=2779,
)
_RANKHISTOGRAM_BUCKET = _descriptor.Descriptor(
name='Bucket',
full_name='featureStatistics.RankHistogram.Bucket',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='low_rank', full_name='featureStatistics.RankHistogram.Bucket.low_rank', index=0,
number=1, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='high_rank', full_name='featureStatistics.RankHistogram.Bucket.high_rank', index=1,
number=2, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='deprecated_count', full_name='featureStatistics.RankHistogram.Bucket.deprecated_count', index=2,
number=3, type=4, cpp_type=4, label=1,
has_default_value=False, default_value=0,
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=_descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b('\030\001'))),
_descriptor.FieldDescriptor(
name='label', full_name='featureStatistics.RankHistogram.Bucket.label', index=3,
number=4, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='sample_count', full_name='featureStatistics.RankHistogram.Bucket.sample_count', index=4,
number=5, type=1, cpp_type=5, label=1,
has_default_value=False, default_value=float(0),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=2871,
serialized_end=2983,
)
_RANKHISTOGRAM = _descriptor.Descriptor(
name='RankHistogram',
full_name='featureStatistics.RankHistogram',
filename=None,
file=DESCRIPTOR,
containing_type=None,
fields=[
_descriptor.FieldDescriptor(
name='buckets', full_name='featureStatistics.RankHistogram.buckets', index=0,
number=1, type=11, cpp_type=10, label=3,
has_default_value=False, default_value=[],
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
_descriptor.FieldDescriptor(
name='name', full_name='featureStatistics.RankHistogram.name', index=1,
number=2, type=9, cpp_type=9, label=1,
has_default_value=False, default_value=_b("").decode('utf-8'),
message_type=None, enum_type=None, containing_type=None,
is_extension=False, extension_scope=None,
options=None),
],
extensions=[
],
nested_types=[_RANKHISTOGRAM_BUCKET, ],
enum_types=[
],
options=None,
is_extendable=False,
syntax='proto3',
extension_ranges=[],
oneofs=[
],
serialized_start=2782,
serialized_end=2983,
)
_DATASETFEATURESTATISTICSLIST.fields_by_name['datasets'].message_type = _DATASETFEATURESTATISTICS
_DATASETFEATURESTATISTICS.fields_by_name['features'].message_type = _FEATURENAMESTATISTICS
_FEATURENAMESTATISTICS.fields_by_name['type'].enum_type = _FEATURENAMESTATISTICS_TYPE
_FEATURENAMESTATISTICS.fields_by_name['num_stats'].message_type = _NUMERICSTATISTICS
_FEATURENAMESTATISTICS.fields_by_name['string_stats'].message_type = _STRINGSTATISTICS
_FEATURENAMESTATISTICS.fields_by_name['bytes_stats'].message_type = _BYTESSTATISTICS
_FEATURENAMESTATISTICS.fields_by_name['custom_stats'].message_type = _CUSTOMSTATISTIC
_FEATURENAMESTATISTICS_TYPE.containing_type = _FEATURENAMESTATISTICS
_FEATURENAMESTATISTICS.oneofs_by_name['stats'].fields.append(
_FEATURENAMESTATISTICS.fields_by_name['num_stats'])
_FEATURENAMESTATISTICS.fields_by_name['num_stats'].containing_oneof = _FEATURENAMESTATISTICS.oneofs_by_name['stats']
_FEATURENAMESTATISTICS.oneofs_by_name['stats'].fields.append(
_FEATURENAMESTATISTICS.fields_by_name['string_stats'])
_FEATURENAMESTATISTICS.fields_by_name['string_stats'].containing_oneof = _FEATURENAMESTATISTICS.oneofs_by_name['stats']
_FEATURENAMESTATISTICS.oneofs_by_name['stats'].fields.append(
_FEATURENAMESTATISTICS.fields_by_name['bytes_stats'])
_FEATURENAMESTATISTICS.fields_by_name['bytes_stats'].containing_oneof = _FEATURENAMESTATISTICS.oneofs_by_name['stats']
_CUSTOMSTATISTIC.fields_by_name['histogram'].message_type = _HISTOGRAM
_CUSTOMSTATISTIC.oneofs_by_name['val'].fields.append(
_CUSTOMSTATISTIC.fields_by_name['num'])
_CUSTOMSTATISTIC.fields_by_name['num'].containing_oneof = _CUSTOMSTATISTIC.oneofs_by_name['val']
_CUSTOMSTATISTIC.oneofs_by_name['val'].fields.append(
_CUSTOMSTATISTIC.fields_by_name['str'])
_CUSTOMSTATISTIC.fields_by_name['str'].containing_oneof = _CUSTOMSTATISTIC.oneofs_by_name['val']
_CUSTOMSTATISTIC.oneofs_by_name['val'].fields.append(
_CUSTOMSTATISTIC.fields_by_name['histogram'])
_CUSTOMSTATISTIC.fields_by_name['histogram'].containing_oneof = _CUSTOMSTATISTIC.oneofs_by_name['val']
_NUMERICSTATISTICS.fields_by_name['common_stats'].message_type = _COMMONSTATISTICS
_NUMERICSTATISTICS.fields_by_name['histograms'].message_type = _HISTOGRAM
_NUMERICSTATISTICS.fields_by_name['weighted_numeric_stats'].message_type = _WEIGHTEDNUMERICSTATISTICS
_STRINGSTATISTICS_FREQANDVALUE.containing_type = _STRINGSTATISTICS
_STRINGSTATISTICS.fields_by_name['common_stats'].message_type = _COMMONSTATISTICS
_STRINGSTATISTICS.fields_by_name['top_values'].message_type = _STRINGSTATISTICS_FREQANDVALUE
_STRINGSTATISTICS.fields_by_name['rank_histogram'].message_type = _RANKHISTOGRAM
_STRINGSTATISTICS.fields_by_name['weighted_string_stats'].message_type = _WEIGHTEDSTRINGSTATISTICS
_WEIGHTEDNUMERICSTATISTICS.fields_by_name['histograms'].message_type = _HISTOGRAM
_WEIGHTEDSTRINGSTATISTICS.fields_by_name['top_values'].message_type = _STRINGSTATISTICS_FREQANDVALUE
_WEIGHTEDSTRINGSTATISTICS.fields_by_name['rank_histogram'].message_type = _RANKHISTOGRAM
_BYTESSTATISTICS.fields_by_name['common_stats'].message_type = _COMMONSTATISTICS
_COMMONSTATISTICS.fields_by_name['num_values_histogram'].message_type = _HISTOGRAM
_COMMONSTATISTICS.fields_by_name['weighted_common_stats'].message_type = _WEIGHTEDCOMMONSTATISTICS
_COMMONSTATISTICS.fields_by_name['feature_list_length_histogram'].message_type = _HISTOGRAM
_HISTOGRAM_BUCKET.containing_type = _HISTOGRAM
_HISTOGRAM.fields_by_name['buckets'].message_type = _HISTOGRAM_BUCKET
_HISTOGRAM.fields_by_name['type'].enum_type = _HISTOGRAM_HISTOGRAMTYPE
_HISTOGRAM_HISTOGRAMTYPE.containing_type = _HISTOGRAM
_RANKHISTOGRAM_BUCKET.containing_type = _RANKHISTOGRAM
_RANKHISTOGRAM.fields_by_name['buckets'].message_type = _RANKHISTOGRAM_BUCKET
DESCRIPTOR.message_types_by_name['DatasetFeatureStatisticsList'] = _DATASETFEATURESTATISTICSLIST
DESCRIPTOR.message_types_by_name['DatasetFeatureStatistics'] = _DATASETFEATURESTATISTICS
DESCRIPTOR.message_types_by_name['FeatureNameStatistics'] = _FEATURENAMESTATISTICS
DESCRIPTOR.message_types_by_name['WeightedCommonStatistics'] = _WEIGHTEDCOMMONSTATISTICS
DESCRIPTOR.message_types_by_name['CustomStatistic'] = _CUSTOMSTATISTIC
DESCRIPTOR.message_types_by_name['NumericStatistics'] = _NUMERICSTATISTICS
DESCRIPTOR.message_types_by_name['StringStatistics'] = _STRINGSTATISTICS
DESCRIPTOR.message_types_by_name['WeightedNumericStatistics'] = _WEIGHTEDNUMERICSTATISTICS
DESCRIPTOR.message_types_by_name['WeightedStringStatistics'] = _WEIGHTEDSTRINGSTATISTICS
DESCRIPTOR.message_types_by_name['BytesStatistics'] = _BYTESSTATISTICS
DESCRIPTOR.message_types_by_name['CommonStatistics'] = _COMMONSTATISTICS
DESCRIPTOR.message_types_by_name['Histogram'] = _HISTOGRAM
DESCRIPTOR.message_types_by_name['RankHistogram'] = _RANKHISTOGRAM
_sym_db.RegisterFileDescriptor(DESCRIPTOR)
DatasetFeatureStatisticsList = _reflection.GeneratedProtocolMessageType('DatasetFeatureStatisticsList', (_message.Message,), dict(
DESCRIPTOR = _DATASETFEATURESTATISTICSLIST,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.DatasetFeatureStatisticsList)
))
_sym_db.RegisterMessage(DatasetFeatureStatisticsList)
DatasetFeatureStatistics = _reflection.GeneratedProtocolMessageType('DatasetFeatureStatistics', (_message.Message,), dict(
DESCRIPTOR = _DATASETFEATURESTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.DatasetFeatureStatistics)
))
_sym_db.RegisterMessage(DatasetFeatureStatistics)
FeatureNameStatistics = _reflection.GeneratedProtocolMessageType('FeatureNameStatistics', (_message.Message,), dict(
DESCRIPTOR = _FEATURENAMESTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.FeatureNameStatistics)
))
_sym_db.RegisterMessage(FeatureNameStatistics)
WeightedCommonStatistics = _reflection.GeneratedProtocolMessageType('WeightedCommonStatistics', (_message.Message,), dict(
DESCRIPTOR = _WEIGHTEDCOMMONSTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.WeightedCommonStatistics)
))
_sym_db.RegisterMessage(WeightedCommonStatistics)
CustomStatistic = _reflection.GeneratedProtocolMessageType('CustomStatistic', (_message.Message,), dict(
DESCRIPTOR = _CUSTOMSTATISTIC,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.CustomStatistic)
))
_sym_db.RegisterMessage(CustomStatistic)
NumericStatistics = _reflection.GeneratedProtocolMessageType('NumericStatistics', (_message.Message,), dict(
DESCRIPTOR = _NUMERICSTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.NumericStatistics)
))
_sym_db.RegisterMessage(NumericStatistics)
StringStatistics = _reflection.GeneratedProtocolMessageType('StringStatistics', (_message.Message,), dict(
FreqAndValue = _reflection.GeneratedProtocolMessageType('FreqAndValue', (_message.Message,), dict(
DESCRIPTOR = _STRINGSTATISTICS_FREQANDVALUE,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.StringStatistics.FreqAndValue)
))
,
DESCRIPTOR = _STRINGSTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.StringStatistics)
))
_sym_db.RegisterMessage(StringStatistics)
_sym_db.RegisterMessage(StringStatistics.FreqAndValue)
WeightedNumericStatistics = _reflection.GeneratedProtocolMessageType('WeightedNumericStatistics', (_message.Message,), dict(
DESCRIPTOR = _WEIGHTEDNUMERICSTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.WeightedNumericStatistics)
))
_sym_db.RegisterMessage(WeightedNumericStatistics)
WeightedStringStatistics = _reflection.GeneratedProtocolMessageType('WeightedStringStatistics', (_message.Message,), dict(
DESCRIPTOR = _WEIGHTEDSTRINGSTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.WeightedStringStatistics)
))
_sym_db.RegisterMessage(WeightedStringStatistics)
BytesStatistics = _reflection.GeneratedProtocolMessageType('BytesStatistics', (_message.Message,), dict(
DESCRIPTOR = _BYTESSTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.BytesStatistics)
))
_sym_db.RegisterMessage(BytesStatistics)
CommonStatistics = _reflection.GeneratedProtocolMessageType('CommonStatistics', (_message.Message,), dict(
DESCRIPTOR = _COMMONSTATISTICS,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.CommonStatistics)
))
_sym_db.RegisterMessage(CommonStatistics)
Histogram = _reflection.GeneratedProtocolMessageType('Histogram', (_message.Message,), dict(
Bucket = _reflection.GeneratedProtocolMessageType('Bucket', (_message.Message,), dict(
DESCRIPTOR = _HISTOGRAM_BUCKET,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.Histogram.Bucket)
))
,
DESCRIPTOR = _HISTOGRAM,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.Histogram)
))
_sym_db.RegisterMessage(Histogram)
_sym_db.RegisterMessage(Histogram.Bucket)
RankHistogram = _reflection.GeneratedProtocolMessageType('RankHistogram', (_message.Message,), dict(
Bucket = _reflection.GeneratedProtocolMessageType('Bucket', (_message.Message,), dict(
DESCRIPTOR = _RANKHISTOGRAM_BUCKET,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.RankHistogram.Bucket)
))
,
DESCRIPTOR = _RANKHISTOGRAM,
__module__ = 'feature_statistics_pb2'
# @@protoc_insertion_point(class_scope:featureStatistics.RankHistogram)
))
_sym_db.RegisterMessage(RankHistogram)
_sym_db.RegisterMessage(RankHistogram.Bucket)
_STRINGSTATISTICS_FREQANDVALUE.fields_by_name['deprecated_freq'].has_options = True
_STRINGSTATISTICS_FREQANDVALUE.fields_by_name['deprecated_freq']._options = _descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b('\030\001'))
_HISTOGRAM_BUCKET.fields_by_name['deprecated_count'].has_options = True
_HISTOGRAM_BUCKET.fields_by_name['deprecated_count']._options = _descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b('\030\001'))
_RANKHISTOGRAM_BUCKET.fields_by_name['deprecated_count'].has_options = True
_RANKHISTOGRAM_BUCKET.fields_by_name['deprecated_count']._options = _descriptor._ParseOptions(descriptor_pb2.FieldOptions(), _b('\030\001'))
# @@protoc_insertion_point(module_scope)
|
run_modelService.py | #from BaseHTTPServer import BaseHTTPRequestHandler,HTTPServer
from http.server import BaseHTTPRequestHandler,HTTPServer
#from SocketServer import ThreadingMixIn
from socketserver import ThreadingMixIn
'''
from augustus.core.ModelLoader import ModelLoader
from augustus.strict import modelLoader
'''
import pandas as pd
import os,sys
from os.path import expanduser
import platform
import numpy as np
import configparser
import threading
import subprocess
import argparse
import re
import cgi
from datetime import datetime
import json
import sys
from datetime import datetime
user_records = {}
class LocalModelData(object):
models = {}
class HTTPRequestHandler(BaseHTTPRequestHandler):
def do_POST(self):
print("PYTHON ######## REQUEST ####### STARTED")
if None != re.search('/AION/', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
data = self.rfile.read(length)
model = self.path.split('/')[-2]
operation = self.path.split('/')[-1]
data = json.loads(data)
dataStr = json.dumps(data)
if operation.lower() == 'predict':
predict_path = os.path.join(os.path.dirname(os.path.abspath(__file__)),'AION','aion_predict.py')
outputStr = subprocess.check_output([sys.executable,predict_path,dataStr])
outputStr = outputStr.decode('utf-8')
outputStr = re.search(r'predictions:(.*)',str(outputStr), re.IGNORECASE).group(1)
outputStr = outputStr.strip()
resp = outputStr
elif operation.lower() == 'explain':
predict_path = os.path.join(os.path.dirname(os.path.abspath(__file__)),'AION','aion_xai.py')
outputStr = subprocess.check_output([sys.executable,predict_path,'local',dataStr])
outputStr = outputStr.decode('utf-8')
outputStr = re.search(r'aion_ai_explanation:(.*)',str(outputStr), re.IGNORECASE).group(1)
outputStr = outputStr.strip()
resp = outputStr
resp=resp+"\n"
resp=resp.encode()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
elif None != re.search('/AION/pattern_anomaly_predict', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
model = self.path.split('/')[-1]
data = self.rfile.read(length)
data = json.loads(data)
anomaly = False
remarks = ''
clusterid = -1
home = expanduser("~")
if platform.system() == 'Windows':
configfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'datadetails.json')
filename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'clickstream.json')
clusterfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'stateClustering.csv')
probfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'stateTransitionProbability.csv')
else:
configfilename = os.path.join(home,'HCLT','AION','target',model,'datadetails.json')
filename = os.path.join(home,'HCLT','AION','target',model,'clickstream.json')
clusterfilename = os.path.join(home,'HCLT','AION','target',model,'stateClustering.csv')
probfilename = os.path.join(home,'HCLT','AION','target',model,'stateTransitionProbability.csv')
dfclus = pd.read_csv(clusterfilename)
dfprod = pd.read_csv(probfilename)
f = open(configfilename, "r")
configSettings = f.read()
f.close()
configSettingsJson = json.loads(configSettings)
activity = configSettingsJson['activity']
sessionid = configSettingsJson['sessionid']
f = open(filename, "r")
configSettings = f.read()
f.close()
configSettingsJson = json.loads(configSettings)
groupswitching = configSettingsJson['groupswitching']
page_threshold = configSettingsJson['transitionprobability']
chain_count = configSettingsJson['transitionsequence']
chain_probability = configSettingsJson['sequencethreshold']
currentactivity = data[activity]
if bool(user_records):
sessionid = data[sessionid]
print(sessionid,user_records['SessionID'])
if sessionid != user_records['SessionID']:
user_records['SessionID'] = sessionid
prevactivity = ''
user_records['probarry'] = []
user_records['prevclusterid'] = -1
user_records['NoOfClusterHopping'] = 0
user_records['pageclicks'] = 1
else:
prevactivity = user_records['Activity']
user_records['Activity'] = currentactivity
pageswitch = True
if prevactivity == currentactivity or prevactivity == '':
probability = 0
pageswitch = False
remarks = ''
else:
user_records['pageclicks'] += 1
df1 = dfprod[(dfprod['State'] == prevactivity) & (dfprod['NextState'] == currentactivity)]
if df1.empty:
remarks = 'Anomaly Detected - User in unusual state'
anomaly = True
clusterid = -1
probability = 0
user_records['probarry'].append(probability)
n=int(chain_count)
num_list = user_records['probarry'][-n:]
avg = sum(num_list)/len(num_list)
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
else:
probability = df1['Probability'].iloc[0]
user_records['probarry'].append(probability)
n=int(chain_count)
num_list = user_records['probarry'][-n:]
davg = sum(num_list)/len(num_list)
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
remarks = ''
if user_records['prevclusterid'] != -1:
if probability == 0 and user_records['prevclusterid'] != clusterid:
user_records['NoOfClusterHopping'] = user_records['NoOfClusterHopping']+1
if user_records['pageclicks'] == 1:
remarks = 'Anomaly Detected - Frequent Cluster Hopping'
anomaly = True
else:
remarks = 'Cluster Hopping Detected'
user_records['pageclicks'] = 0
if user_records['NoOfClusterHopping'] > int(groupswitching) and anomaly == False:
remarks = 'Anomaly Detected - Multiple Cluster Hopping'
anomaly = True
elif probability == 0:
remarks = 'Anomaly Detected - Unusual State Transition Detected'
anomaly = True
elif probability <= float(page_threshold):
remarks = 'Anomaly Detected - In-frequent State Transition Detected'
anomaly = True
else:
#print(pageswitch)
#print(probability)
if pageswitch == True:
if probability == 0:
remarks = 'Anomaly Detected - Unusual State Transition Detected'
anomaly = True
elif probability <= float(page_threshold):
remarks = 'Anomaly Detected - In-frequent State Transition Detected'
anomaly = True
else:
remarks = ''
if davg < float(chain_probability):
if anomaly == False:
remarks = 'Anomaly Detected - In-frequent Pattern Detected'
anomaly = True
else:
user_records['SessionID'] = data[sessionid]
user_records['Activity'] = data[activity]
user_records['probability'] = 0
user_records['probarry'] = []
user_records['chainprobability'] = 0
user_records['prevclusterid'] = -1
user_records['NoOfClusterHopping'] = 0
user_records['pageclicks'] = 1
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
user_records['prevclusterid'] = clusterid
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"status":"SUCCESS","data":{"Anomaly":"'+str(anomaly)+'","Remarks":"'+str(remarks)+'"}}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
else:
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"Anomaly":"Error","Remarks":"'+str(Int)+'"}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
elif None != re.search('/AION/pattern_anomaly_settings', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
data = self.rfile.read(length)
#print(data)
#keyList = list(data.keys())
#print(keyList[0])
model = self.path.split('/')[-1]
#print(model)
data = json.loads(data)
#dataStr = json.dumps(data)
groupswitching = data['groupswitching']
transitionprobability = data['transitionprobability']
transitionsequence = data['transitionsequence']
sequencethreshold = data['sequencethreshold']
home = expanduser("~")
if platform.system() == 'Windows':
filename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'clickstream.json')
else:
filename = os.path.join(home,'HCLT','AION','target',model,'clickstream.json')
#print(filename)
data = {}
data['groupswitching'] = groupswitching
data['transitionprobability'] = transitionprobability
data['transitionsequence'] = transitionsequence
data['sequencethreshold'] = sequencethreshold
updatedConfig = json.dumps(data)
with open(filename, "w") as fpWrite:
fpWrite.write(updatedConfig)
fpWrite.close()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"Status":"SUCCESS"}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
else:
print("python ==> else1")
self.send_response(403)
self.send_header('Content-Type', 'application/json')
self.end_headers()
print("PYTHON ######## REQUEST ####### ENDED")
return
def do_GET(self):
print("PYTHON ######## REQUEST ####### STARTED")
if None != re.search('/AION/predict', self.path):
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
model = self.path.split('/')[-1]
display_path = os.path.join(os.path.dirname(os.path.abspath(__file__)),'target',model,'display.json')
displaymsg = 'Data in JSON Format'
if(os.path.isfile(display_path)):
with open(display_path) as file:
config = json.load(file)
file.close()
features = config['modelFeatures']
datajson={}
for feature in features:
datajson[feature] = 'Value'
displaymsg = json.dumps(datajson)
msg="""
URL:{url}
RequestType: POST
Content-Type=application/json
Body: {displaymsg}
""".format(url=self.path,displaymsg=displaymsg)
self.wfile.write(msg.encode())
else:
self.send_response(403)
self.send_header('Content-Type', 'application/json')
self.end_headers()
return
class ThreadedHTTPServer(ThreadingMixIn, HTTPServer):
allow_reuse_address = True
def shutdown(self):
self.socket.close()
HTTPServer.shutdown(self)
class SimpleHttpServer():
def __init__(self, ip, port):
self.server = ThreadedHTTPServer((ip,port), HTTPRequestHandler)
def start(self):
self.server_thread = threading.Thread(target=self.server.serve_forever)
self.server_thread.daemon = True
self.server_thread.start()
def waitForThread(self):
self.server_thread.join()
def stop(self):
self.server.shutdown()
self.waitForThread()
if __name__=='__main__':
parser = argparse.ArgumentParser(description='HTTP Server')
parser.add_argument('port', type=int, help='Listening port for HTTP Server')
parser.add_argument('ip', help='HTTP Server IP')
args = parser.parse_args()
server = SimpleHttpServer(args.ip, args.port)
#delete file
#create file
#write just msg as "started"
print('HTTP Server Running...........')
#file close
server.start()
server.waitForThread()
|
runService.py | #from BaseHTTPServer import BaseHTTPRequestHandler,HTTPServer
from http.server import BaseHTTPRequestHandler,HTTPServer
#from SocketServer import ThreadingMixIn
from socketserver import ThreadingMixIn
'''
from augustus.core.ModelLoader import ModelLoader
from augustus.strict import modelLoader
'''
import pandas as pd
from datetime import datetime
import os,sys
from os.path import expanduser
import platform
import numpy as np
import configparser
import threading
import subprocess
import argparse
import re
import cgi
import time
from datetime import datetime
import json
import sys
from datetime import datetime
import sqlite3
from os.path import expanduser
from pathlib import Path
from io import BytesIO
DEPLOY_DATABASE_PATH = os.path.join(os.path.join(os.path.dirname(__file__)),'database')
targetPath = Path(DEPLOY_DATABASE_PATH)
targetPath.mkdir(parents=True, exist_ok=True)
modelVersion = 'run_1'
version = 1
class sqlite_db():
def __init__(self, location, database_file=None):
if not isinstance(location, Path):
location = Path(location)
if database_file:
self.database_name = database_file
else:
self.database_name = location.stem + '.db'
db_file = str(location/self.database_name)
self.conn = sqlite3.connect(db_file)
self.cursor = self.conn.cursor()
self.tables = []
def table_exists(self, name):
if name in self.tables:
return True
elif name:
query = f"SELECT name FROM sqlite_master WHERE type='table' AND name='{name}';"
listOfTables = self.cursor.execute(query).fetchall()
if len(listOfTables) > 0 :
self.tables.append(name)
return True
return False
def read(self, table_name,condition=''):
if condition == '':
return pd.read_sql_query(f"SELECT * FROM {table_name}", self.conn)
else:
return pd.read_sql_query(f"SELECT * FROM {table_name} WHERE {condition}", self.conn)
def create_table(self,name, columns, dtypes):
query = f'CREATE TABLE IF NOT EXISTS {name} ('
for column, data_type in zip(columns, dtypes):
query += f"'{column}' TEXT,"
query = query[:-1]
query += ');'
self.conn.execute(query)
return True
def update(self,table_name,updates,condition):
update_query = f'UPDATE {table_name} SET {updates} WHERE {condition}'
self.cursor.execute(update_query)
self.conn.commit()
return True
def write(self,data, table_name):
if not self.table_exists(table_name):
self.create_table(table_name, data.columns, data.dtypes)
tuple_data = list(data.itertuples(index=False, name=None))
insert_query = f'INSERT INTO {table_name} VALUES('
for i in range(len(data.columns)):
insert_query += '?,'
insert_query = insert_query[:-1] + ')'
self.cursor.executemany(insert_query, tuple_data)
self.conn.commit()
return True
def delete(self, name):
pass
def close(self):
self.conn.close()
user_records = {}
class LocalModelData(object):
models = {}
class HTTPRequestHandler(BaseHTTPRequestHandler):
def do_POST(self):
print("PYTHON ######## REQUEST ####### STARTED")
if None != re.search('/AION/', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
#data = cgi.parse_qs(self.rfile.read(length), keep_blank_values=1)
data = self.rfile.read(length)
model = self.path.split('/')[-2]
operation = self.path.split('/')[-1]
#data = json.loads(data)
#dataStr = json.dumps(data)
home = expanduser("~")
dataStr = data
sqlite_dbObj = sqlite_db(DEPLOY_DATABASE_PATH,'deploy.db')
model_path = os.path.join(os.path.dirname(__file__),modelVersion)
DATA_FILE_PATH = os.path.join(os.path.dirname(__file__),'temp')
Path(DATA_FILE_PATH).mkdir(parents=True, exist_ok=True)
isdir = os.path.isdir(model_path)
if isdir:
if operation.lower() == 'predict':
if not sqlite_dbObj.table_exists('servingDetails'):
data = {'usecase':model,'noOfPredictCalls':0,'noOfDriftCalls':0}
data = pd.DataFrame(data, index=[0])
sqlite_dbObj.create_table('servingDetails',data.columns, data.dtypes)
df2 = pd.read_json(BytesIO(dataStr), orient ='records')
if not sqlite_dbObj.table_exists('prodData'):
sqlite_dbObj.create_table('prodData',df2.columns, df2.dtypes)
sqlite_dbObj.write(df2,'prodData')
data = sqlite_dbObj.read('servingDetails',"usecase = '"+model+"'")
if len(data) == 0:
data = {'usecase':model,'noOfPredictCalls':1,'noOfDriftCalls':0}
data = pd.DataFrame(data, index=[0])
sqlite_dbObj.write(data,'servingDetails')
else:
noofPredictCalls = int(data['noOfPredictCalls'].iloc[0])+1
sqlite_dbObj.update('servingDetails',"noOfPredictCalls = '"+str(noofPredictCalls)+"'","usecase = '"+model+"'")
predict_path = os.path.join(model_path,'aion_predict.py')
outputStr = subprocess.check_output([sys.executable,predict_path,dataStr])
outputStr = outputStr.decode('utf-8')
outputStr = re.search(r'predictions:(.*)',str(outputStr), re.IGNORECASE).group(1)
outputStr = outputStr.strip()
resp = outputStr
elif operation.lower() == 'monitoring':
if not sqlite_dbObj.table_exists('monitoring'):
data = {'usecase':model,'status':'No Drift','Msg':'No Input Drift Found','RecordTime':'Time','version':'1'}
data = pd.DataFrame(data, index=[0])
sqlite_dbObj.create_table('monitoring',data.columns, data.dtypes)
trainingDataPath = os.path.join(os.path.dirname(__file__),modelVersion,'data','preprocesseddata.csv')
data = sqlite_dbObj.read('prodData')
filetimestamp = str(int(time.time()))
dataFile = os.path.join(DATA_FILE_PATH, 'AION_' + filetimestamp+'.csv')
data.to_csv(dataFile, index=False)
predict_path = os.path.join(model_path,'aion_ipdrift.py')
inputJSON={'trainingDataLocation':trainingDataPath,'currentDataLocation':dataFile}
outputStr = subprocess.check_output([sys.executable,predict_path,json.dumps(inputJSON)])
outputStr = outputStr.decode('utf-8')
outputStr = re.search(r'drift:(.*)',str(outputStr), re.IGNORECASE).group(1)
outputStr = outputStr.strip()
outputData = json.loads(outputStr)
status = outputData['status']
if status == 'SUCCESS':
Msg = str(outputData['data'])
else:
Msg = 'Error during drift analysis'
now = datetime.now() # current date and time
date_time = now.strftime("%m/%d/%Y, %H:%M:%S")
data = {'usecase':model,'status':status,'Msg':Msg,'RecordTime':date_time,'version':version}
data = pd.DataFrame(data, index=[0])
sqlite_dbObj.write(data,'monitoring')
resp = outputStr
resp=resp+"\n"
resp=resp.encode()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
elif None != re.search('/AION/pattern_anomaly_predict', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
model = self.path.split('/')[-1]
data = self.rfile.read(length)
data = json.loads(data)
anomaly = False
remarks = ''
clusterid = -1
home = expanduser("~")
if platform.system() == 'Windows':
configfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'datadetails.json')
filename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'clickstream.json')
clusterfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'stateClustering.csv')
probfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'stateTransitionProbability.csv')
else:
configfilename = os.path.join(home,'HCLT','AION','target',model,'datadetails.json')
filename = os.path.join(home,'HCLT','AION','target',model,'clickstream.json')
clusterfilename = os.path.join(home,'HCLT','AION','target',model,'stateClustering.csv')
probfilename = os.path.join(home,'HCLT','AION','target',model,'stateTransitionProbability.csv')
dfclus = pd.read_csv(clusterfilename)
dfprod = pd.read_csv(probfilename)
f = open(configfilename, "r")
configSettings = f.read()
f.close()
configSettingsJson = json.loads(configSettings)
activity = configSettingsJson['activity']
sessionid = configSettingsJson['sessionid']
f = open(filename, "r")
configSettings = f.read()
f.close()
configSettingsJson = json.loads(configSettings)
groupswitching = configSettingsJson['groupswitching']
page_threshold = configSettingsJson['transitionprobability']
chain_count = configSettingsJson['transitionsequence']
chain_probability = configSettingsJson['sequencethreshold']
currentactivity = data[activity]
if bool(user_records):
sessionid = data[sessionid]
print(sessionid,user_records['SessionID'])
if sessionid != user_records['SessionID']:
user_records['SessionID'] = sessionid
prevactivity = ''
user_records['probarry'] = []
user_records['prevclusterid'] = -1
user_records['NoOfClusterHopping'] = 0
user_records['pageclicks'] = 1
else:
prevactivity = user_records['Activity']
user_records['Activity'] = currentactivity
pageswitch = True
if prevactivity == currentactivity or prevactivity == '':
probability = 0
pageswitch = False
remarks = ''
else:
user_records['pageclicks'] += 1
df1 = dfprod[(dfprod['State'] == prevactivity) & (dfprod['NextState'] == currentactivity)]
if df1.empty:
remarks = 'Anomaly Detected - User in unusual state'
anomaly = True
clusterid = -1
probability = 0
user_records['probarry'].append(probability)
n=int(chain_count)
num_list = user_records['probarry'][-n:]
avg = sum(num_list)/len(num_list)
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
else:
probability = df1['Probability'].iloc[0]
user_records['probarry'].append(probability)
n=int(chain_count)
num_list = user_records['probarry'][-n:]
davg = sum(num_list)/len(num_list)
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
remarks = ''
if user_records['prevclusterid'] != -1:
if probability == 0 and user_records['prevclusterid'] != clusterid:
user_records['NoOfClusterHopping'] = user_records['NoOfClusterHopping']+1
if user_records['pageclicks'] == 1:
remarks = 'Anomaly Detected - Frequent Cluster Hopping'
anomaly = True
else:
remarks = 'Cluster Hopping Detected'
user_records['pageclicks'] = 0
if user_records['NoOfClusterHopping'] > int(groupswitching) and anomaly == False:
remarks = 'Anomaly Detected - Multiple Cluster Hopping'
anomaly = True
elif probability == 0:
remarks = 'Anomaly Detected - Unusual State Transition Detected'
anomaly = True
elif probability <= float(page_threshold):
remarks = 'Anomaly Detected - In-frequent State Transition Detected'
anomaly = True
else:
#print(pageswitch)
#print(probability)
if pageswitch == True:
if probability == 0:
remarks = 'Anomaly Detected - Unusual State Transition Detected'
anomaly = True
elif probability <= float(page_threshold):
remarks = 'Anomaly Detected - In-frequent State Transition Detected'
anomaly = True
else:
remarks = ''
if davg < float(chain_probability):
if anomaly == False:
remarks = 'Anomaly Detected - In-frequent Pattern Detected'
anomaly = True
else:
user_records['SessionID'] = data[sessionid]
user_records['Activity'] = data[activity]
user_records['probability'] = 0
user_records['probarry'] = []
user_records['chainprobability'] = 0
user_records['prevclusterid'] = -1
user_records['NoOfClusterHopping'] = 0
user_records['pageclicks'] = 1
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
user_records['prevclusterid'] = clusterid
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"status":"SUCCESS","data":{"Anomaly":"'+str(anomaly)+'","Remarks":"'+str(remarks)+'"}}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
else:
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"Anomaly":"Error","Remarks":"'+str(Int)+'"}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
elif None != re.search('/AION/pattern_anomaly_settings', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
data = self.rfile.read(length)
#print(data)
#keyList = list(data.keys())
#print(keyList[0])
model = self.path.split('/')[-1]
#print(model)
data = json.loads(data)
#dataStr = json.dumps(data)
groupswitching = data['groupswitching']
transitionprobability = data['transitionprobability']
transitionsequence = data['transitionsequence']
sequencethreshold = data['sequencethreshold']
home = expanduser("~")
if platform.system() == 'Windows':
filename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'clickstream.json')
else:
filename = os.path.join(home,'HCLT','AION','target',model,'clickstream.json')
#print(filename)
data = {}
data['groupswitching'] = groupswitching
data['transitionprobability'] = transitionprobability
data['transitionsequence'] = transitionsequence
data['sequencethreshold'] = sequencethreshold
updatedConfig = json.dumps(data)
with open(filename, "w") as fpWrite:
fpWrite.write(updatedConfig)
fpWrite.close()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"Status":"SUCCESS"}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
else:
print("python ==> else1")
self.send_response(403)
self.send_header('Content-Type', 'application/json')
self.end_headers()
print("PYTHON ######## REQUEST ####### ENDED")
return
def do_GET(self):
print("PYTHON ######## REQUEST ####### STARTED")
if None != re.search('/AION/', self.path) or None != re.search('/aion/', self.path):
usecase = self.path.split('/')[-2]
operation = self.path.split('/')[-1]
if operation.lower() == 'metrices':
sqlite_dbObj = sqlite_db(DEPLOY_DATABASE_PATH,'deploy.db')
if sqlite_dbObj.table_exists('servingDetails'):
df1 = sqlite_dbObj.read('servingDetails',"usecase = '"+usecase+"'")
else:
df1 = pd.DataFrame()
if sqlite_dbObj.table_exists('monitoring'):
df2 = sqlite_dbObj.read('monitoring')
else:
df2 = pd.DataFrame()
if sqlite_dbObj.table_exists('modeldetails'):
df3 = sqlite_dbObj.read('modeldetails')
else:
df3 = pd.DataFrame()
msg='<html>\n'
msg+='<head>\n'
msg+='<title>Model Metrices</title>\n'
msg+='</head>\n'
msg+="""<style>table, th, td { border: 1px solid black; border-collapse: collapse;}</style>"""
msg+='<body>\n'
msg+='<h2>Model Metrices - Deployed Version '+str(version)+'</h2>'
msg+='<br/>\n'
msg+='<table style="width:80%">\n'
msg+="""<tr>
<th>Model</th>
<th>Version</th>
<th>ScoreType</th>
<th>Score</th>
</tr
"""
for idx in reversed(df3.index):
msg += "<tr>\n"
msg += "<td>"+str(df3.usecase[idx])+"</td>\n"
msg += "<td>"+str(df3.version[idx])+"</td>\n"
msg += "<td>"+str(df3.scoreType[idx])+"</td>\n"
msg += "<td>"+str(df3.score[idx])+"</td>\n"
msg += "</tr>\n"
msg += '</table>\n'
msg += '<br/>\n'
msg += '<br/>\n'
msg+='<table style="width:50%">\n'
msg+='<tr>\n'
msg+='<td>No of Predictions</td>\n'
if df1.shape[0] > 0:
msg+='<td>'+str(df1['noOfPredictCalls'].iloc[0])+'</td>\n'
else:
msg+='<td>0</td>\n'
msg+='</tr>\n'
msg+='<tr>\n'
msg+='<td>No of Ground Truth</td>\n'
msg+='<td>0</td>\n'
msg+='</tr>\n'
msg += '</table>\n'
msg += '<br/>\n'
msg+='<table style="width:100%">\n'
msg+="""<tr>
<th>UseCase</th>
<th>Version</th>
<th>Status</th>
<th>Message</th>
<th>Time</th>
</tr>
"""
for idx in reversed(df2.index):
msg += "<tr>\n"
msg += "<td>"+str(df2.usecase[idx])+"</td>\n"
msg += "<td>"+str(df3.version[idx])+"</td>\n"
msg += "<td>"+str(df2.status[idx])+"</td>\n"
msg += "<td>"+str(df2.Msg[idx])+"</td>\n"
msg += "<td>"+str(df2.RecordTime[idx])+"</td>\n"
msg += "</tr>\n"
msg += '</table>\n'
msg += '</body>\n'
msg += '</html>\n'
self.send_response(200)
self.send_response(200)
self.send_header('Content-Type', 'text/html')
self.end_headers()
self.wfile.write(msg.encode())
else:
self.send_response(403)
self.send_header('Content-Type', 'application/json')
self.end_headers()
return
class ThreadedHTTPServer(ThreadingMixIn, HTTPServer):
allow_reuse_address = True
def shutdown(self):
self.socket.close()
HTTPServer.shutdown(self)
class file_status():
def __init__(self,file):
self.files_status = {}
self.initializeFileStatus(file)
def initializeFileStatus(self, file):
self.files_status = {'path': file, 'time':file.stat().st_mtime}
def is_file_changed(self):
if self.files_status['path'].stat().st_mtime > self.files_status['time']:
self.files_status['time'] = self.files_status['path'].stat().st_mtime
return True
return False
def run(self):
while( True):
time.sleep(30)
if self.is_file_changed():
readRun()
class SimpleHttpServer():
def __init__(self, ip, port,model_file_path):
self.server = ThreadedHTTPServer((ip,port), HTTPRequestHandler)
self.status_checker = file_status(model_file_path)
def start(self):
self.server_thread = threading.Thread(target=self.server.serve_forever)
self.server_thread.daemon = True
self.server_thread.start()
self.status_thread = threading.Thread(target=self.status_checker.run)
self.status_thread.start()
def waitForThread(self):
self.server_thread.join()
def stop(self):
self.server.shutdown()
self.waitForThread()
def readRun(boot=False):
filename = os.path.join(os.path.join(os.path.dirname(__file__)),'run')
f = open (filename, "r")
data = json.loads(f.read())
global modelVersion
global version
modelVersion = 'run_'+str(data['version'])
version = str(data['version'])
sqlite_dbObj = sqlite_db(DEPLOY_DATABASE_PATH,'deploy.db')
dataa = {'usecase':data['usecase'],'version':data['version'],'scoreType':data['scoreType'],'score':data['score']}
data = pd.DataFrame(dataa, index=[0])
if not sqlite_dbObj.table_exists('modeldetails'):
sqlite_dbObj.create_table('modeldetails',data.columns, data.dtypes)
rdata = sqlite_dbObj.read('modeldetails',"version = '"+dataa['version']+"'")
if (rdata.shape[0]) == 0 or (not boot):
sqlite_dbObj.write(data,'modeldetails')
readRun(boot=True)
if __name__=='__main__':
filename = os.path.join(os.path.join(os.path.dirname(__file__)),'run')
parser = argparse.ArgumentParser(description='HTTP Server')
parser.add_argument('port', type=int, help='Listening port for HTTP Server')
parser.add_argument('ip', help='HTTP Server IP')
args = parser.parse_args()
server = SimpleHttpServer(args.ip, args.port,Path(filename))
#delete file
#create file
#write just msg as "started"
print('HTTP Server Running...........')
#file close
server.start()
server.waitForThread()
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
model_service.py | '''
from AION_185 import aion_prediction
from AION_185 import featureslist
from AION_185 import aion_drift
from AION_185 import aion_performance
'''
#from BaseHTTPServer import BaseHTTPRequestHandler,HTTPServer
from http.server import BaseHTTPRequestHandler,HTTPServer
#from SocketServer import ThreadingMixIn
from socketserver import ThreadingMixIn
'''
from augustus.core.ModelLoader import ModelLoader
from augustus.strict import modelLoader
'''
import pandas as pd
import os,sys
from os.path import expanduser
import platform
import numpy as np
import configparser
import threading
import subprocess
import argparse
import re
import cgi
from datetime import datetime
import json
import sys
from datetime import datetime
user_records = {}
class LocalModelData(object):
models = {}
class HTTPRequestHandler(BaseHTTPRequestHandler):
def do_POST(self):
print("PYTHON ######## REQUEST ####### STARTED")
if None != re.search('/AION/predict', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
#data = cgi.parse_qs(self.rfile.read(length), keep_blank_values=1)
data = self.rfile.read(length)
data = json.loads(data)
dataStr = json.dumps(data)
outputStr = aion_prediction.predict(dataStr)
outputStr = outputStr.strip()
resp = outputStr
resp=resp+"\n"
resp=resp.encode()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
elif None != re.search('/AION/features', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
outputStr = featureslist.getfeatures()
outputStr = outputStr.strip()
resp = outputStr
resp=resp+"\n"
resp=resp.encode()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
elif None != re.search('/AION/monitoring', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
data = self.rfile.read(length)
model = self.path.split('/')[-1]
data = json.loads(data)
dataStr = json.dumps(data)
outputStr = aion_drift.drift(dataStr)
outputStr = outputStr.strip()
resp = outputStr
resp=resp+"\n"
resp=resp.encode()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
elif None != re.search('/AION/performance', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
data = self.rfile.read(length)
data = json.loads(data)
dataStr = json.dumps(data)
outputStr = aion_performance.drift(dataStr)
outputStr = outputStr.strip()
resp = outputStr
resp=resp+"\n"
resp=resp.encode()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
elif None != re.search('/AION/pattern_anomaly_predict', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
model = self.path.split('/')[-1]
data = self.rfile.read(length)
data = json.loads(data)
anomaly = False
remarks = ''
clusterid = -1
home = expanduser("~")
if platform.system() == 'Windows':
configfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'datadetails.json')
filename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'clickstream.json')
clusterfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'stateClustering.csv')
probfilename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'stateTransitionProbability.csv')
else:
configfilename = os.path.join(home,'HCLT','AION','target',model,'datadetails.json')
filename = os.path.join(home,'HCLT','AION','target',model,'clickstream.json')
clusterfilename = os.path.join(home,'HCLT','AION','target',model,'stateClustering.csv')
probfilename = os.path.join(home,'HCLT','AION','target',model,'stateTransitionProbability.csv')
dfclus = pd.read_csv(clusterfilename)
dfprod = pd.read_csv(probfilename)
f = open(configfilename, "r")
configSettings = f.read()
f.close()
configSettingsJson = json.loads(configSettings)
activity = configSettingsJson['activity']
sessionid = configSettingsJson['sessionid']
f = open(filename, "r")
configSettings = f.read()
f.close()
configSettingsJson = json.loads(configSettings)
groupswitching = configSettingsJson['groupswitching']
page_threshold = configSettingsJson['transitionprobability']
chain_count = configSettingsJson['transitionsequence']
chain_probability = configSettingsJson['sequencethreshold']
currentactivity = data[activity]
if bool(user_records):
sessionid = data[sessionid]
print(sessionid,user_records['SessionID'])
if sessionid != user_records['SessionID']:
user_records['SessionID'] = sessionid
prevactivity = ''
user_records['probarry'] = []
user_records['prevclusterid'] = -1
user_records['NoOfClusterHopping'] = 0
user_records['pageclicks'] = 1
else:
prevactivity = user_records['Activity']
user_records['Activity'] = currentactivity
pageswitch = True
if prevactivity == currentactivity or prevactivity == '':
probability = 0
pageswitch = False
remarks = ''
else:
user_records['pageclicks'] += 1
df1 = dfprod[(dfprod['State'] == prevactivity) & (dfprod['NextState'] == currentactivity)]
if df1.empty:
remarks = 'Anomaly Detected - User in unusual state'
anomaly = True
clusterid = -1
probability = 0
user_records['probarry'].append(probability)
n=int(chain_count)
num_list = user_records['probarry'][-n:]
avg = sum(num_list)/len(num_list)
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
else:
probability = df1['Probability'].iloc[0]
user_records['probarry'].append(probability)
n=int(chain_count)
num_list = user_records['probarry'][-n:]
davg = sum(num_list)/len(num_list)
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
remarks = ''
if user_records['prevclusterid'] != -1:
if probability == 0 and user_records['prevclusterid'] != clusterid:
user_records['NoOfClusterHopping'] = user_records['NoOfClusterHopping']+1
if user_records['pageclicks'] == 1:
remarks = 'Anomaly Detected - Frequent Cluster Hopping'
anomaly = True
else:
remarks = 'Cluster Hopping Detected'
user_records['pageclicks'] = 0
if user_records['NoOfClusterHopping'] > int(groupswitching) and anomaly == False:
remarks = 'Anomaly Detected - Multiple Cluster Hopping'
anomaly = True
elif probability == 0:
remarks = 'Anomaly Detected - Unusual State Transition Detected'
anomaly = True
elif probability <= float(page_threshold):
remarks = 'Anomaly Detected - In-frequent State Transition Detected'
anomaly = True
else:
#print(pageswitch)
#print(probability)
if pageswitch == True:
if probability == 0:
remarks = 'Anomaly Detected - Unusual State Transition Detected'
anomaly = True
elif probability <= float(page_threshold):
remarks = 'Anomaly Detected - In-frequent State Transition Detected'
anomaly = True
else:
remarks = ''
if davg < float(chain_probability):
if anomaly == False:
remarks = 'Anomaly Detected - In-frequent Pattern Detected'
anomaly = True
else:
user_records['SessionID'] = data[sessionid]
user_records['Activity'] = data[activity]
user_records['probability'] = 0
user_records['probarry'] = []
user_records['chainprobability'] = 0
user_records['prevclusterid'] = -1
user_records['NoOfClusterHopping'] = 0
user_records['pageclicks'] = 1
for index, row in dfclus.iterrows():
clusterlist = row["clusterlist"]
if currentactivity in clusterlist:
clusterid = row["clusterid"]
user_records['prevclusterid'] = clusterid
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"status":"SUCCESS","data":{"Anomaly":"'+str(anomaly)+'","Remarks":"'+str(remarks)+'"}}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
else:
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"Anomaly":"Error","Remarks":"'+str(Int)+'"}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
elif None != re.search('/AION/pattern_anomaly_settings', self.path):
ctype, pdict = cgi.parse_header(self.headers.get('content-type'))
if ctype == 'application/json':
length = int(self.headers.get('content-length'))
data = self.rfile.read(length)
#print(data)
#keyList = list(data.keys())
#print(keyList[0])
model = self.path.split('/')[-1]
#print(model)
data = json.loads(data)
#dataStr = json.dumps(data)
groupswitching = data['groupswitching']
transitionprobability = data['transitionprobability']
transitionsequence = data['transitionsequence']
sequencethreshold = data['sequencethreshold']
home = expanduser("~")
if platform.system() == 'Windows':
filename = os.path.join(home,'AppData','Local','HCLT','AION','target',model,'clickstream.json')
else:
filename = os.path.join(home,'HCLT','AION','target',model,'clickstream.json')
#print(filename)
data = {}
data['groupswitching'] = groupswitching
data['transitionprobability'] = transitionprobability
data['transitionsequence'] = transitionsequence
data['sequencethreshold'] = sequencethreshold
updatedConfig = json.dumps(data)
with open(filename, "w") as fpWrite:
fpWrite.write(updatedConfig)
fpWrite.close()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
resp = '{"Status":"SUCCESS"}'
resp=resp+"\n"
resp=resp.encode()
self.wfile.write(resp)
else:
print("python ==> else2")
data = {}
else:
print("python ==> else1")
self.send_response(403)
self.send_header('Content-Type', 'application/json')
self.end_headers()
print("PYTHON ######## REQUEST ####### ENDED")
return
def do_GET(self):
print("PYTHON ######## REQUEST ####### STARTED")
if None != re.search('/AION/predict', self.path):
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
features = featureslist.getfeatures()
displaymsg = 'Data in JSON Format'
config = json.loads(features)
features = config['features']
datajson={}
for feature in features:
if feature['Type'].lower() != 'target':
datajson[feature['feature']] = 'Value'
displaymsg = json.dumps(datajson)
msg="""
URL:{url}
RequestType: POST
Content-Type=application/json
Body: {displaymsg}
""".format(url=self.path,displaymsg=displaymsg)
self.wfile.write(msg.encode())
elif None != re.search('/AION/monitoring', self.path):
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
displaymsg='{"trainingDataLocation":"Reference Data File Path","currentDataLocation":"Latest Data File Path"}'
msg="""
URL:{url}
RequestType: POST
Content-Type=application/json
Body: {displaymsg}""".format(url=self.path,displaymsg=displaymsg)
self.wfile.write(msg.encode())
elif None != re.search('/AION/features', self.path):
outputStr = featureslist.getfeatures()
outputStr = outputStr.strip()
resp = outputStr
resp=resp+"\n"
resp=resp.encode()
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(resp)
else:
msg="""
URL for prediction: /AION/predict
URL for features List: /AION/features
URL for monitoring: /AION/monitoring
URL for performance: /AION/performance"""
self.send_response(200)
self.send_header('Content-Type', 'application/json')
self.end_headers()
self.wfile.write(msg.encode())
return
class ThreadedHTTPServer(ThreadingMixIn, HTTPServer):
allow_reuse_address = True
def shutdown(self):
self.socket.close()
HTTPServer.shutdown(self)
class SimpleHttpServer():
def __init__(self, ip, port):
self.server = ThreadedHTTPServer((ip,port), HTTPRequestHandler)
def start(self):
self.server_thread = threading.Thread(target=self.server.serve_forever)
self.server_thread.daemon = True
self.server_thread.start()
def waitForThread(self):
self.server_thread.join()
def stop(self):
self.server.shutdown()
self.waitForThread()
if __name__=='__main__':
parser = argparse.ArgumentParser(description='HTTP Server')
parser.add_argument('port', type=int, help='Listening port for HTTP Server')
parser.add_argument('ip', help='HTTP Server IP')
args = parser.parse_args()
server = SimpleHttpServer(args.ip, args.port)
#delete file
#create file
#write just msg as "started"
print('HTTP Server Running...........')
#file close
server.start()
server.waitForThread()
|
dl_aion_predict.py | import sys
import os
import pickle
import json
import traceback
import warnings
warnings.filterwarnings("ignore")
import numpy as np
import pandas as pd
import dask.dataframe as dd
import scipy
from pandas import json_normalize
import dask.distributed
from dask_ml.wrappers import ParallelPostFit
class incBatchPredictor():
def __init__(self):
self.home = os.path.dirname(os.path.abspath(__file__))
self.configPath = os.path.join(self.home, 'etc', 'config.json')
self.configDict = {}
self.incProfilerPath = ''
self.incSelectorPath = ''
self.modelPath = ''
self.incProfiler = None
self.incSelector = None
self.model = None
self.targetFeature = None
self.trainingFeatures = None
self.modelName = ''
self.problemType = ''
self.algorithm = ''
self.dataLocation = ""
self.nworkers = None
self.npartitions = None
self.threadsPerWorker = None
def get_nworkers(self):
return self.nworkers
def get_npartitions(self):
return self.npartitions
def get_threadsPerWorker(self):
return self.threadsPerWorker
def readData(self,data):
try:
if os.path.splitext(data)[1] in [".tsv",".csv",".data"]:
df = dd.read_csv(data, # sep=r'\s*,\s*',
assume_missing=True,
parse_dates=True, infer_datetime_format=True,
sample=1000000,
# dtype={'caliper': 'object',
# 'timestamp': 'object'},
# dtype='object',
na_values=['-','?']
)
df = df.repartition(self.npartitions)
else:
if os.path.splitext(data)[1] == ".json":
with open(data,'r',encoding='utf-8') as f:
jsonData = json.load(f)
else:
jsonData = json.loads(data)
for key in jsonData:
if type(jsonData[key]) == str:
try:
x = eval(jsonData[key])
if type(x) == int:
jsonData[key] = int(jsonData[key])
print("check inside ==int")
if type(x) == float:
jsonData[key] = float(jsonData[key])
except:
pass
df = json_normalize(jsonData)
df = df.replace(r'^\s*$', np.NaN, regex=True)
df = dd.from_pandas(df, npartitions=self.npartitions)
except KeyError as e:
output = {"status":"FAIL","message":str(e).strip('"')}
print(json.dumps(output))
except Exception as e:
output = {"status":"FAIL","message":str(e).strip('"')}
print(json.dumps(output))
return df
def readConfig(self):
with open(self.configPath, 'r', encoding= 'utf8') as f:
self.configDict = json.load(f)
self.targetFeature = self.configDict['targetFeature']
self.trainingFeatures = self.configDict['trainingFeatures']
self.modelName = self.configDict["modelName"]
self.modelVersion = self.configDict["modelVersion"]
self.dataLocation = self.configDict["dataLocation"]
self.deployLocation = self.configDict["deployLocation"]
self.incProfilerPath = self.configDict["profilerLocation"]
self.incSelectorPath = self.configDict["selectorLocation"]
self.problemType = self.configDict['analysisType']
self.algorithm = self.configDict["algorithm"]
self.modelPath = self.configDict["modelLocation"]
self.scoringCriteria = self.configDict['scoringCriteria']
self.nworkers = int(self.configDict["n_workers"])
self.npartitions = int(self.configDict["npartitions"])
self.threadsPerWorker = int(self.configDict["threads_per_worker"])
def pickleLoad(self, file):
if os.path.exists(file):
with open(file, 'rb') as f:
model = pickle.load(f)
return model
else:
return None
def loadSavedModels(self):
self.incProfiler = self.pickleLoad(os.path.join(self.home, 'model',self.incProfilerPath))
if self.incSelectorPath != '':
self.incSelector = self.pickleLoad(os.path.join(self.home, 'model',self.incSelectorPath))
self.model = self.pickleLoad(os.path.join(self.home, 'model',self.modelPath))
def dataFramePreProcess(self, df):
df = df.replace(r'^\s*$', np.NaN, regex=True)
df = df.replace('-', np.nan)
df = df.replace('?', np.nan)
return df
def profiler(self, df):
X = self.dataFramePreProcess(df)
if 'self.targetFeature' in X:
X = X.drop(self.targetFeature, axis=1)
X = self.incProfiler.transform(X)
if self.incSelectorPath != '':
X = self.incSelector.transform(X.to_dask_array(lengths=True))
# X = dd.from_dask_array(X)
return X
def trainedModel(self,X):
ParallelPostFit(estimator=self.model)
# preds = self.model.predict(X)
if self.algorithm=="Distributed Light Gradient Boosting (LightGBM)":
X = X.to_dask_array(lengths=True)
preds = self.model.predict(X).compute()
return preds
def apply_output_format(self,df,modeloutput):
label_maping = None
if self.problemType.lower() == 'regression':
if not isinstance(modeloutput, np.ndarray):
modeloutput = modeloutput.to_numpy()
dask_arr = dd.from_array(modeloutput)
dask_arr.name = 'prediction'
df = df.merge(dask_arr.to_frame())
df['rounded_prediction'] = df['prediction'].round(2)
elif self.problemType.lower() == 'classification':
print("type: ", type(modeloutput))
if not isinstance(modeloutput, np.ndarray):
modeloutput = modeloutput.to_numpy()
dask_arr = dd.from_array(modeloutput)
dask_arr.name = "prediction"
df = df.merge(dask_arr.to_frame())
with open(self.deployLocation + "/etc/" + "label_mapping.json") as jsonfile:
label_maping = json.load(jsonfile)
df["prediction"] = df["prediction"].astype(int)
df["prediction"] = df["prediction"].astype(str)
df["prediction_label"] = df["prediction"].map(label_maping)
if df["prediction_label"].dtype == None:
df["prediction_label"] = df["prediction"]
outputjson = df.compute().to_json(orient='records')
outputjson = {"status":"SUCCESS","data":json.loads(outputjson)}
return(json.dumps(outputjson))
def predict(self,data):
try:
self.readConfig()
df = self.readData(data)
dfOrg = df.copy()
if len(self.configDict)!=0:
self.loadSavedModels()
df = self.profiler(df)
modeloutput = self.trainedModel(df)
# dfOrg = dfOrg[self.allFtrs]
output = self.apply_output_format(dfOrg, modeloutput)
else:
pass
except Exception as e:
print(traceback.format_exc())
output = {"status":"FAIL","message":str(e).strip('"')}
return output
if __name__ == "__main__":
incBPobj = incBatchPredictor()
incBPobj.readConfig()
nWorkers = incBPobj.get_nworkers()
threads_per_worker = incBPobj.get_threadsPerWorker()
cluster = dask.distributed.LocalCluster(n_workers=nWorkers,
threads_per_worker=threads_per_worker)
client = dask.distributed.Client(cluster)
output = incBPobj.predict(sys.argv[1])
print("predictions:",output)
client.close()
cluster.close() |
heBinary.py | # -*- coding: utf-8 -*-
"""
Created on Sat Sep 10 23:57:56 2022
@author: jayaramakrishnans
"""
import numpy as np
import pandas as pd
from secrets import token_bytes
from ppxgboost import PaillierAPI as paillier
from ppxgboost import BoosterParser as boostparser
from ppxgboost import PPBooster as ppbooster
from ppxgboost.PPBooster import MetaData
from ppxgboost.PPKey import PPBoostKey
# from ope.pyope.ope import OPE
from pyope.ope import OPE
import joblib
import logging
import os
from flask import Flask,request,jsonify,render_template
# import pickle
from flask import Flask, request, jsonify
import json
import jsonpickle
app = Flask(__name__)
class server_ppxgboost:
def __init__(self):
# self.problemtype=problemtype
self.confdata=None
print("Inside server_ppxgboost_1\n")
## Loading config file
def configload(self):
print("Inside server_ppxgboost_1,configload\n")
cwd = os.path.abspath(os.path.dirname(__file__))
file_name='config.json'
try:
config_file=os.path.normpath(os.path.join(cwd,'config',file_name))
except Exception as e:
print("config path error. Error Msg: \n",e)
with open(config_file, 'r') as file:
data = json.load(file)
model_name=str(data["model_name"])
# version=str(data["version"])
return model_name
## Load server xgboost model from ../model dir.
def model_load( self,path):
print("Inside server_ppxgboost_1,model_load\n")
cwd = os.path.abspath(os.path.dirname(__file__))
file_name=path
try:
model_location=os.path.normpath(os.path.join(cwd,'model',file_name))
except Exception as e:
print("Model path error. Error Msg: \n",e)
# print(path)
loaded_model = joblib.load(model_location)
return loaded_model
## Generate Encrypted prediction fn
def ppxgboostpredict_server(self,model,ppBoostKey,clientdata,min_max):
xgboost_model = model
meta_min_max = MetaData(min_max)
p_trees, feature_set, min_max = boostparser.model_to_trees(xgboost_model, min_max)
enc_trees = ppbooster.enc_xgboost_model(ppBoostKey, p_trees, MetaData(min_max))
enc_client_data=clientdata
# enc_predictions = ppbooster.predict_binary(enc_trees, X_test) # should rename the function
enc_predictions = ppbooster.predict_binary(p_trees, enc_client_data)
return enc_predictions
## XGBoost wrapper for native model (native model to scikit learn xgboost model)
def xgboostwrappertonative(self,wrappermodel):
print("Inside server_ppxgboost_7,xgboostwrappertonative= \n",wrappermodel)
nativemodel = wrappermodel.get_booster()
return nativemodel
def training_dataset_parser(self,train_data: pd.DataFrame):
"""
:param train_data: dataframe training data
:return: minimum of the training dataset, and maximum of the training dataset.
"""
return {'min': np.min(pd.DataFrame.min(train_data)), 'max': np.max(pd.DataFrame.max(train_data))}
## Homomorphic secure main server
cls_obj=server_ppxgboost()
@app.route('/homomorphicprediction_server_api',methods=['GET','POST'])
def main_server():
data = request.get_json(force=True)
response_data = json.dumps(data)
json_in= json.loads(response_data)
values = json_in['values']
features=json_in['features']
ppBoostKey=jsonpickle.decode(json_in['key'])
encrypted_clientdata=pd.DataFrame(values,columns =features)
## Create encrypted predition from model
model=None
min_max = {'min': 0, 'max': 1000}
model_name = cls_obj.configload()
# model_name=usecase_name
model_location=model_name
model_xgboost = cls_obj.model_load(model_location)
try:
## For sklearn based xgboost model to native model
model = cls_obj.xgboostwrappertonative(model_xgboost)
except:
## For native xgboost,we dont need to get booster.
model= model_xgboost
## FFor logging
cwd = os.path.abspath(os.path.dirname(__file__))
# model_name=model_name
file_name = model_name.rsplit('.', 1)
file_name=file_name[0]
file_name=file_name+".log"
try:
hm_log=os.path.normpath(os.path.join(cwd,'logs',file_name))
os.makedirs(os.path.dirname(hm_log), exist_ok=True)
except Exception as e:
print("Log path error. Error Msg: \n",e)
logging.basicConfig(filename=hm_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
log = logging.getLogger('AION')
log.info('Server binary class encryption service started...')
print("Encrypted client data is ....\n")
log.info("Encrypted client data is (received by server): \n"+str(encrypted_clientdata))
enc_predictions = cls_obj.ppxgboostpredict_server(model,ppBoostKey,encrypted_clientdata,min_max)
log.info("server side encrypted prediction: \n"+str(enc_predictions))
## Serialize the ppboost encrypted prediction by jsonpickle, normal pikle lib not working.
enc_predictions_json = jsonpickle.encode(enc_predictions)
# enc_predictions_json = enc_predictions.to_json()
return enc_predictions_json
if __name__ == '__main__':
#app.run(debug=True)
app.run(host="localhost", port=9000, debug=True) |
heRegression.py | # -*- coding: utf-8 -*-
"""
Created on Sat Sep 10 23:57:56 2022
"""
import numpy as np
import pandas as pd
from secrets import token_bytes
from ppxgboost import PaillierAPI as paillier
from ppxgboost import BoosterParser as boostparser
from ppxgboost import PPBooster as ppbooster
from ppxgboost.PPBooster import MetaData
from ppxgboost.PPKey import PPBoostKey
# from ope.pyope.ope import OPE
from pyope.ope import OPE
import joblib
import logging
import os
from flask import Flask,request,jsonify,render_template
# import pickle
from flask import Flask, request, jsonify
import json
import jsonpickle
import os.path
import time
import subprocess
import sys
from os.path import expanduser
import ntpath
import shutil
import platform
from pathlib import Path
home = expanduser("~")
if platform.system() == 'Windows':
LOG_FILE_PATH = os.path.join(home,'AppData','Local','HCLT','AION','logs')
else:
LOG_FILE_PATH = os.path.join(home,'HCLT','AION','logs')
app = Flask(__name__)
class server_ppxgboost:
def __init__(self):
# self.problemtype=problemtype
self.confdata=None
## Loading config file
def configload(self):
cwd = os.path.abspath(os.path.dirname(__file__))
file_name='config.json'
try:
config_file=os.path.normpath(os.path.join(cwd,'config',file_name))
except Exception as e:
print("config path error. Error Msg: \n",e)
with open(config_file, 'r') as file:
data = json.load(file)
model_name=str(data["model_name"])
# version=str(data["version"])
return model_name
## Load server xgboost model from ../model dir.
def model_load( self,path):
cwd = os.path.abspath(os.path.dirname(__file__))
file_name=path
try:
model_location=os.path.normpath(os.path.join(cwd,'model',file_name))
except Exception as e:
print("Model path error. Error Msg: \n",e)
# print(path)
loaded_model = joblib.load(model_location)
return loaded_model
## Generate Encrypted prediction fn
def ppxgboostpredict_server(self,model,ppBoostKey,clientdata,min_max):
xgboost_model = model
meta_min_max = MetaData(min_max)
p_trees, feature_set, min_max = boostparser.model_to_trees(xgboost_model, min_max)
enc_trees = ppbooster.enc_xgboost_model(ppBoostKey, p_trees, MetaData(min_max))
enc_client_data=clientdata
# enc_predictions = ppbooster.predict_binary(enc_trees, X_test) # should rename the function
enc_predictions = ppbooster.predict_binary(p_trees, enc_client_data)
return enc_predictions
## XGBoost wrapper for native model (native model to scikit learn xgboost model)
def xgboostwrappertonative(self,wrappermodel):
nativemodel = wrappermodel.get_booster()
return nativemodel
def training_dataset_parser(self,train_data: pd.DataFrame):
"""
:param train_data: dataframe training data
:return: minimum of the training dataset, and maximum of the training dataset.
"""
return {'min': np.min(pd.DataFrame.min(train_data)), 'max': np.max(pd.DataFrame.max(train_data))}
## Homomorphic secure main server
cls_obj=server_ppxgboost()
@app.route('/homomorphicprediction_server_api',methods=['GET','POST'])
def main_server():
data = request.get_json(force=True)
response_data = json.dumps(data)
json_in= json.loads(response_data)
values = json_in['values']
features=json_in['features']
ppBoostKey=jsonpickle.decode(json_in['key'])
encrypted_clientdata=pd.DataFrame(values,columns =features)
## Create encrypted predition from model
model=None
min_max = {'min': 0, 'max': 1000}
model_name = cls_obj.configload()
# model_name=usecase_name
model_location=model_name
model_xgboost = cls_obj.model_load(model_location)
try:
## For sklearn based xgboost model to native model
model = cls_obj.xgboostwrappertonative(model_xgboost)
except:
## For native xgboost,we dont need to get booster.
model= model_xgboost
## FFor logging
cwd = os.path.abspath(os.path.dirname(__file__))
# model_name=model_name
file_name = model_name.rsplit('.', 1)
file_name=file_name[0]
file_name=file_name+".log"
try:
hm_log=os.path.normpath(os.path.join(cwd,'logs',file_name))
os.makedirs(os.path.dirname(hm_log), exist_ok=True)
except Exception as e:
print("Log path error. Error Msg: \n",e)
logging.basicConfig(filename=hm_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
log = logging.getLogger('AION')
log.info('Server regression encryption service started...')
print("Encrypted client data is ....\n")
log.info("Encrypted client data is (received by server): \n"+str(encrypted_clientdata))
print("Client side encrypted data: \n",encrypted_clientdata)
enc_predictions = cls_obj.ppxgboostpredict_server(model,ppBoostKey,encrypted_clientdata,min_max)
log.info("server side encrypted prediction: \n"+str(enc_predictions))
## Serialize the ppboost encrypted prediction by jsonpickle, normal pikle lib not working.
enc_predictions_json = jsonpickle.encode(enc_predictions)
# enc_predictions_json = enc_predictions.to_json()
return enc_predictions_json
if __name__ == '__main__':
#app.run(debug=True)
app.run(host="localhost", port=9000, debug=True) |
heMulticlass.py | # -*- coding: utf-8 -*-
"""
Created on Sat Sep 10 23:57:56 2022
"""
import numpy as np
import sqlite3
import sys
import pandas as pd
from secrets import token_bytes
from ppxgboost import PaillierAPI as paillier
from ppxgboost import BoosterParser as boostparser
from ppxgboost import PPBooster as ppbooster
from ppxgboost.PPBooster import MetaData
from ppxgboost.PPKey import PPBoostKey
# from ope.pyope.ope import OPE
from pyope.ope import OPE
import joblib
import logging
import os
from flask import Flask,request,jsonify,render_template
# import pickle
from flask import Flask, request, jsonify
import json
import jsonpickle
import time
from pathlib import Path
app = Flask(__name__)
class server_ppxgboost:
def __init__(self):
# self.problemtype=problemtype
print("Inside server_ppxgboost_1,init\n")
self.confdata=None
## Loading config file
def configload(self):
cwd = os.path.abspath(os.path.dirname(__file__))
file_name='secure_config.json'
try:
config_file=os.path.normpath(os.path.join(cwd,'etc',file_name))
except Exception as e:
print("config path error. Error Msg: \n",e)
with open(config_file, 'r') as file:
data = json.load(file)
model_name=str(data["model_name"])
# version=str(data["version"])
return model_name
## Load server xgboost model from ../model dir.
def model_load( self,path):
cwd = os.path.abspath(os.path.dirname(__file__))
file_name=path
try:
model_location=os.path.normpath(os.path.join(cwd,'model',file_name))
except Exception as e:
print("Model path error. Error Msg: \n",e)
# print(path)
loaded_model = joblib.load(model_location)
return loaded_model
def create_connection(self,db_file):
conn = None
try:
conn = sqlite3.connect(db_file)
conn.execute('''CREATE TABLE IF NOT EXISTS modelinfo (key BLOB NOT NULL,encrypttree BLOB NOT NULL)''')
except Exception as e:
print(e)
return conn
def count_encrypt_model(self,conn):
try:
sql = "select count(*) from modelinfo"
cur = conn.cursor()
cur.execute(sql)
cur_result = cur.fetchone()
return cur_result[0]
except Exception as e:
print(e)
def create_encryptmodel(self,conn,modeldetails):
sql = ''' INSERT INTO modelinfo(key,encrypttree) VALUES(?,?) '''
cur = conn.cursor()
cur.execute(sql, modeldetails)
conn.commit()
return cur.lastrowid
def search_encryptmodel(self,conn,key):
try:
sql = "SELECT encrypttree FROM modelinfo where key='"+key+"'"
cursor = conn.execute(sql)
for row in cursor:
return row[0]
return ''
except Exception as e:
print(e)
def ppxgboostpredict_server(self,model,ppBoostKey,clientdata,min_max,noofclasses):
try:
db_file = Path(__file__).parent/'logs'/'encryptedModels'
conn = self.create_connection(db_file)
enc_trees = self.search_encryptmodel(conn,jsonpickle.encode(ppBoostKey))
if enc_trees != '':
enc_trees = jsonpickle.decode(enc_trees)
else:
if self.count_encrypt_model(conn) >= 5:
outputjson = {"status":"ERROR","msg":"Maximum Number of Encrypted"}
return json.dumps(outputjson)
xgboost_model = model
meta_min_max = MetaData(min_max)
p_trees, feature_set, min_max = boostparser.model_to_trees(xgboost_model, min_max)
enc_trees = ppbooster.enc_xgboost_model(ppBoostKey, p_trees, MetaData(min_max))
modelinfo = (jsonpickle.encode(ppBoostKey),jsonpickle.encode(enc_trees))
self.create_encryptmodel(conn,modelinfo)
enc_client_data=clientdata
# try:
# num_classes = model.n_classes_
# except:
# num_classes = noofclasses
num_classes = noofclasses
if num_classes == 0:
nc_predictions = ppbooster.predict_binary(p_trees, enc_client_data)
else:
enc_predictions = ppbooster.predict_multiclass(enc_trees, num_classes, enc_client_data)
enc_predictions_json = jsonpickle.encode(enc_predictions)
outputjson = {"status":"SUCCESS","data":enc_predictions_json}
return json.dumps(outputjson)
except Exception as e:
outputjson = {"status":"ERROR","msg":str(e)}
return json.dumps(outputjson)
## XGBoost wrapper for native model (native model to scikit learn xgboost model)
def xgboostwrappertonative(self,wrappermodel):
nativemodel = wrappermodel.get_booster()
try:
noOfClasses = wrappermodel.n_classes_
except Exception as e:
print(e)
noOfClasses = 0
return nativemodel,noOfClasses
def training_dataset_parser(self,train_data: pd.DataFrame):
"""
:param train_data: dataframe training data
:return: minimum of the training dataset, and maximum of the training dataset.
"""
return {'min': np.min(pd.DataFrame.min(train_data)), 'max': np.max(pd.DataFrame.max(train_data))}
## Homomorphic secure main server
cls_obj=server_ppxgboost()
def spredict(data):
try:
json_in= json.loads(data)
values = json_in['values']
features=json_in['features']
ppBoostKey=jsonpickle.decode(json_in['key'])
encrypted_clientdata=pd.DataFrame(values,columns =features)
model=None
min_max = {'min': 0, 'max': 1000}
model_name = cls_obj.configload()
model_location=model_name
model_xgboost = cls_obj.model_load(model_location)
try:
model,noofclasses = cls_obj.xgboostwrappertonative(model_xgboost)
except Exception as e:
print(e)
model= model_xgboost
noofclasses = 0
cwd = os.path.abspath(os.path.dirname(__file__))
# model_name=model_name
file_name = model_name.rsplit('.', 1)
file_name=file_name[0]
file_name=file_name+".log"
try:
hm_log=os.path.normpath(os.path.join(cwd,'logs',file_name))
os.makedirs(os.path.dirname(hm_log), exist_ok=True)
except Exception as e:
print("Log path error. Error Msg: \n",e)
logging.basicConfig(filename=hm_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
log = logging.getLogger('AION')
log.info('Server multiclass classification encryption service started...')
log.info("Encrypted client data is (received by server): \n"+str(encrypted_clientdata))
output = cls_obj.ppxgboostpredict_server(model,ppBoostKey,encrypted_clientdata,min_max,noofclasses)
print("predictions:",output)
print("Inside server_ppxgboost_8,output= \n",output)
return output
except Exception as e:
outputjson = {"status":"ERROR","msg":str(e)}
output = json.dumps(outputjson)
print("predictions:",output)
return output
if __name__ == "__main__":
output = spredict(sys.argv[1]) |
aion_hebinaryclient.py | # -*- coding: utf-8 -*-
# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import logging
import os
import sys
from logging import INFO
from script.heBinary import client_ppxgboost
from script.aion_predict import selector
from script.inputprofiler import inputprofiler
## Client main class for binary classification
class aion_hebinaryclient:
def __init__(self):
self.confdata=None
def configLoad(self,jsonfile):
import json
jsonfile=str(jsonfile)
with open(jsonfile, 'r') as file:
self.confdata = json.load(file)
return self.confdata
def dataload(self,datapath):
df = pd.read_csv(datapath)
## Data preprocess in test dataset, In aion, aion profiler will handle it.
# df =df[~df.isin([np.nan, np.inf, -np.inf]).any(1)]
# df=df.reset_index(drop=True)
profilerobj = inputprofiler()
df = profilerobj.run(df)
selectobj = selector()
df = selectobj.run(df)
return df
# Start Flower server for n rounds of federated learning
if __name__ == "__main__":
classobj=aion_hebinaryclient()
data_location=str(sys.argv[1])
df=classobj.dataload(data_location)
# print("df: \n",df)
aionhmcobj=client_ppxgboost(df)
ppxgboost_pred=aionhmcobj.main_client()
print("final decrypted prediction at client side: \n",ppxgboost_pred)
# classobj.log.info("At client end, homomorphic prediction df: \n"+str(ppxgboost_pred))
# classobj.log.info("Aion homomorphic client encrypted prediction df: \n"+str(ppxgboost_pred))
|
aion_hemulticlient.py | # -*- coding: utf-8 -*-
# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import logging
import os
import sys
from logging import INFO
from script.heMulticlass import client_ppxgboost
from script.aion_predict import selector
from script.inputprofiler import inputprofiler
import argparse
class aion_hemulticlient:
def __init__(self):
self.confdata=None
def dataload(self,datapath):
df = pd.read_csv(datapath)
## Data preprocess in test dataset, In aion, aion profiler will handle it.
# df =df[~df.isin([np.nan, np.inf, -np.inf]).any(1)]
# df=df.reset_index(drop=True)
profilerobj = inputprofiler()
df = profilerobj.run(df)
selectobj = selector()
df = selectobj.run(df)
return df
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('-i', '--inputDataLocation', help='Input Data Path')
parser.add_argument('-k', '--keyGenerate', help='True')
parser.add_argument('-e', '--endPoint', help='Service End Point')
args = parser.parse_args()
if args.inputDataLocation:
dataLocation=args.inputDataLocation
if args.keyGenerate:
keyGenerate = args.keyGenerate
else:
keyGenerate='False'
print(keyGenerate)
if args.endPoint:
endPoint=args.endPoint
else:
raise('End Point Not Defined')
classobj=aion_hemulticlient()
df=classobj.dataload(dataLocation)
aionhmcobj=client_ppxgboost(df,keyGenerate,endPoint)
ppxgboost_pred=aionhmcobj.main_client()
print("final decrypted prediction at client side: \n",ppxgboost_pred)
|
aion_heregressionclient.py | # -*- coding: utf-8 -*-
# -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import logging
import os
import sys
from logging import INFO
from script.heRegression import client_ppxgboost
from script.aion_predict import selector
from script.inputprofiler import inputprofiler
import argparse
class aion_hemulticlient:
def __init__(self):
self.confdata=None
def dataload(self,datapath):
df = pd.read_csv(datapath)
## Data preprocess in test dataset, In aion, aion profiler will handle it.
# df =df[~df.isin([np.nan, np.inf, -np.inf]).any(1)]
# df=df.reset_index(drop=True)
profilerobj = inputprofiler()
df = profilerobj.run(df)
selectobj = selector()
df = selectobj.run(df)
return df
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('-i', '--inputDataLocation', help='Input Data Path')
parser.add_argument('-k', '--keyGenerate', help='True')
parser.add_argument('-e', '--endPoint', help='Service End Point')
args = parser.parse_args()
if args.inputDataLocation:
dataLocation=args.inputDataLocation
if args.keyGenerate:
keyGenerate = args.keyGenerate
else:
keyGenerate='False'
print(keyGenerate)
if args.endPoint:
endPoint=args.endPoint
else:
raise('End Point Not Defined')
classobj=aion_hemulticlient()
df=classobj.dataload(dataLocation)
aionhmcobj=client_ppxgboost(df,keyGenerate,endPoint)
ppxgboost_pred=aionhmcobj.main_client()
print("final decrypted prediction at client side: \n",ppxgboost_pred)
|
heBinary.py | # -*- coding: utf-8 -*-
import pandas as pd
from sklearn.model_selection import train_test_split
import numpy as np
from secrets import token_bytes
from ppxgboost import PaillierAPI as paillier
from ppxgboost import BoosterParser as boostparser
from ppxgboost import PPBooster as ppbooster
from ppxgboost.PPBooster import MetaData
from ppxgboost.PPKey import PPBoostKey
# from ope.pyope.ope import OPE
from pyope.ope import OPE
import sys
sys.path.insert(0, '..')
import logging
from logging import INFO
import pickle
import requests
import json
# from json import JSONEncoder
import jsonpickle
import os
##Aion main client class for ppxgboost based encryption,decryption
class client_ppxgboost:
def __init__(self,data):
self.data=data
self.prediction=None
## For logging
cwd = os.path.abspath(os.path.dirname(__file__))
# model_name=model_name
file_name = "he_binaryclass"
file_name=file_name+".log"
try:
hm_log=os.path.normpath(os.path.join(cwd,'logs',file_name))
os.makedirs(os.path.dirname(hm_log), exist_ok=True)
except Exception as e:
print("Log path error. Error Msg: \n",e)
logging.basicConfig(filename=hm_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
self.log = logging.getLogger('AION')
self.log.info('Client binary class classification homomorphic encryption service started...')
## Loading configuration parameters, Not used now.
def configload(self):
try:
data=self.confdata
usecase_name=str(data["usecase_name"])
version=str(data["version"])
problem_type=data["problem_type"]
model_location=str(data["model_location"])
data_location=str(data["data_location"])
selected_feature=data["selected_feature"]
if (type(selected_feature) is str):
selected_feature=selected_feature.split(',')
target_feature=data["target_feature"]
client_encryption_accuracy=str(data["client_encryption_accuracy"])
test_size=int(data["test_size"])
test_size=test_size/100
except Exception as e:
self.log.info("Reading server config file issue. Err.Msg: %s "+str(e))
return usecase_name,data_location,model_location,problem_type,version,selected_feature,target_feature,client_encryption_accuracy,test_size
## Load the model, Not used at client now.
def model_load(self, path):
loaded_model = pickle.load(open(path, 'rb'))
return loaded_model
#Generating secure key
def generate_ppboostkey(self):
public_key, private_key = paillier.he_key_gen()
prf_key = token_bytes(16)
OPE_key = token_bytes(16)
encrypter = OPE(OPE_key)
ppBoostKey = PPBoostKey(public_key, prf_key, encrypter)
return public_key,private_key,prf_key,OPE_key,encrypter,ppBoostKey
## Binary client prediction (decrypted prediction)
def ppxgboostbinary_predict(self,enc_predictions,private_key):
dec = ppbooster.client_decrypt(private_key, enc_predictions)
##For binary classification
res = ppbooster.client_decrypt_prediction_binary(private_key, enc_predictions)
res_df=pd.DataFrame(res)
return res
# class ppkeyEncoder(JSONEncoder):
# def default(self,o):
# return o.__dict__
## Function to connect secure server via flask restapi (send enc data and receive enc prediction.)
def connect_xgboostserver(self,ppBoostKey,encrypted_xtest):
url = 'http://localhost:9000//homomorphicprediction_server_api'
enc_dict={}
# df_list=[encrypted_xtest.columns.values.tolist()]+df.values.tolist()
enc_dict['values']=encrypted_xtest.values.tolist()
enc_dict['features']=encrypted_xtest.columns.values.tolist()
enc_dict['key']= jsonpickle.encode(ppBoostKey)
json_out=json.dumps(enc_dict,indent=4)
headers = {
'content-type': "application/json",
'cache-control': "no-cache"
}
r = requests.post(url,data=json_out,headers=headers)
enc_predictions_obj=jsonpickle.decode(r.content)
return enc_predictions_obj
## Create PaillierAPI based encrypted user given data , here, testdata=userdata
def generate_encrypted_testdata(self,prf_key,encrypter,testdata,min_max):
feature_set_testdata=set(testdata.columns)
ppbooster.enc_input_vector(prf_key, encrypter, feature_set_testdata, testdata, MetaData(min_max))
return testdata
## Create min and max of testdata df for pailler encryption,decryption
def training_dataset_parser(self, client_data: pd.DataFrame):
"""
:param client_data: dataframe training data
:return: minimum of the training dataset, and maximum of the training dataset.
"""
return {'min': np.min(pd.DataFrame.min(client_data)), 'max': np.max(pd.DataFrame.max(client_data))}
## Main client function call for enc data, send data to server, receive enc pred, finally decrypt prediction
def main_client(self):
self.log.info('Client actual data sample (displaying last 10 values) : \n'+str(self.data.tail(10)))
print(" Client actual data sample (displaying last 10 values) : \n",self.data.tail(10))
public_key,private_key,prf_key,OPE_key,encrypter,ppBoostKey = self.generate_ppboostkey()
min_max = self.training_dataset_parser(self.data)
meta_min_max = MetaData(min_max)
encrypted_testdata = self.generate_encrypted_testdata(prf_key,encrypter,self.data,min_max)
# print("Sending encrypted client data to server....\n")
print("\n Client side encrypted input data to server (displaying last 10 rows): \n",encrypted_testdata.tail(10))
self.log.info('Client side encrypted input data to server (displaying last 10 rows): \n'+str(encrypted_testdata.tail(10)))
enc_predictions = self.connect_xgboostserver(ppBoostKey,encrypted_testdata)
print("\n Encrypted prediction from server (displaying last 10 values.): \n",enc_predictions[-10:])
self.log.info('\n Encrypted prediction from server (displaying last 10 values.): \n'+str(enc_predictions[-10:]))
## Decrypted predition
dec = self.ppxgboostbinary_predict(enc_predictions,private_key)
# ppxgboost_pred=pd.DataFrame(list(zip(dec, predictions)),columns =['homomorphic_prediction', 'actual_prediction'])
ppxgboost_pred=pd.DataFrame(dec,columns =['homomorphic_prediction'])
self.log.info("final decrypted prediction at client side:: \n"+str(ppxgboost_pred))
return ppxgboost_pred
## For standalone testing
if __name__ == '__main__':
problemtype='Binary classification'
data=None
targetfeature=None
ppxgboost_client_obj=client_ppxgboost(problemtype,data,targetfeature)
ppxgboost_dec_predictions = ppxgboost_client_obj.main_client()
# print("In main: ppxgboost_dec_predictions: \n",ppxgboost_dec_predictions) |
heRegression.py | # -*- coding: utf-8 -*-
import pandas as pd
from sklearn.model_selection import train_test_split
import numpy as np
from secrets import token_bytes
from ppxgboost import PaillierAPI as paillier
from ppxgboost import BoosterParser as boostparser
from ppxgboost import PPBooster as ppbooster
from ppxgboost.PPBooster import MetaData
from ppxgboost.PPKey import PPBoostKey
# from ope.pyope.ope import OPE
from pyope.ope import OPE
import sys
sys.path.insert(0, '..')
import logging
from logging import INFO
import pickle
import requests
import json
# from json import JSONEncoder
import jsonpickle
import os
from pathlib import Path
##Aion main client class for ppxgboost based encryption,decryption
class client_ppxgboost:
def __init__(self,data,keyGenerate,endPoint):
self.data=data
self.keyGenerate = keyGenerate
self.endPoint = endPoint
self.prediction=None
## For logging
clientDirectory = os.path.abspath(os.path.dirname(__file__))
# model_name=model_name
file_name = "he_regression"
file_name=file_name+".log"
self.keydir=os.path.join(clientDirectory,'..','keys')
os.makedirs(self.keydir, exist_ok=True)
try:
hm_log=os.path.normpath(os.path.join(clientDirectory,'logs',file_name))
os.makedirs(os.path.dirname(hm_log), exist_ok=True)
except Exception as e:
print("Log path error. Error Msg: \n",e)
logging.basicConfig(filename=hm_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
self.log = logging.getLogger('AION')
self.log.info('Client Regression homomorphic encryption service started...')
## Loading configuration parameters, Not used now.
def configload(self):
try:
data=self.confdata
usecase_name=str(data["usecase_name"])
version=str(data["version"])
problem_type=data["problem_type"]
model_location=str(data["model_location"])
data_location=str(data["data_location"])
selected_feature=data["selected_feature"]
if (type(selected_feature) is str):
selected_feature=selected_feature.split(',')
target_feature=data["target_feature"]
client_encryption_accuracy=str(data["client_encryption_accuracy"])
test_size=int(data["test_size"])
test_size=test_size/100
except Exception as e:
self.log.info("Reading server config file issue. Err.Msg: %s "+str(e))
return usecase_name,data_location,model_location,problem_type,version,selected_feature,target_feature,client_encryption_accuracy,test_size
## Load the model, Not used at client now.
def model_load(self, path):
loaded_model = pickle.load(open(path, 'rb'))
return loaded_model
#Generating secure key
def generate_ppboostkey(self):
try:
public_key_file = Path(__file__).parent.parent/'keys'/'public.k'
private_key_file = Path(__file__).parent.parent/'keys'/'private.k'
prf_key_file = Path(__file__).parent.parent/'keys'/'prf.k'
ope_key_file = Path(__file__).parent.parent/'keys'/'ope.k'
encryptor_key_file = Path(__file__).parent.parent/'keys'/'encryptor.k'
boostkey_key_file = Path(__file__).parent.parent/'keys'/'boostkey.k'
if not boostkey_key_file.exists() or self.keyGenerate == 'True':
public_key, private_key = paillier.he_key_gen()
pub_file = open(public_key_file, 'w')
pub_file.write(jsonpickle.encode(public_key))
pri_file = open(private_key_file, 'w')
pri_file.write(jsonpickle.encode(private_key))
prf_key = token_bytes(16)
OPE_key = token_bytes(16)
prf_file = open(prf_key_file, 'w')
prf_file.write(jsonpickle.encode(prf_key))
ope_file = open(ope_key_file, 'w')
ope_file.write(jsonpickle.encode(OPE_key))
encrypter = OPE(OPE_key)
enc_file = open(encryptor_key_file, 'w')
enc_file.write(jsonpickle.encode(encrypter))
ppBoostKey = PPBoostKey(public_key, prf_key, encrypter)
boost_file = open(boostkey_key_file, 'w')
boost_file.write(jsonpickle.encode(ppBoostKey))
else:
pub_file = open(public_key_file, 'r')
public_key = jsonpickle.decode(pub_file.read())
pub_file.close()
pri_file = open(private_key_file, 'r')
private_key = jsonpickle.decode(pri_file.read())
pri_file.close()
prf_file = open(prf_key_file, 'r')
prf_key = jsonpickle.decode(prf_file.read())
prf_file.close()
ope_file = open(ope_key_file, 'r')
OPE_key = jsonpickle.decode(ope_file.read())
ope_file.close()
enc_file = open(encryptor_key_file, 'r')
encrypter = jsonpickle.decode(enc_file.read())
enc_file.close()
boost_file = open(boostkey_key_file, 'r')
ppBoostKey = jsonpickle.decode(boost_file.read())
boost_file.close()
return public_key,private_key,prf_key,OPE_key,encrypter,ppBoostKey
except Exception as e:
print(e)
def ppxgboostreg_predict(self,enc_predictions,private_key):
dec = []
for p in enc_predictions:
dec.append(paillier.decrypt(private_key, p))
dec_df=pd.DataFrame(dec)
return dec
# class ppkeyEncoder(JSONEncoder):
# def default(self,o):
# return o.__dict__
## Function to connect secure server via flask restapi (send enc data and receive enc prediction.)
def connect_xgboostserver(self,ppBoostKey,encrypted_xtest):
url = self.endPoint
enc_dict={}
# df_list=[encrypted_xtest.columns.values.tolist()]+df.values.tolist()
enc_dict['values']=encrypted_xtest.values.tolist()
enc_dict['features']=encrypted_xtest.columns.values.tolist()
enc_dict['key']= jsonpickle.encode(ppBoostKey)
json_out=json.dumps(enc_dict,indent=4)
headers = {
'content-type': "application/json",
'cache-control': "no-cache"
}
response = requests.post(url,auth=('admin','aion'),data=json_out,headers=headers)
#print(response.content)
outputStr=response.content
outputStr = outputStr.decode('utf-8')
outputStr = outputStr.strip()
predict_dict = json.loads(str(outputStr))
if (predict_dict['status'] == 'SUCCESS'):
data = predict_dict['data']
enc_predictions_ob=jsonpickle.decode(data)
return enc_predictions_ob
else:
print('Error')
## Create PaillierAPI based encrypted user given data , here, testdata=userdata
def generate_encrypted_testdata(self,prf_key,encrypter,testdata,min_max):
feature_set_testdata=set(testdata.columns)
ppbooster.enc_input_vector(prf_key, encrypter, feature_set_testdata, testdata, MetaData(min_max))
return testdata
## Create min and max of testdata df for pailler encryption,decryption
def training_dataset_parser(self, client_data: pd.DataFrame):
"""
:param client_data: dataframe training data
:return: minimum of the training dataset, and maximum of the training dataset.
"""
return {'min': np.min(pd.DataFrame.min(client_data)), 'max': np.max(pd.DataFrame.max(client_data))}
## Main client function call for enc data, send data to server, receive enc pred, finally decrypt prediction
def main_client(self):
self.log.info('Client actual data sample (displaying last 10 values) : \n'+str(self.data.tail(10)))
print(" Client actual data sample (displaying last 10 values) : \n",self.data.tail(10))
public_key,private_key,prf_key,OPE_key,encrypter,ppBoostKey = self.generate_ppboostkey()
min_max = self.training_dataset_parser(self.data)
meta_min_max = MetaData(min_max)
encrypted_testdata = self.generate_encrypted_testdata(prf_key,encrypter,self.data,min_max)
# print("Sending encrypted client data to server....\n")
print("\n Client side encrypted input data to server (displaying last 10 rows): \n",encrypted_testdata.tail(10))
self.log.info('Client side encrypted input data to server (displaying last 10 rows): \n'+str(encrypted_testdata.tail(10)))
enc_predictions = self.connect_xgboostserver(ppBoostKey,encrypted_testdata)
print("\n Encrypted prediction from server (displaying last 10 values.): \n",enc_predictions[-10:])
self.log.info('\n Encrypted prediction from server (displaying last 10 values.): \n'+str(enc_predictions[-10:]))
## Decrypted predition
dec = self.ppxgboostreg_predict(enc_predictions,private_key)
# ppxgboost_pred=pd.DataFrame(list(zip(dec, predictions)),columns =['homomorphic_prediction', 'actual_prediction'])
ppxgboost_pred=pd.DataFrame(dec,columns =['homomorphic_prediction'])
# print("final decrypted prediction at client side: \n",ppxgboost_pred)
self.log.info("Final decrypted prediction at client side:: \n"+str(ppxgboost_pred))
return ppxgboost_pred
## For standalone testing
if __name__ == '__main__':
problemtype='regression'
data=None
targetfeature=None
ppxgboost_client_obj=client_ppxgboost(problemtype,data,targetfeature)
ppxgboost_dec_predictions = ppxgboost_client_obj.main_client()
print("In main: ppxgboost_dec_predictions: \n",ppxgboost_dec_predictions) |
heMulticlass.py | # -*- coding: utf-8 -*-
import pandas as pd
from sklearn.model_selection import train_test_split
import numpy as np
from secrets import token_bytes
from ppxgboost import PaillierAPI as paillier
from ppxgboost import BoosterParser as boostparser
from ppxgboost import PPBooster as ppbooster
from ppxgboost.PPBooster import MetaData
from ppxgboost.PPKey import PPBoostKey
# from ope.pyope.ope import OPE
from pyope.ope import OPE
import sys
sys.path.insert(0, '..')
import logging
from logging import INFO
import pickle
import requests
import json
# from json import JSONEncoder
import jsonpickle
import os
from pathlib import Path
##Aion main client class for ppxgboost based encryption,decryption
class client_ppxgboost:
def __init__(self,data,keyGenerate,endPoint):
self.data=data
self.keyGenerate = keyGenerate
self.endPoint = endPoint
self.prediction=None
## For logging
clientDirectory = os.path.abspath(os.path.dirname(__file__))
# model_name=model_name
file_name = "he_multiclass"
file_name=file_name+".log"
self.keydir=os.path.join(clientDirectory,'..','keys')
os.makedirs(self.keydir, exist_ok=True)
try:
hm_log=os.path.normpath(os.path.join(clientDirectory,'logs',file_name))
os.makedirs(os.path.dirname(hm_log), exist_ok=True)
except Exception as e:
print("Log path error. Error Msg: \n",e)
logging.basicConfig(filename=hm_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
self.log = logging.getLogger('AION')
self.log.info('Client Multi class classification homomorphic encryption service started...')
## Loading configuration parameters, Not used now.
def configload(self):
try:
data=self.confdata
usecase_name=str(data["usecase_name"])
version=str(data["version"])
problem_type=data["problem_type"]
model_location=str(data["model_location"])
data_location=str(data["data_location"])
selected_feature=data["selected_feature"]
if (type(selected_feature) is str):
selected_feature=selected_feature.split(',')
target_feature=data["target_feature"]
client_encryption_accuracy=str(data["client_encryption_accuracy"])
test_size=int(data["test_size"])
test_size=test_size/100
except Exception as e:
self.log.info("Reading server config file issue. Err.Msg: %s "+str(e))
return usecase_name,data_location,model_location,problem_type,version,selected_feature,target_feature,client_encryption_accuracy,test_size
## Load the model, Not used at client now.
def model_load(self, path):
loaded_model = pickle.load(open(path, 'rb'))
return loaded_model
#Generating secure key
def generate_ppboostkey(self):
try:
public_key_file = Path(__file__).parent.parent/'keys'/'public.k'
private_key_file = Path(__file__).parent.parent/'keys'/'private.k'
prf_key_file = Path(__file__).parent.parent/'keys'/'prf.k'
ope_key_file = Path(__file__).parent.parent/'keys'/'ope.k'
encryptor_key_file = Path(__file__).parent.parent/'keys'/'encryptor.k'
boostkey_key_file = Path(__file__).parent.parent/'keys'/'boostkey.k'
if not boostkey_key_file.exists() or self.keyGenerate == 'True':
public_key, private_key = paillier.he_key_gen()
pub_file = open(public_key_file, 'w')
pub_file.write(jsonpickle.encode(public_key))
pri_file = open(private_key_file, 'w')
pri_file.write(jsonpickle.encode(private_key))
prf_key = token_bytes(16)
OPE_key = token_bytes(16)
prf_file = open(prf_key_file, 'w')
prf_file.write(jsonpickle.encode(prf_key))
ope_file = open(ope_key_file, 'w')
ope_file.write(jsonpickle.encode(OPE_key))
encrypter = OPE(OPE_key)
enc_file = open(encryptor_key_file, 'w')
enc_file.write(jsonpickle.encode(encrypter))
ppBoostKey = PPBoostKey(public_key, prf_key, encrypter)
boost_file = open(boostkey_key_file, 'w')
boost_file.write(jsonpickle.encode(ppBoostKey))
else:
pub_file = open(public_key_file, 'r')
public_key = jsonpickle.decode(pub_file.read())
pub_file.close()
pri_file = open(private_key_file, 'r')
private_key = jsonpickle.decode(pri_file.read())
pri_file.close()
prf_file = open(prf_key_file, 'r')
prf_key = jsonpickle.decode(prf_file.read())
prf_file.close()
ope_file = open(ope_key_file, 'r')
OPE_key = jsonpickle.decode(ope_file.read())
ope_file.close()
enc_file = open(encryptor_key_file, 'r')
encrypter = jsonpickle.decode(enc_file.read())
enc_file.close()
boost_file = open(boostkey_key_file, 'r')
ppBoostKey = jsonpickle.decode(boost_file.read())
boost_file.close()
return public_key,private_key,prf_key,OPE_key,encrypter,ppBoostKey
except Exception as e:
print(e)
## PPboost multiclass prediction fn
def ppxgboostmulticlass_predict(self,enc_predictions,private_key):
##For binary classification
# res = ppbooster.client_decrypt_prediction_binary(private_key, enc_predictions)
## For multiclass classification
res = ppbooster.client_decrypt_prediction_multiclass(private_key, enc_predictions)
return res
# class ppkeyEncoder(JSONEncoder):
# def default(self,o):
# return o.__dict__
## Function to connect secure server via flask restapi (send enc data and receive enc prediction.)
def connect_xgboostserver(self,ppBoostKey,encrypted_xtest):
url = self.endPoint
enc_dict={}
# df_list=[encrypted_xtest.columns.values.tolist()]+df.values.tolist()
enc_dict['values']=encrypted_xtest.values.tolist()
enc_dict['features']=encrypted_xtest.columns.values.tolist()
enc_dict['key']= jsonpickle.encode(ppBoostKey)
json_out=json.dumps(enc_dict,indent=4)
headers = {
'content-type': "application/json",
'cache-control': "no-cache"
}
response = requests.post(url,auth=('admin','aion'),data=json_out,headers=headers)
#print(response.content)
outputStr=response.content
outputStr = outputStr.decode('utf-8')
outputStr = outputStr.strip()
predict_dict = json.loads(str(outputStr))
if (predict_dict['status'] == 'SUCCESS'):
data = predict_dict['data']
enc_predictions_ob=jsonpickle.decode(data)
return enc_predictions_ob
else:
print('Error')
## Create PaillierAPI based encrypted user given data , here, testdata=userdata
def generate_encrypted_testdata(self,prf_key,encrypter,testdata,min_max):
feature_set_testdata=set(testdata.columns)
ppbooster.enc_input_vector(prf_key, encrypter, feature_set_testdata, testdata, MetaData(min_max))
return testdata
## Create min and max of testdata df for pailler encryption,decryption
def training_dataset_parser(self, client_data: pd.DataFrame):
"""
:param client_data: dataframe training data
:return: minimum of the training dataset, and maximum of the training dataset.
"""
return {'min': np.min(pd.DataFrame.min(client_data)), 'max': np.max(pd.DataFrame.max(client_data))}
## Main client function call for enc data, send data to server, receive enc pred, finally decrypt prediction
def main_client(self):
self.log.info('Client actual data sample (displaying last 10 values) : \n'+str(self.data.tail(10)))
#print(" Client actual data sample (displaying last 10 values) : \n",self.data.tail(10))
public_key,private_key,prf_key,OPE_key,encrypter,ppBoostKey = self.generate_ppboostkey()
min_max = self.training_dataset_parser(self.data)
meta_min_max = MetaData(min_max)
#print('++++++++++++++++++++++++++++')
encrypted_testdata = self.generate_encrypted_testdata(prf_key,encrypter,self.data,min_max)
# print("Sending encrypted client data to server....\n")
#print("\n Client side encrypted input data to server (displaying last 10 rows): \n",encrypted_testdata.tail(10))
self.log.info('Client side encrypted input data to server (displaying last 10 rows): \n'+str(encrypted_testdata.tail(10)))
enc_predictions = self.connect_xgboostserver(ppBoostKey,encrypted_testdata)
#print("\n Encrypted prediction from server (displaying last 10 values.): \n",enc_predictions[-10:])
#self.log.info('\n Encrypted prediction from server (displaying last 10 values.): \n'+str(enc_predictions[-10:]))
## Decrypted predition
dec = self.ppxgboostmulticlass_predict(enc_predictions,private_key)
# ppxgboost_pred=pd.DataFrame(list(zip(dec, predictions)),columns =['homomorphic_prediction', 'actual_prediction'])
ppxgboost_pred=pd.DataFrame(dec,columns =['homomorphic_prediction'])
self.log.info("final decrypted prediction at client side:: \n"+str(ppxgboost_pred))
return ppxgboost_pred
## For standalone testing
if __name__ == '__main__':
problemtype='Multi class classification'
data=None
targetfeature=None
ppxgboost_client_obj=client_ppxgboost(problemtype,data,targetfeature)
ppxgboost_dec_predictions = ppxgboost_client_obj.main_client()
# print("In main: ppxgboost_dec_predictions: \n",ppxgboost_dec_predictions) |
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
explainable_ai.py | from script.inputprofiler import inputprofiler
def preprocessing(data):
profilerobj = inputprofiler()
data = profilerobj.run(data)
data = data.astype(np.float64)
return(data)
import matplotlib.pyplot as plt
try:
from sklearn.externals import joblib
except:
import joblib
import os,sys
import pandas as pd
from alibi.explainers import ALE,plot_ale
import io
import json
import urllib, base64
import numpy as np
from scipy.stats import linregress
from statistics import mean
def get_ranked_values(explanation):
ranked_shap_vals = []
for cls_idx in range(len(explanation.shap_values)):
this_ranking = (
explanation.raw['importances'][str(cls_idx)]['ranked_effect'],
explanation.raw['importances'][str(cls_idx)]['names']
)
ranked_shap_vals.append(this_ranking)
return ranked_shap_vals
def feature_importance_using_shap(model,X,featuresNames,classes,x_test,x_test_waterfall):
from alibi.explainers import KernelShap
import shap
shap.initjs()
if hasattr(model, "decision_function"):
pred_fcn = model.decision_function
elif hasattr(model, "predict_proba"):
pred_fcn = model.predict_proba
else:
pred_fcn = model.predict
try:
svm_explainer = KernelShap(pred_fcn,feature_names=featuresNames)
xtest = x_test[0].reshape(1, -1)
svm_explainer.fit(X,n_background_samples=100)
svm_explanation = svm_explainer.explain(xtest)
try:
idx = 0
instance = x_test[0][None, :]
pred = model.predict(instance)
class_idx = pred.item()
if isinstance(svm_explainer.expected_value,np.ndarray):
forceplot = shap.force_plot(svm_explainer.expected_value[class_idx],svm_explanation.shap_values[class_idx][idx,:],instance,feature_names=featuresNames,matplotlib=True,show=False)
else:
forceplot = shap.force_plot(svm_explainer.expected_value,svm_explanation.shap_values[0][idx,:],instance,feature_names=featuresNames,matplotlib=True,show=False)
plt.tight_layout(pad = 0)
image = io.BytesIO()
plt.savefig(image, format='png')
image.seek(0)
string = base64.b64encode(image.read())
image_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except Exception as inst:
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
image_64 = ''
try:
plt.clf()
if isinstance(svm_explainer.expected_value,np.ndarray):
r = shap.multioutput_decision_plot(svm_explainer.expected_value.tolist(),
svm_explanation.shap_values,
idx,
feature_names=featuresNames,
feature_order='importance',
highlight=[class_idx],
legend_labels=classes,
return_objects=True,
legend_location='lower right',show=False)
else:
expectedvalues = [svm_explainer.expected_value]
r = shap.multioutput_decision_plot(expectedvalues,
svm_explanation.shap_values,
idx,
feature_names=featuresNames,
highlight = [0],
return_objects=True,
legend_labels=['Value'],
feature_order='importance',
show=False)
plt.tight_layout(pad = 0)
image = io.BytesIO()
plt.savefig(image, format='png')
image.seek(0)
string = base64.b64encode(image.read())
image2_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
image2_64 = ''
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
image2_64 = ''
image_64 = ''
try:
plt.clf()
x_test_waterfall = x_test_waterfall[featuresNames]
explainer = shap.Explainer(model.predict, x_test_waterfall, feature_names=featuresNames)
shap_values = explainer(x_test)
r = shap.plots.waterfall(shap_values[0], show=False)
image = io.BytesIO()
plt.savefig(image, format='png', bbox_inches='tight')
image.seek(0)
string = base64.b64encode(image.read())
image3_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
image3_64 = ''
return (image_64, image2_64, image3_64)
def feature_importance(xtrain,ytrain,xfeatures,yfeature,problemType):
if problemType == 'classification':
from sklearn.feature_selection import SelectFromModel
from sklearn.ensemble import ExtraTreesClassifier
selector = SelectFromModel(ExtraTreesClassifier())
selector.fit(xtrain,ytrain)
values = selector.estimator_.feature_importances_
elif problemType == 'regression':
from sklearn.feature_selection import SelectFromModel
from sklearn.linear_model import Lasso
selector = SelectFromModel(Lasso())
selector.fit(xtrain,ytrain)
values = np.abs(selector.estimator_.coef_)
labels = xtrain.columns.tolist()
dft = pd.DataFrame()
dft['labels'] = labels
dft['values'] = values
maxrecord = dft.iloc[dft['values'].argmax()]
mostimportantfeature = maxrecord['labels']
f_imp = dft.to_json(orient='records')
return(f_imp,mostimportantfeature)
def get_trust_score(prdictfn,proba_fun,X_train,y_train):
from alibi.confidence import TrustScore
ts = TrustScore(k_filter=10,alpha=.05,filter_type='distance_knn',leaf_size=40,metric='euclidean',dist_filter_type='point')
ts.fit(X_train, y_train, classes=3)
y_pred = prdictfn(X_train)
#y_prod = proba_fun(X_train)
#probas = y_prod[range(len(y_pred)), y_pred]
score, closest_class = ts.score(X_train, y_pred,k=2,dist_type='point')
return(mean(score))
def getCounterFactuals(model,prdictfn,features,x_train,categories):
from alibi.explainers import CounterFactualProto
cat_vars_ord = {}
categoryList=categories.keys().tolist()
categoryCountList=categories.tolist()
for i in range(0,len(categoryCountList)):
cat_vars_ord[categoryList[i]] = categoryCountList[i]
print(cat_vars_ord)
X = x_train[0].reshape((1,) + x_train[0].shape)
shape = X.shape
print(shape)
beta = .01
c_init = 1.
c_steps = 5
max_iterations = 500
rng = (-1., 1.) # scale features between -1 and 1
feature_range = (x_train.min(axis=0), x_train.max(axis=0))
cf = CounterFactualProto(prdictfn,shape,cat_vars=cat_vars_ord)
explanation = cf.explain(X)
print(explanation)
def getAnchorTabularofFirstRecord(predict_fn,features,X_train,X_test,labelMap, class_percent=None):
threshold = 0.95
from alibi.explainers import AnchorTabular
explainer = AnchorTabular(predict_fn, features)
explainer.fit(X_train.values)
X_test = X_test.values
anchors = []
for idx in range(len(X_test)):
prediction = explainer.predictor(X_test[idx].reshape(1, -1))[0]
if len(labelMap) > 0:
predictionstr = list(labelMap.keys())[list(labelMap.values()).index(prediction)]
else:
predictionstr = prediction
explanation = explainer.explain(X_test[idx],threshold=threshold)
if str(explanation.anchor) == '[]':
if class_percent and class_percent.get(prediction, 0.0) > threshold:
anchor = f"Cannot explain the prediction of this class ({predictionstr}) since there is no salient subset of features that is necessary for the prediction to hold. This behaviour is typical when the data is very imbalanced and is seen for the majority class in a classification problem."
else:
anchor = f'Can not get the explanation for {predictionstr}.'
precision = explanation.precision[0]
else:
anchor = '%s' % (' AND '.join(explanation.anchor))
precision = explanation.precision
coverage = explanation.coverage
anchorjson = {}
anchorjson['features'] = eval(str(features))
anchorjson['values'] = eval(str(list(X_test[idx])))
anchorjson['prediction'] = str(predictionstr)
anchorjson['precision'] = str(round(precision,2))
anchorjson['anchor'] = anchor
anchors.append(anchorjson)
print(anchors)
try:
return(json.dumps(anchors))
except Exception as e:
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return(json.dumps({}))
def ale_analysis():
displaypath = os.path.join(os.path.dirname(os.path.abspath(__file__)),"etc","display.json")
with open(displaypath) as file:
config = json.load(file)
file.close()
model = joblib.load(os.path.join(os.path.dirname(os.path.abspath(__file__)),"model",config['saved_model']))
predict_fn = lambda x: model.predict(x)
predictproba_fn = lambda x: model.predict_proba(x)
dathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)),'data','postprocesseddata.csv.gz')
dataFrame=pd.read_csv(dathPath,compression='gzip')
#dataFrame = pd.read_csv(dathPath)
testdathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)),'data','predicteddata.csv.gz')
testdataFrame=pd.read_csv(testdathPath,compression='gzip')
#testdataFrame = pd.read_csv(testdathPath)
features = config['modelFeatures']#['SepalWidthCm','PetalLengthCm']
targetfeature = config['targetFeature']#'Species'
labelMap = config['labelMaps']
targetData = dataFrame[targetfeature]
if config['problemType'].lower() == 'regression':
X_train = dataFrame[features]
X_test = testdataFrame.head(5)
X_test = X_test[features]
else:
valueCount=targetData.value_counts()
class_percent = (valueCount/ len(targetData)).to_dict()
categoryList=valueCount.keys().tolist()
class_names = categoryList
X_train = dataFrame[features]
X_test = testdataFrame.groupby('predict').first().reset_index()
X_test = X_test[features]
f_imp,m_imp_f = feature_importance(X_train,targetData,features,targetfeature,config['problemType'].lower())
if hasattr(model, "decision_function"):
logit_fun_lr = model.decision_function
try:
logit_ale_lr = ALE(logit_fun_lr, feature_names=features, target_names=categoryList)
logit_exp_lr = logit_ale_lr.explain(X_train.values)
values = logit_exp_lr.ale_values
feature = logit_exp_lr.feature_names
feature_values = logit_exp_lr.feature_values
lines= []
sentences = []
for x in range(0,len(feature)):
f_value = feature_values[x]
value = values[x]
lines= []
for y in range(0,len(class_names)):
line = []
for z in value:
cordinate = z[y]
line.append(cordinate)
lines.append(line)
line = lines[0]
slope1, intercept1, r_value, p_value, std_err = linregress(f_value,line)
line = lines[1]
slope2, intercept2, r_value, p_value, std_err = linregress(f_value,line)
xi = (intercept1-intercept2) / (slope2-slope1)
xi = round(xi,2)
lastvalues = {}
i = 0
for line in lines:
value = line[len(line)-1]
lastvalues[class_names[i]] = value
i = i+1
Keymax = max(lastvalues, key=lastvalues.get)
Keymin = min(lastvalues, key=lastvalues.get)
Keymaxclass = list(labelMap.keys())[list(labelMap.values()).index(Keymax)]
Keyminclass = list(labelMap.keys())[list(labelMap.values()).index(Keymin)]
sentense = '<b>Effect of '+str(feature[x])+'</b><br>For data samples having <b>'+str(feature[x])+'</b> >= <b>~'+str(xi)+'</b> ,there is a very high chance that they are of class <b>'+str(Keymaxclass)+'</b> '+targetfeature+'. For data samples having <b>'+str(feature[x])+'</b> < <b>~'+str(xi)+'</b> there is a very high change that they are of class <b>'+str(Keyminclass)+'</b> '+targetfeature+'.'
sentences.append(sentense)
except:
sentense = ''
sentences.append(sentense)
xi = 0
elif hasattr(model, "predict_proba"):
logit_fun_lr = model.predict_proba
logit_ale_lr = ALE(logit_fun_lr, feature_names=features, target_names=categoryList)
logit_exp_lr = logit_ale_lr.explain(X_train.values)
values = logit_exp_lr.ale_values
feature = logit_exp_lr.feature_names
feature_values = logit_exp_lr.feature_values
lines= []
sentences = []
sentense = 'Graphs gives a feature value how much more(less) probability does the model assign to each class relative to mean prediction. This also means that any increase in relative probability of one class must result into a decrease in probability of another class.'
sentences.append(sentense)
xi = 0
elif hasattr(model, "predict"):
logit_fun_lr = model.predict
logit_ale_lr = ALE(logit_fun_lr, feature_names=features, target_names=['Value'])
logit_exp_lr = logit_ale_lr.explain(X_train.values)
values = logit_exp_lr.ale_values
feature = logit_exp_lr.feature_names
feature_values = logit_exp_lr.feature_values
lines= []
sentences = []
sentense = 'The ALE value corresponding to that feature value is difference to the mean effect of that feature. Put differently, the ALE value is the relative feature effect on the prediction at that feature value.'
sentences.append(sentense)
xi = 0
if (len(features)%2 ==0):
n_cols = int(len(features)/2)
else:
n_cols = int(len(features)/2)+1
figheight = n_cols*3
try:
plot_ale(logit_exp_lr,n_cols=2, fig_kw={'figwidth': 8, 'figheight': figheight})
plt.tight_layout(pad = 0)
image = io.BytesIO()
plt.savefig(image, format='png')
image.seek(0)
string = base64.b64encode(image.read())
image_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except:
image_64 = ''
#score = get_trust_score(model.predict,proba_fun_lr,X_train.values,targetData.values)
if config['problemType'].lower() == 'classification':
anchorjson = getAnchorTabularofFirstRecord(predict_fn,features,X_train,X_test,labelMap,class_percent)
else:
anchorjson = getAnchorTabularofFirstRecord(predict_fn,features,X_train,X_test,labelMap)
#anchors=[]
#anchorjson = json.dumps(anchors)
#feature_importance_using_shap(model,X_train.values,features,class_names)
#getCounterFactuals(model,predictproba_fn,features,X_train.values,valueCount)
output_json = {"status":"SUCCESS","data":{"data":image_64,"most_influencedfeature":m_imp_f,"interceptionpoint":xi,"sentences":sentences,"feature_importance":json.loads(f_imp),"anchorjson":json.loads(anchorjson)}}
output_json = json.dumps(output_json)
print("aion_ai_explanation:",output_json)
return(output_json)
def local_analysis(jsonData):
jsonData = json.loads(jsonData)
displaypath = os.path.join(os.path.dirname(os.path.abspath(__file__)),"etc","display.json")
with open(displaypath) as file:
config = json.load(file)
file.close()
model = joblib.load(os.path.join(os.path.dirname(os.path.abspath(__file__)),"model",config['saved_model']))
predict_fn = lambda x: model.predict(x)
dathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)),'data','postprocesseddata.csv.gz')
dataFrame=pd.read_csv(dathPath,compression='gzip')
testdathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data', 'predicteddata.csv.gz')
testdataFrame = pd.read_csv(testdathPath, compression='gzip')
features = config['modelFeatures']#['SepalWidthCm','PetalLengthCm']
targetfeature = config['targetFeature']#'Species'
targetData = dataFrame[targetfeature]
valueCount=targetData.value_counts()
categoryList=valueCount.keys().tolist()
class_names = categoryList
#class_names = class_names.sort()
X_train = dataFrame[features]
from pandas import json_normalize
df_test = json_normalize(jsonData)
df_test = preprocessing(df_test)
df_test = df_test[features]
from alibi.explainers import AnchorTabular
explainer = AnchorTabular(predict_fn, features)
explainer.fit(X_train.values)
df_test = df_test.values
prediction = explainer.predictor(df_test.reshape(1, -1))[0]
labelMap = config['labelMaps']
if len(labelMap) > 0:
prediction = list(labelMap.keys())[list(labelMap.values()).index(prediction)]
else:
prediction = str(prediction)
try:
explanation = explainer.explain(df_test,threshold=0.85)
if str(explanation.anchor) == '[]':
anchor = 'NA'
precision = str(round(explanation.precision[0],2))
else:
anchor = '%s' % (' AND '.join(explanation.anchor))
precision = str(round(explanation.precision,2))
coverage = explanation.coverage
except Exception as e:
print(e)
anchor = 'NA'
precision = 0
coverage = 0
df_test_waterfall = testdataFrame
forceplot,multidecisionplot,waterfallplot = feature_importance_using_shap(model,X_train.head(300).values,features,class_names,df_test,df_test_waterfall)
output_json = {"status":"SUCCESS","data":{"anchor":anchor,"precision":precision,"coverage":coverage,"prediction":prediction,"forceplot":forceplot,"multidecisionplot":multidecisionplot,"waterfallplot":waterfallplot}}
#print(output_json)
output_json = json.dumps(output_json)
print("aion_ai_explanation:",output_json)
return(output_json)
if __name__ == '__main__':
analysis_type = sys.argv[1]
if analysis_type.lower() == 'global':
ale_analysis()
if analysis_type.lower() == 'local':
data = sys.argv[2]
local_analysis(data) |
explainabledl_ai.py | from script.inputprofiler import inputprofiler
def preprocessing(data):
profilerobj = inputprofiler()
data = profilerobj.run(data)
data = data.astype(np.float64)
return(data)
import matplotlib.pyplot as plt
try:
from sklearn.externals import joblib
except:
import joblib
import os,sys
import pandas as pd
from alibi.explainers import ALE,plot_ale
import io
import json
import urllib, base64
import numpy as np
from scipy.stats import linregress
from statistics import mean
from tensorflow.keras.models import load_model
from tensorflow.keras import backend as K
import tensorflow as tf
tf.compat.v1.disable_eager_execution()
def recall_m(y_true, y_pred):
true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))
possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))
recall = true_positives / (possible_positives + K.epsilon())
return recall
def precision_m(y_true, y_pred):
true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))
predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))
precision = true_positives / (predicted_positives + K.epsilon())
return precision
def f1_m(y_true, y_pred):
precision = precision_m(y_true, y_pred)
recall = recall_m(y_true, y_pred)
return 2*((precision*recall)/(precision+recall+K.epsilon()))
def rmse_m(y_true, y_pred):
return K.sqrt(K.mean(K.square(y_pred - y_true), axis=-1))
def r_square(y_true, y_pred):
SS_res = K.sum(K.square(y_true-y_pred))
SS_tot = K.sum(K.square(y_true-K.mean(y_true)))
return (1 - SS_res/(SS_tot+K.epsilon()))
def get_ranked_values(explanation):
ranked_shap_vals = []
for cls_idx in range(len(explanation.shap_values)):
this_ranking = (
explanation.raw['importances'][str(cls_idx)]['ranked_effect'],
explanation.raw['importances'][str(cls_idx)]['names']
)
ranked_shap_vals.append(this_ranking)
return ranked_shap_vals
def feature_importance_using_shap(model,X,featuresNames,classes,x_test,problemType,modelname,x_test_waterfall):
from alibi.explainers import KernelShap
import shap
shap.initjs()
if hasattr(model, "decision_function") and problemType.lower() == 'classification':
pred_fcn = model.decision_function
elif hasattr(model, "predict_proba") and problemType.lower() == 'classification':
pred_fcn = lambda x: model.predict_proba(np.expand_dims(x, axis=2))
else:
if modelname == 'Neural Network':
pred_fcn = lambda x: model.predict(x)
else:
pred_fcn = lambda x: model.predict(np.expand_dims(x, axis=2))
svm_explainer = KernelShap(pred_fcn,feature_names=featuresNames)
xtest = x_test[0].reshape(1, -1)
svm_explainer.fit(X,n_background_samples=100)
svm_explanation = svm_explainer.explain(xtest)
try:
idx = 0
instance = x_test[0][None, :]
if problemType.lower() == 'classification':
if modelname == 'Neural Network':
instance = x_test
else:
instance = np.expand_dims(x_test, axis=2)
pred = np.argmax(model.predict(instance),axis=1)
class_idx = pred.item()
else:
instance = np.expand_dims(x_test, axis=2)
pred = model.predict(instance)
class_idx = 0
if isinstance(svm_explainer.expected_value,np.ndarray):
forceplot = shap.force_plot(svm_explainer.expected_value[class_idx],svm_explanation.shap_values[class_idx][idx,:],instance,feature_names=featuresNames,matplotlib=True,show=False)
else:
forceplot = shap.force_plot(svm_explainer.expected_value,svm_explanation.shap_values[0][idx,:],instance,feature_names=featuresNames,matplotlib=True,show=False)
plt.tight_layout(pad = 0)
image = io.BytesIO()
plt.savefig(image, format='png')
image.seek(0)
string = base64.b64encode(image.read())
image_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
image_64 = ''
try:
plt.clf()
if isinstance(svm_explainer.expected_value,np.ndarray):
r = shap.multioutput_decision_plot(svm_explainer.expected_value.tolist(),
svm_explanation.shap_values,
idx,
feature_names=featuresNames,
feature_order='importance',
highlight=[class_idx],
legend_labels=classes,
return_objects=True,
legend_location='lower right',show=False)
else:
expectedvalues = [svm_explainer.expected_value]
r = shap.multioutput_decision_plot(expectedvalues,
svm_explanation.shap_values,
idx,
feature_names=featuresNames,
highlight = [0],
return_objects=True,
legend_labels=['Value'],
feature_order='importance',
show=False)
plt.tight_layout(pad = 0)
image = io.BytesIO()
plt.savefig(image, format='png')
image.seek(0)
string = base64.b64encode(image.read())
image2_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
image2_64 = ''
try:
plt.clf()
explainer = shap.DeepExplainer(model, X)
shap_values = explainer.shap_values(x_test_waterfall.values)
exp = shap.plots._waterfall.waterfall_legacy(explainer.expected_value[0].numpy(), shap_values[0][0],feature_names=featuresNames,show=False)
image = io.BytesIO()
plt.savefig(image, format='png', bbox_inches='tight')
image.seek(0)
string = base64.b64encode(image.read())
image3_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
image3_64 = ''
return(image_64,image2_64,image3_64)
def feature_importance(xtrain,ytrain,xfeatures,yfeature,problemType):
if problemType == 'classification':
from sklearn.feature_selection import SelectFromModel
from sklearn.ensemble import ExtraTreesClassifier
selector = SelectFromModel(ExtraTreesClassifier())
selector.fit(xtrain,ytrain)
values = selector.estimator_.feature_importances_
elif problemType == 'regression':
from sklearn.feature_selection import SelectFromModel
from sklearn.linear_model import Lasso
selector = SelectFromModel(Lasso())
selector.fit(xtrain,ytrain)
values = np.abs(selector.estimator_.coef_)
labels = xtrain.columns.tolist()
dft = pd.DataFrame()
dft['labels'] = labels
dft['values'] = values
maxrecord = dft.iloc[dft['values'].argmax()]
mostimportantfeature = maxrecord['labels']
f_imp = dft.to_json(orient='records')
return(f_imp,mostimportantfeature)
def get_trust_score(prdictfn,proba_fun,X_train,y_train):
from alibi.confidence import TrustScore
ts = TrustScore(k_filter=10,alpha=.05,filter_type='distance_knn',leaf_size=40,metric='euclidean',dist_filter_type='point')
ts.fit(X_train, y_train, classes=3)
y_pred = prdictfn(X_train)
#y_prod = proba_fun(X_train)
#probas = y_prod[range(len(y_pred)), y_pred]
score, closest_class = ts.score(X_train, y_pred,k=2,dist_type='point')
return(mean(score))
def getCounterFactuals(model,prdictfn,features,x_train,categories):
from alibi.explainers import CounterFactualProto
cat_vars_ord = {}
categoryList=categories.keys().tolist()
categoryCountList=categories.tolist()
for i in range(0,len(categoryCountList)):
cat_vars_ord[categoryList[i]] = categoryCountList[i]
print(cat_vars_ord)
X = x_train[0].reshape((1,) + x_train[0].shape)
shape = X.shape
print(shape)
beta = .01
c_init = 1.
c_steps = 5
max_iterations = 500
rng = (-1., 1.) # scale features between -1 and 1
feature_range = (x_train.min(axis=0), x_train.max(axis=0))
cf = CounterFactualProto(prdictfn,shape,cat_vars=cat_vars_ord)
explanation = cf.explain(X)
print(explanation)
def getAnchorTabularofFirstRecord(predict_fn,features,X_train,X_test,labelMap, class_percent=None):
threshold = 0.95
from alibi.explainers import AnchorTabular
explainer = AnchorTabular(predict_fn, features)
explainer.fit(X_train.values)
X_test = X_test.values
anchors = []
for idx in range(len(X_test)):
prediction = explainer.predictor(X_test[idx].reshape(1, -1))[0]
if isinstance(prediction,np.ndarray):
prediction = prediction[0]
if len(labelMap) > 0:
predictionstr = list(labelMap.keys())[list(labelMap.values()).index(prediction)]
else:
predictionstr = str(prediction)
try:
explanation = explainer.explain(X_test[idx],threshold=threshold)
if str(explanation.anchor) == '[]':
if class_percent and class_percent.get(prediction, 0.0) > threshold:
anchor = f"Cannot explain the prediction of this class ({predictionstr}) since there is no salient subset of features that is necessary for the prediction to hold. This behaviour is typical when the data is very imbalanced and is seen for the majority class in a classification problem."
else:
anchor = f'Can not get the explaination for {predictionstr}.'
precision = explanation.precision[0]
else:
anchor = '%s' % (' AND '.join(explanation.anchor))
precision = explanation.precision
coverage = explanation.coverage
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
anchor = 'Reason Not found'
precision = 0
anchorjson = {}
anchorjson['features'] = eval(str(features))
anchorjson['values'] = eval(str(list(X_test[idx])))
anchorjson['prediction'] = predictionstr
anchorjson['precision'] = precision
anchorjson['anchor'] = anchor
anchors.append(anchorjson)
return(json.dumps(anchors))
def ale_analysis():
displaypath = os.path.join(os.path.dirname(os.path.abspath(__file__)),"etc","display.json")
with open(displaypath) as file:
config = json.load(file)
file.close()
scoreParam = config['scoreParam']
filename_dl = os.path.join(os.path.dirname(os.path.abspath(__file__)),"model",config['saved_model'])
if(scoreParam.lower() == 'rmse'):
model = load_model(filename_dl,custom_objects={"rmse": rmse_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[rmse_m])
elif(scoreParam.lower() == 'r2'):
model = load_model(filename_dl,custom_objects={"r2": r_square},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[r_square])
elif(scoreParam.lower() == 'recall'):
model = load_model(filename_dl,custom_objects={"recall": recall_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[recall_m])
elif(scoreParam.lower() == 'precision'):
model = load_model(filename_dl,custom_objects={"precision": precision_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[precision_m])
elif(scoreParam.lower() == 'roc_auc'):
model = load_model(filename_dl,compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[tf.keras.metrics.AUC()])
elif(scoreParam.lower() == 'f1_score'):
model = load_model(filename_dl,custom_objects={"f1_score": f1_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[f1_m])
else:
model = load_model(filename_dl)
if config['modelname'] == 'Neural Network':
predict_fn = lambda x: model.predict(x)
else:
predict_fn = lambda x: model.predict(np.expand_dims(x, axis=2))
predictproba_fn = lambda x: model.predict_proba(np.expand_dims(x, axis=2))
dathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)),'data','postprocesseddata.csv.gz')
dataFrame=pd.read_csv(dathPath,compression='gzip')
testdathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)),'data','predicteddata.csv.gz')
testdataFrame=pd.read_csv(testdathPath,compression='gzip')
features = config['modelFeatures']#['SepalWidthCm','PetalLengthCm']
targetfeature = config['targetFeature']#'Species'
labelMap = config['labelMaps']
targetData = dataFrame[targetfeature]
if config['problemType'].lower() == 'regression':
X_train = dataFrame[features]
X_test = testdataFrame.head(5)
X_test = X_test[features]
else:
valueCount=targetData.value_counts()
class_percent = (valueCount/ len(targetData)).to_dict()
categoryList=valueCount.keys().tolist()
class_names = categoryList
X_train = dataFrame[features]
X_test = testdataFrame.groupby('predict').first().reset_index()
X_test = X_test[features]
f_imp,m_imp_f = feature_importance(X_train,targetData,features,targetfeature,config['problemType'].lower())
if hasattr(model, "decision_function") and config['problemType'].lower() == 'classification':
logit_fun_lr = model.decision_function
try:
logit_ale_lr = ALE(logit_fun_lr, feature_names=features, target_names=categoryList)
logit_exp_lr = logit_ale_lr.explain(X_train.values)
values = logit_exp_lr.ale_values
feature = logit_exp_lr.feature_names
feature_values = logit_exp_lr.feature_values
lines= []
sentences = []
for x in range(0,len(feature)):
f_value = feature_values[x]
value = values[x]
lines= []
for y in range(0,len(class_names)):
line = []
for z in value:
cordinate = z[y]
line.append(cordinate)
lines.append(line)
line = lines[0]
slope1, intercept1, r_value, p_value, std_err = linregress(f_value,line)
line = lines[1]
slope2, intercept2, r_value, p_value, std_err = linregress(f_value,line)
xi = (intercept1-intercept2) / (slope2-slope1)
xi = round(xi,2)
lastvalues = {}
i = 0
for line in lines:
value = line[len(line)-1]
lastvalues[class_names[i]] = value
i = i+1
Keymax = max(lastvalues, key=lastvalues.get)
Keymin = min(lastvalues, key=lastvalues.get)
Keymaxclass = list(labelMap.keys())[list(labelMap.values()).index(Keymax)]
Keyminclass = list(labelMap.keys())[list(labelMap.values()).index(Keymin)]
sentense = '<b>Effect of '+str(feature[x])+'</b><br>For data samples having <b>'+str(feature[x])+'</b> >= <b>~'+str(xi)+'</b> ,there is a very high chance that they are of class <b>'+str(Keymaxclass)+'</b> '+targetfeature+'. For data samples having <b>'+str(feature[x])+'</b> < <b>~'+str(xi)+'</b> there is a very high change that they are of class <b>'+str(Keyminclass)+'</b> '+targetfeature+'.'
sentences.append(sentense)
except:
sentense = ''
sentences.append(sentense)
xi = 0
elif hasattr(model, "predict_proba") and config['problemType'].lower() == 'classification':
logit_fun_lr = lambda x: model.predict_proba(np.expand_dims(x, axis=2))
logit_ale_lr = ALE(logit_fun_lr, feature_names=features, target_names=categoryList)
print(model.__class__)
try:
logit_exp_lr = logit_ale_lr.explain(X_train.values)
except:
X = np.expand_dims(X_train, axis=2)
logit_exp_lr = logit_ale_lr.explain(X)
values = logit_exp_lr.ale_values
feature = logit_exp_lr.feature_names
feature_values = logit_exp_lr.feature_values
lines= []
sentences = []
sentense = 'Graphs gives a feature value how much more(less) probability does the model assign to each class relative to mean prediction. This also means that any increase in relative probability of one class must result into a decrease in probability of another class.'
sentences.append(sentense)
xi = 0
elif hasattr(model, "predict"):
try:
if config['modelname'] == 'Neural Network':
logit_fun_lr = lambda x: model.predict(x)
else:
logit_fun_lr = lambda x: model.predict(np.expand_dims(x, axis=2))
logit_ale_lr = ALE(logit_fun_lr, feature_names=features, target_names=['Value'])
logit_exp_lr = logit_ale_lr.explain(X_train.values)
values = logit_exp_lr.ale_values
feature = logit_exp_lr.feature_names
feature_values = logit_exp_lr.feature_values
lines= []
sentences = []
sentense = 'The ALE value corresponding to that feature value is difference to the mean effect of that feature. Put differently, the ALE value is the relative feature effect on the prediction at that feature value.'
sentences.append(sentense)
xi = 0
except:
xi = 0
sentences = []
if (len(features)%2 ==0):
n_cols = int(len(features)/2)
else:
n_cols = int(len(features)/2)+1
figheight = n_cols*3
try:
plot_ale(logit_exp_lr,n_cols=2, fig_kw={'figwidth': 8, 'figheight': figheight})
plt.tight_layout(pad = 0)
image = io.BytesIO()
plt.savefig(image, format='png')
image.seek(0)
string = base64.b64encode(image.read())
image_64 = 'data:image/png;base64,' + urllib.parse.quote(string)
except:
image_64 = ''
#score = get_trust_score(model.predict,proba_fun_lr,X_train.values,targetData.values)
if config['problemType'].lower() == 'classification':
anchorjson = getAnchorTabularofFirstRecord(predict_fn,features,X_train,X_test,labelMap,class_percent)
else:
anchorjson = getAnchorTabularofFirstRecord(predict_fn,features,X_train,X_test,labelMap)
#anchors=[]
#anchorjson = json.dumps(anchors)
#feature_importance_using_shap(model,X_train.values,features,class_names)
#getCounterFactuals(model,predictproba_fn,features,X_train.values,valueCount)
output_json = {"status":"SUCCESS","data":{"data":image_64,"most_influencedfeature":m_imp_f,"interceptionpoint":xi,"sentences":sentences,"feature_importance":json.loads(f_imp),"anchorjson":json.loads(anchorjson)}}
output_json = json.dumps(output_json)
print("aion_ai_explanation:",output_json)
return(output_json)
def local_analysis(jsonData):
jsonData = json.loads(jsonData)
displaypath = os.path.join(os.path.dirname(os.path.abspath(__file__)),"etc","display.json")
with open(displaypath) as file:
config = json.load(file)
file.close()
scoreParam = config['scoreParam']
filename_dl = os.path.join(os.path.dirname(os.path.abspath(__file__)),"model",config['saved_model'])
if(scoreParam.lower() == 'rmse'):
model = load_model(filename_dl,custom_objects={"rmse": rmse_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[rmse_m])
elif(scoreParam.lower() == 'r2'):
model = load_model(filename_dl,custom_objects={"r2": r_square},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[r_square])
elif(scoreParam.lower() == 'recall'):
model = load_model(filename_dl,custom_objects={"recall": recall_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[recall_m])
elif(scoreParam.lower() == 'precision'):
model = load_model(filename_dl,custom_objects={"precision": precision_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[precision_m])
elif(scoreParam.lower() == 'roc_auc'):
model = load_model(filename_dl,compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[tf.keras.metrics.AUC()])
elif(scoreParam.lower() == 'f1_score'):
model = load_model(filename_dl,custom_objects={"f1_score": f1_m},compile=False)
model.compile(loss='binary_crossentropy',optimizer='Nadam', metrics=[f1_m])
else:
model = load_model(filename_dl)
if config['modelname'] == 'Neural Network':
predict_fn = lambda x: model.predict(x)
else:
predict_fn = lambda x: model.predict(np.expand_dims(x, axis=2))
dathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)),'data','postprocesseddata.csv.gz')
dataFrame=pd.read_csv(dathPath,compression='gzip')
testdathPath = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'data', 'predicteddata.csv.gz')
testdataFrame = pd.read_csv(testdathPath,compression='gzip')
features = config['modelFeatures']#['SepalWidthCm','PetalLengthCm']
targetfeature = config['targetFeature']#'Species'
targetData = dataFrame[targetfeature]
valueCount=targetData.value_counts()
categoryList=valueCount.keys().tolist()
class_names = categoryList
#class_names = class_names.sort()
X_train = dataFrame[features]
from pandas import json_normalize
df_test = json_normalize(jsonData)
df_test = preprocessing(df_test)
df_test = df_test[features]
from alibi.explainers import AnchorTabular
explainer = AnchorTabular(predict_fn, features)
explainer.fit(X_train.values)
df_test = df_test.values
prediction = explainer.predictor(df_test.reshape(1, -1))[0]
if isinstance(prediction,np.ndarray):
prediction = prediction[0]
labelMap = config['labelMaps']
if len(labelMap) > 0:
prediction = list(labelMap.keys())[list(labelMap.values()).index(prediction)]
else:
prediction = str(prediction)
try:
explanation = explainer.explain(df_test.reshape(1, -1),threshold=0.85)
if str(explanation.anchor) == '[]':
anchor = 'NA'
precision = explanation.precision[0]
else:
anchor = '%s' % (' AND '.join(explanation.anchor))
precision = explanation.precision
coverage = explanation.coverage
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
anchor = 'NA'
precision = 0
coverage = 0
df_test_waterfall = testdataFrame
forceplot,multidecisionplot,waterfallplot = feature_importance_using_shap(model,X_train.head(300).values,features,class_names,df_test,config['problemType'],config['modelname'],df_test_waterfall)
output_json = {"status":"SUCCESS","data":{"anchor":anchor,"precision":precision,"coverage":coverage,"prediction":prediction,"forceplot":forceplot,"multidecisionplot":multidecisionplot,'waterfallplot':waterfallplot}}
print(output_json)
output_json = json.dumps(output_json)
print("aion_ai_explanation:",output_json)
return(output_json)
if __name__ == '__main__':
analysis_type = sys.argv[1]
if analysis_type.lower() == 'global':
ale_analysis()
if analysis_type.lower() == 'local':
data = sys.argv[2]
#jsonData = json.loads(data)
local_analysis(data) |
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
dataProfiler.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import io
import json
import logging
import pandas as pd
import sys
import numpy as np
from pathlib import Path
from word2number import w2n
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import OrdinalEncoder
from sklearn.preprocessing import OneHotEncoder
from sklearn.impute import SimpleImputer, KNNImputer
from sklearn.pipeline import Pipeline, FeatureUnion
from sklearn.preprocessing import FunctionTransformer
from sklearn.preprocessing import MinMaxScaler,StandardScaler
from sklearn.preprocessing import PowerTransformer
from sklearn.compose import ColumnTransformer
from sklearn.base import TransformerMixin
from sklearn.ensemble import IsolationForest
from category_encoders import TargetEncoder
import scipy
try:
import transformations.data_profiler_functions as cs
except:
import data_profiler_functions as cs
if 'AION' in sys.modules:
try:
from appbe.app_config import DEBUG_ENABLED
except:
DEBUG_ENABLED = False
else:
DEBUG_ENABLED = False
log_suffix = f'[{Path(__file__).stem}] '
class profiler():
def __init__(self, xtrain, ytrain=None, target=None, encode_target = False, config={}, keep_unprocessed=[],data_path=None,log=None):
if not isinstance(xtrain, pd.DataFrame):
raise ValueError(f'{log_suffix}supported data type is pandas.DataFrame but provide data is of {type(xtrain)} type')
if xtrain.empty:
raise ValueError(f'{log_suffix}Data frame is empty')
if target and target in xtrain.columns:
self.target = xtrain[target]
xtrain.drop(target, axis=1, inplace=True)
self.target_name = target
elif ytrain:
self.target = ytrain
self.target_name = 'target'
else:
self.target = pd.Series()
self.target_name = None
self.data_path = data_path
self.encode_target = encode_target
self.label_encoder = None
self.data = xtrain
self.keep_unprocessed = keep_unprocessed
self.colm_type = {}
for colm, infer_type in zip(self.data.columns, self.data.dtypes):
self.colm_type[colm] = infer_type
self.numeric_feature = []
self.cat_feature = []
self.text_feature = []
self.wordToNumericFeatures = []
self.added_features = []
self.pipeline = []
self.dropped_features = {}
self.train_features_type={}
self.__update_type()
self.config = config
self.featureDict = config.get('featureDict', [])
self.output_columns = []
self.feature_expender = []
self.text_to_num = {}
self.force_numeric_conv = []
if log:
self.log = log
else:
self.log = logging.getLogger('eion')
self.type_conversion = {}
def log_dataframe(self, msg=None):
buffer = io.StringIO()
self.data.info(buf=buffer)
if msg:
log_text = f'Data frame after {msg}:'
else:
log_text = 'Data frame:'
log_text += '\n\t'+str(self.data.head(2)).replace('\n','\n\t')
log_text += ('\n\t' + buffer.getvalue().replace('\n','\n\t'))
self.log.info(log_text)
def transform(self):
if self.is_target_available():
if self.target_name:
self.log.info(f"Target feature name: '{self.target_name}'")
self.log.info(f"Target feature size: {len(self.target)}")
else:
self.log.info(f"Target feature not present")
self.log_dataframe()
print(self.data.info())
try:
self.process()
except Exception as e:
self.log.error(e, exc_info=True)
raise
pipe = FeatureUnion(self.pipeline)
try:
if self.text_feature:
from text.textProfiler import set_pretrained_model
set_pretrained_model(pipe)
conversion_method = self.get_conversion_method()
process_data = pipe.fit_transform(self.data, y=self.target)
# save for testing
if DEBUG_ENABLED:
if isinstance(process_data, scipy.sparse.spmatrix):
process_data = process_data.toarray()
df = pd.DataFrame(process_data)
df.to_csv('debug_preprocessed.csv', index=False)
if self.text_feature and conversion_method == 'latentsemanticanalysis':
n_size = self.get_tf_idf_output_size( pipe)
dimensions = self.get_tf_idf_dimensions()
if n_size < dimensions or n_size > dimensions:
dimensions = n_size
from sklearn.decomposition import TruncatedSVD
reducer = TruncatedSVD( n_components = dimensions)
reduced_data = reducer.fit_transform( process_data[:,-n_size:])
text_process_idx = [t[0] for t in pipe.transformer_list].index('text_process')
pipe.transformer_list[text_process_idx][1].steps.append(('feature_reducer',reducer))
if isinstance(process_data, scipy.sparse.spmatrix):
process_data = process_data.toarray()
process_data = np.concatenate((process_data[:,:-n_size], reduced_data), axis=1)
last_step = self.feature_expender.pop()
self.feature_expender.append({'feature_reducer':list(last_step.values())[0]})
except EOFError as e:
if "Compressed file ended before the end-of-stream marker was reached" in str(e):
raise EOFError('Pretrained model is not downloaded properly')
self.update_output_features_names(pipe)
if isinstance(process_data, scipy.sparse.spmatrix):
process_data = process_data.toarray()
df = pd.DataFrame(process_data, index=self.data.index, columns=self.output_columns)
if self.is_target_available() and self.target_name:
df[self.target_name] = self.target
if self.keep_unprocessed:
df[self.keep_unprocessed] = self.data[self.keep_unprocessed]
self.log_numerical_fill()
self.log_categorical_fill()
self.log_normalization()
return df, pipe, self.label_encoder
def log_type_conversion(self):
if self.log:
self.log.info('----------- Inspecting Features -----------')
self.log.info('----------- Type Conversion -----------')
count = 0
for k, v in self.type_conversion.items():
if v[0] != v[1]:
self.log.info(f'-------> {k} -> from {v[0]} to {v[1]} : {v[2]}')
self.log.info('Status:- |... Feature inspection done')
def check_config(self):
removeDuplicate = self.config.get('removeDuplicate', False)
self.config['removeDuplicate'] = cs.get_boolean(removeDuplicate)
self.config['misValueRatio'] = float(self.config.get('misValueRatio', cs.default_config['misValueRatio']))
self.config['numericFeatureRatio'] = float(self.config.get('numericFeatureRatio', cs.default_config['numericFeatureRatio']))
self.config['categoryMaxLabel'] = int(self.config.get('categoryMaxLabel', cs.default_config['categoryMaxLabel']))
featureDict = self.config.get('featureDict', [])
if isinstance(featureDict, dict):
self.config['featureDict'] = []
if isinstance(featureDict, str):
self.config['featureDict'] = []
def process(self):
#remove duplicate not required at the time of prediction
self.check_config()
self.remove_constant_feature()
self.remove_empty_feature(self.config['misValueRatio'])
self.remove_index_features()
self.dropna()
if self.config['removeDuplicate']:
self.drop_duplicate()
#self.check_categorical_features()
#self.string_to_numeric()
self.process_target()
self.train_features_type = {k:v for k,v in zip(self.data.columns, self.data.dtypes)}
self.parse_process_step_config()
self.process_drop_fillna()
self.log_type_conversion()
self.update_num_fill_dict()
if DEBUG_ENABLED:
print(self.num_fill_method_dict)
self.update_cat_fill_dict()
self.create_pipeline()
self.text_pipeline(self.config)
self.apply_outlier()
if DEBUG_ENABLED:
self.log.info(self.process_method)
self.log.info(self.pipeline)
def is_target_available(self):
return (isinstance(self.target, pd.Series) and not self.target.empty) or len(self.target)
def process_target(self, operation='encode', arg=None):
if self.is_target_available():
# drop null values
self.__update_index( self.target.notna(), 'target')
if self.encode_target:
self.label_encoder = LabelEncoder()
self.target = self.label_encoder.fit_transform(self.target)
return self.label_encoder
return None
def is_target_column(self, column):
return column == self.target_name
def fill_default_steps(self):
num_fill_method = cs.get_one_true_option(self.config.get('numericalFillMethod',{}))
normalization_method = cs.get_one_true_option(self.config.get('normalization',{}),'none')
for colm in self.numeric_feature:
if num_fill_method:
self.fill_missing_value_method(colm, num_fill_method.lower())
if normalization_method:
self.fill_normalizer_method(colm, normalization_method.lower())
cat_fill_method = cs.get_one_true_option(self.config.get('categoricalFillMethod',{}))
cat_encode_method = cs.get_one_true_option(self.config.get('categoryEncoding',{}))
for colm in self.cat_feature:
if cat_fill_method:
self.fill_missing_value_method(colm, cat_fill_method.lower())
if cat_encode_method:
self.fill_encoder_value_method(colm, cat_encode_method.lower(), default=True)
def parse_process_step_config(self):
self.process_method = {}
user_provided_data_type = {}
for feat_conf in self.featureDict:
colm = feat_conf.get('feature', '')
if not self.is_target_column(colm):
if colm in self.data.columns:
user_provided_data_type[colm] = feat_conf['type']
if user_provided_data_type:
self.update_user_provided_type(user_provided_data_type)
self.fill_default_steps()
for feat_conf in self.featureDict:
colm = feat_conf.get('feature', '')
if not self.is_target_column(colm):
if colm in self.data.columns:
if feat_conf.get('fillMethod', None):
self.fill_missing_value_method(colm, feat_conf['fillMethod'].lower())
if feat_conf.get('categoryEncoding', None):
self.fill_encoder_value_method(colm, feat_conf['categoryEncoding'].lower())
if feat_conf.get('normalization', None):
self.fill_normalizer_method(colm, feat_conf['normalization'].lower())
if feat_conf.get('outlier', None):
self.fill_outlier_method(colm, feat_conf['outlier'].lower())
if feat_conf.get('outlierOperation', None):
self.fill_outlier_process(colm, feat_conf['outlierOperation'].lower())
def get_tf_idf_dimensions(self):
dim = cs.get_one_true_option(self.config.get('embeddingSize',{}).get('TF_IDF',{}), 'default')
return {'default': 300, '50d':50, '100d':100, '200d':200, '300d':300}[dim]
def get_tf_idf_output_size(self, pipe):
start_index = {}
for feat_expender in self.feature_expender:
if feat_expender:
step_name = list(feat_expender.keys())[0]
index = list(feat_expender.values())[0]
for transformer_step in pipe.transformer_list:
if transformer_step[1].steps[-1][0] in step_name:
start_index[index] = {transformer_step[1].steps[-1][0]: transformer_step[1].steps[-1][1].get_feature_names_out()}
if start_index:
for key,value in start_index.items():
for k,v in value.items():
if k == 'vectorizer':
return len(v)
return 0
def update_output_features_names(self, pipe):
columns = self.output_columns
start_index = {}
index_shifter = 0
for feat_expender in self.feature_expender:
if feat_expender:
step_name = list(feat_expender.keys())[0]
for key,value in start_index.items():
for k,v in value.items():
index_shifter += len(v)
index = list(feat_expender.values())[0]
for transformer_step in pipe.transformer_list:
if transformer_step[1].steps[-1][0] in step_name:
start_index[index + index_shifter] = {transformer_step[1].steps[-1][0]: transformer_step[1].steps[-1][1].get_feature_names_out()}
#print(start_index)
if start_index:
for key,value in start_index.items():
for k,v in value.items():
if k == 'vectorizer':
v = [f'{x}_vect' for x in v]
self.output_columns[key:key] = v
self.added_features = [*self.added_features, *v]
def text_pipeline(self, conf_json):
if self.text_feature:
from text.textProfiler import textProfiler
from text.textProfiler import textCombine
pipeList = []
text_pipe = Pipeline([
('selector', ColumnTransformer([
("selector", "passthrough", self.text_feature)
], remainder="drop")),
("text_fillNa",SimpleImputer(strategy='constant', fill_value='')),
("merge_text_feature", textCombine())])
obj = textProfiler()
pipeList = obj.textProfiler(conf_json, pipeList, self.data_path)
last_step = "merge_text_feature"
for pipe_elem in pipeList:
text_pipe.steps.append((pipe_elem[0], pipe_elem[1]))
last_step = pipe_elem[0]
text_transformer = ('text_process', text_pipe)
self.pipeline.append(text_transformer)
self.feature_expender.append({last_step:len(self.output_columns)})
def create_pipeline(self):
num_pipe = {}
for k,v in self.num_fill_method_dict.items():
for k1,v1 in v.items():
if k1 and k1 != 'none':
num_pipe[f'{k}_{k1}'] = Pipeline([
('selector', ColumnTransformer([
("selector", "passthrough", v1)
], remainder="drop")),
(k, self.get_num_imputer(k)),
(k1, self.get_num_scaler(k1))
])
else:
num_pipe[f'{k}_{k1}'] = Pipeline([
('selector', ColumnTransformer([
("selector", "passthrough", v1)
], remainder="drop")),
(k, self.get_num_imputer(k))
])
self.output_columns.extend(v1)
cat_pipe = {}
for k,v in self.cat_fill_method_dict.items():
for k1,v1 in v.items():
cat_pipe[f'{k}_{k1}'] = Pipeline([
('selector', ColumnTransformer([
("selector", "passthrough", v1)
], remainder="drop")),
(k, self.get_cat_imputer(k)),
(k1, self.get_cat_encoder(k1))
])
if k1 not in ['onehotencoding']:
self.output_columns.extend(v1)
else:
self.feature_expender.append({k1:len(self.output_columns)})
for key, pipe in num_pipe.items():
self.pipeline.append((key, pipe))
for key, pipe in cat_pipe.items():
self.pipeline.append((key, pipe))
"Drop: feature during training but replace with zero during prediction "
def process_drop_fillna(self):
drop_column = []
if 'numFill' in self.process_method.keys():
for col, method in self.process_method['numFill'].items():
if method == 'drop':
self.process_method['numFill'][col] = 'zero'
drop_column.append(col)
if 'catFill' in self.process_method.keys():
for col, method in self.process_method['catFill'].items():
if method == 'drop':
self.process_method['catFill'][col] = 'zero'
drop_column.append(col)
if drop_column:
self.data.dropna(subset=drop_column, inplace=True)
def update_num_fill_dict(self):
self.num_fill_method_dict = {}
if 'numFill' in self.process_method.keys():
for f in cs.supported_method['fillNa']['numeric']:
self.num_fill_method_dict[f] = {}
for en in cs.supported_method['normalization']:
self.num_fill_method_dict[f][en] = []
for col in self.numeric_feature:
numFillDict = self.process_method.get('numFill',{})
normalizationDict = self.process_method.get('normalization',{})
if f == numFillDict.get(col, '') and en == normalizationDict.get(col,''):
self.num_fill_method_dict[f][en].append(col)
if not self.num_fill_method_dict[f][en] :
del self.num_fill_method_dict[f][en]
if not self.num_fill_method_dict[f]:
del self.num_fill_method_dict[f]
def update_cat_fill_dict(self):
self.cat_fill_method_dict = {}
if 'catFill' in self.process_method.keys():
for f in cs.supported_method['fillNa']['categorical']:
self.cat_fill_method_dict[f] = {}
for en in cs.supported_method['categoryEncoding']:
self.cat_fill_method_dict[f][en] = []
for col in self.cat_feature:
catFillDict = self.process_method.get('catFill',{})
catEncoderDict = self.process_method.get('catEncoder',{})
if f == catFillDict.get(col, '') and en == catEncoderDict.get(col,''):
self.cat_fill_method_dict[f][en].append(col)
if not self.cat_fill_method_dict[f][en] :
del self.cat_fill_method_dict[f][en]
if not self.cat_fill_method_dict[f]:
del self.cat_fill_method_dict[f]
def __update_type(self):
self.numeric_feature = list( set(self.data.select_dtypes(include='number').columns.tolist()) - set(self.keep_unprocessed))
self.cat_feature = list( set(self.data.select_dtypes(include='category').columns.tolist()) - set(self.keep_unprocessed))
self.text_feature = list( set(self.data.select_dtypes(include='object').columns.tolist()) - set(self.keep_unprocessed))
self.datetime_feature = list( set(self.data.select_dtypes(include='datetime').columns.tolist()) - set(self.keep_unprocessed))
def update_user_provided_type(self, data_types):
allowed_types = ['numerical','categorical', 'text']
skipped_types = ['date','index']
type_mapping = {'numerical': np.dtype('float'), 'float': np.dtype('float'),'categorical': 'category', 'text':np.dtype('object'),'date':'datetime64[ns]','index': np.dtype('int64'),}
mapped_type = {k:type_mapping[v] for k,v in data_types.items() if v in allowed_types}
skipped_features = [k for k,v in data_types.items() if v in skipped_types]
if skipped_features:
self.keep_unprocessed.extend( skipped_features)
self.keep_unprocessed = list(set(self.keep_unprocessed))
self.update_type(mapped_type, 'user provided data type')
def get_type(self, as_list=False):
if as_list:
return [self.colm_type.values()]
else:
return self.colm_type
def update_type(self, data_types={}, reason=''):
invalid_features = [x for x in data_types.keys() if x not in self.data.columns]
if invalid_features:
valid_feat = list(set(data_types.keys()) - set(invalid_features))
valid_feat_type = {k:v for k,v in data_types if k in valid_feat}
else:
valid_feat_type = data_types
for k,v in valid_feat_type.items():
if v != self.colm_type[k].name:
try:
self.data.astype({k:v})
self.colm_type.update({k:self.data[k].dtype})
self.type_conversion[k] = (self.colm_type[k] , v, 'Done', reason)
except:
self.type_conversion[k] = (self.colm_type[k] , v, 'Fail', reason)
if v == np.dtype('float64') and self.colm_type[k].name == 'object':
if self.check_numeric( k):
self.data[ k] = pd.to_numeric(self.data[ k], errors='coerce')
self.type_conversion[k] = (self.colm_type[k] , v, 'Done', reason)
self.force_numeric_conv.append( k)
else:
raise ValueError(f"Can not convert '{k}' feature to 'numeric' as numeric values are less than {self.config['numericFeatureRatio'] * 100}%")
self.data = self.data.astype(valid_feat_type)
self.__update_type()
def check_numeric(self, feature):
col_values = self.data[feature].copy()
col_values = pd.to_numeric(col_values, errors='coerce')
if col_values.count() >= (self.config['numericFeatureRatio'] * len(col_values)):
return True
return False
def string_to_numeric(self):
def to_number(x):
try:
return w2n.word_to_num(x)
except:
return np.nan
for col in self.text_feature:
col_values = self.data[col].copy()
col_values = pd.to_numeric(col_values, errors='coerce')
if col_values.count() >= (self.config['numericFeatureRatio'] * len(col_values)):
self.text_to_num[col] = 'float64'
self.wordToNumericFeatures.append(col)
if self.text_to_num:
columns = list(self.text_to_num.keys())
self.data[columns] = self.data[columns].apply(lambda x: to_number(x), axis=1, result_type='broadcast')
self.update_type(self.text_to_num)
self.log.info('----------- Inspecting Features -----------')
for col in self.text_feature:
self.log.info(f'-------> Feature : {col}')
if col in self.text_to_num:
self.log.info('----------> Numeric Status :Yes')
self.log.info('----------> Data Type Converting to numeric :Yes')
else:
self.log.info('----------> Numeric Status :No')
self.log.info(f'\nStatus:- |... Feature inspection done for numeric data: {len(self.text_to_num)} feature(s) converted to numeric')
self.log.info(f'\nStatus:- |... Feature word to numeric treatment done: {self.text_to_num}')
self.log.info('----------- Inspecting Features End -----------')
def check_categorical_features(self):
num_data = self.data.select_dtypes(include='number')
num_data_unique = num_data.nunique()
num_to_cat_col = {}
for i, value in enumerate(num_data_unique):
if value < self.config['categoryMaxLabel']:
num_to_cat_col[num_data_unique.index[i]] = 'category'
if num_to_cat_col:
self.update_type(num_to_cat_col, 'numerical to categorical')
str_to_cat_col = {}
str_data = self.data.select_dtypes(include='object')
str_data_unique = str_data.nunique()
for i, value in enumerate(str_data_unique):
if value < self.config['categoryMaxLabel']:
str_to_cat_col[str_data_unique.index[i]] = 'category'
for colm in str_data.columns:
if self.data[colm].str.len().max() < cs.default_config['str_to_cat_len_max']:
str_to_cat_col[colm] = 'category'
if str_to_cat_col:
self.update_type(str_to_cat_col, 'text to categorical')
def drop_features(self, features=[], reason='unspecified'):
if isinstance(features, str):
features = [features]
feat_to_remove = [x for x in features if x in self.data.columns]
if feat_to_remove:
self.data.drop(feat_to_remove, axis=1, inplace=True)
for feat in feat_to_remove:
self.dropped_features[feat] = reason
self.log_drop_feature(feat_to_remove, reason)
self.__update_type()
def __update_index(self, indices, reason=''):
if isinstance(indices, (bool, pd.core.series.Series)) and len(indices) == len(self.data):
if not indices.all():
self.data = self.data[indices]
if self.is_target_available():
self.target = self.target[indices]
self.log_update_index((indices == False).sum(), reason)
def dropna(self):
self.data.dropna(how='all',inplace=True)
if self.is_target_available():
self.target = self.target[self.data.index]
def drop_duplicate(self):
index = self.data.duplicated(keep='first')
self.__update_index( ~index, reason='duplicate')
def log_drop_feature(self, columns, reason):
self.log.info(f'---------- Dropping {reason} features ----------')
self.log.info(f'\nStatus:- |... {reason} feature treatment done: {len(columns)} {reason} feature(s) found')
self.log.info(f'-------> Drop Features: {columns}')
self.log.info(f'Data Frame Shape After Dropping (Rows,Columns): {self.data.shape}')
def log_update_index(self,count, reason):
if count:
if reason == 'target':
self.log.info('-------> Null Target Rows Drop:')
self.log.info(f'-------> Dropped rows count: {count}')
elif reason == 'duplicate':
self.log.info('-------> Duplicate Rows Drop:')
self.log.info(f'-------> Dropped rows count: {count}')
elif reason == 'outlier':
self.log.info(f'-------> Dropped rows count: {count}')
self.log.info('Status:- |... Outlier treatment done')
self.log.info(f'-------> Data Frame Shape After Dropping samples(Rows,Columns): {self.data.shape}')
def log_normalization(self):
if self.process_method.get('normalization', None):
self.log.info(f'\nStatus:- !... Normalization treatment done')
for method in cs.supported_method['normalization']:
cols = []
for col, m in self.process_method['normalization'].items():
if m == method:
cols.append(col)
if cols and method != 'none':
self.log.info(f'Running {method} on features: {cols}')
def log_numerical_fill(self):
if self.process_method.get('numFill', None):
self.log.info(f'\nStatus:- !... Fillna for numeric feature done')
for method in cs.supported_method['fillNa']['numeric']:
cols = []
for col, m in self.process_method['numFill'].items():
if m == method:
cols.append(col)
if cols:
self.log.info(f'-------> Running {method} on features: {cols}')
def log_categorical_fill(self):
if self.process_method.get('catFill', None):
self.log.info(f'\nStatus:- !... FillNa for categorical feature done')
for method in cs.supported_method['fillNa']['categorical']:
cols = []
for col, m in self.process_method['catFill'].items():
if m == method:
cols.append(col)
if cols:
self.log.info(f'-------> Running {method} on features: {cols}')
def remove_constant_feature(self):
unique_values = self.data.nunique()
constant_features = []
for i, value in enumerate(unique_values):
if value == 1:
constant_features.append(unique_values.index[i])
if constant_features:
self.drop_features(constant_features, "constant")
def remove_empty_feature(self, misval_ratio=1.0):
missing_ratio = self.data.isnull().sum() / len(self.data)
missing_ratio = {k:v for k,v in zip(self.data.columns, missing_ratio)}
empty_features = [k for k,v in missing_ratio.items() if v > misval_ratio]
if empty_features:
self.drop_features(empty_features, "empty")
def remove_index_features(self):
index_feature = []
for feat in self.numeric_feature:
if self.data[feat].nunique() == len(self.data):
#if (self.data[feat].sum()- sum(self.data.index) == (self.data.iloc[0][feat]-self.data.index[0])*len(self.data)):
# index feature can be time based
count = (self.data[feat] - self.data[feat].shift() == 1).sum()
if len(self.data) - count == 1:
index_feature.append(feat)
self.drop_features(index_feature, "index")
def fill_missing_value_method(self, colm, method):
if colm in self.numeric_feature:
if method in cs.supported_method['fillNa']['numeric']:
if 'numFill' not in self.process_method.keys():
self.process_method['numFill'] = {}
if method == 'na' and self.process_method['numFill'].get(colm, None):
pass # don't overwrite
else:
self.process_method['numFill'][colm] = method
if colm in self.cat_feature:
if method in cs.supported_method['fillNa']['categorical']:
if 'catFill' not in self.process_method.keys():
self.process_method['catFill'] = {}
if method == 'na' and self.process_method['catFill'].get(colm, None):
pass
else:
self.process_method['catFill'][colm] = method
def check_encoding_method(self, method, colm,default=False):
if not self.is_target_available() and (method.lower() == list(cs.target_encoding_method_change.keys())[0]):
method = cs.target_encoding_method_change[method.lower()]
if default:
self.log.info(f"Applying Label encoding instead of Target encoding on feature '{colm}' as target feature is not present")
return method
def fill_encoder_value_method(self,colm, method, default=False):
if colm in self.cat_feature:
if method.lower() in cs.supported_method['categoryEncoding']:
if 'catEncoder' not in self.process_method.keys():
self.process_method['catEncoder'] = {}
if method == 'na' and self.process_method['catEncoder'].get(colm, None):
pass
else:
self.process_method['catEncoder'][colm] = self.check_encoding_method(method, colm,default)
else:
self.log.info(f"-------> categorical encoding method '{method}' is not supported. supported methods are {cs.supported_method['categoryEncoding']}")
def fill_normalizer_method(self,colm, method):
if colm in self.numeric_feature:
if method in cs.supported_method['normalization']:
if 'normalization' not in self.process_method.keys():
self.process_method['normalization'] = {}
if (method == 'na' or method == 'none') and self.process_method['normalization'].get(colm, None):
pass
else:
self.process_method['normalization'][colm] = method
else:
self.log.info(f"-------> Normalization method '{method}' is not supported. supported methods are {cs.supported_method['normalization']}")
def apply_outlier(self):
inlier_indice = np.array([True] * len(self.data))
if self.process_method.get('outlier', None):
self.log.info('-------> Feature wise outlier detection:')
for k,v in self.process_method['outlier'].items():
if k in self.numeric_feature:
if v == 'iqr':
index = cs.findiqrOutlier(self.data[k])
elif v == 'zscore':
index = cs.findzscoreOutlier(self.data[k])
elif v == 'disable':
index = None
if k in self.process_method['outlierOperation'].keys():
if self.process_method['outlierOperation'][k] == 'dropdata':
inlier_indice = np.logical_and(inlier_indice, index)
elif self.process_method['outlierOperation'][k] == 'average':
mean = self.data[k].mean()
index = ~index
self.data.loc[index,[k]] = mean
self.log.info(f'-------> {k}: Replaced by Mean {mean}: total replacement {index.sum()}')
elif self.process_method['outlierOperation'][k] == 'nochange' and v != 'disable':
self.log.info(f'-------> Total outliers in "{k}": {(~index).sum()}')
if self.config.get('outlierDetection',None):
if self.config['outlierDetection'].get('IsolationForest','False') == 'True':
if self.numeric_feature:
index = cs.findiforestOutlier(self.data[self.numeric_feature])
inlier_indice = np.logical_and(inlier_indice, index)
self.log.info(f'-------> Numeric feature based Outlier detection(IsolationForest):')
if inlier_indice.sum() != len(self.data):
self.__update_index(inlier_indice, 'outlier')
def fill_outlier_method(self,colm, method):
if colm in self.numeric_feature:
if method in cs.supported_method['outlier_column_wise']:
if 'outlier' not in self.process_method.keys():
self.process_method['outlier'] = {}
if method not in ['Disable', 'na']:
self.process_method['outlier'][colm] = method
else:
self.log.info(f"-------> outlier detection method '{method}' is not supported for column wise. supported methods are {cs.supported_method['outlier_column_wise']}")
def fill_outlier_process(self,colm, method):
if colm in self.numeric_feature:
if method in cs.supported_method['outlierOperation']:
if 'outlierOperation' not in self.process_method.keys():
self.process_method['outlierOperation'] = {}
self.process_method['outlierOperation'][colm] = method
else:
self.log.info(f"-------> outlier process method '{method}' is not supported for column wise. supported methods are {cs.supported_method['outlierOperation']}")
def get_cat_imputer(self,method):
if method == 'mode':
return SimpleImputer(strategy='most_frequent')
elif method == 'zero':
return SimpleImputer(strategy='constant', fill_value=0)
def get_cat_encoder(self,method):
if method == 'labelencoding':
return OrdinalEncoder()
elif method == 'onehotencoding':
return OneHotEncoder(sparse=False,handle_unknown="ignore")
elif method == 'targetencoding':
if not self.is_target_available():
raise ValueError('Can not apply Target Encoding when target feature is not present')
return TargetEncoder()
def get_num_imputer(self,method):
if method == 'mode':
return SimpleImputer(strategy='most_frequent')
elif method == 'mean':
return SimpleImputer(strategy='mean')
elif method == 'median':
return SimpleImputer(strategy='median')
elif method == 'knnimputer':
return KNNImputer()
elif method == 'zero':
return SimpleImputer(strategy='constant', fill_value=0)
def get_num_scaler(self,method):
if method == 'minmax':
return MinMaxScaler()
elif method == 'standardscaler':
return StandardScaler()
elif method == 'lognormal':
return PowerTransformer(method='yeo-johnson', standardize=False)
def recommenderStartProfiler(self,modelFeatures):
return cs.recommenderStartProfiler(self,modelFeatures)
def folderPreprocessing(self,folderlocation,folderdetails,deployLocation):
return cs.folderPreprocessing(self,folderlocation,folderdetails,deployLocation)
def textSimilarityStartProfiler(self, doc_col_1, doc_col_2):
return cs.textSimilarityStartProfiler(self, doc_col_1, doc_col_2)
def get_conversion_method(self):
return cs.get_one_true_option(self.config.get('textConversionMethod','')).lower()
def set_features(features,profiler=None):
return cs.set_features(features,profiler)
|
data_profiler_functions.py | import os
import sys
import numpy as np
import scipy
import pandas as pd
from pathlib import Path
default_config = {
'misValueRatio': '1.0',
'numericFeatureRatio': '1.0',
'categoryMaxLabel': '20',
'str_to_cat_len_max': 10
}
target_encoding_method_change = {'targetencoding': 'labelencoding'}
supported_method = {
'fillNa':
{
'categorical' : ['mode','zero','na'],
'numeric' : ['median','mean','knnimputer','zero','drop','na'],
},
'categoryEncoding': ['labelencoding','targetencoding','onehotencoding','na','none'],
'normalization': ['standardscaler','minmax','lognormal', 'na','none'],
'outlier_column_wise': ['iqr','zscore', 'disable', 'na'],
'outlierOperation': ['dropdata', 'average', 'nochange']
}
def findiqrOutlier(df):
Q1 = df.quantile(0.25)
Q3 = df.quantile(0.75)
IQR = Q3 - Q1
index = ~((df < (Q1 - 1.5 * IQR)) |(df > (Q3 + 1.5 * IQR)))
return index
def findzscoreOutlier(df):
z = np.abs(scipy.stats.zscore(df))
index = (z < 3)
return index
def findiforestOutlier(df):
from sklearn.ensemble import IsolationForest
isolation_forest = IsolationForest(n_estimators=100)
isolation_forest.fit(df)
y_pred_train = isolation_forest.predict(df)
return y_pred_train == 1
def get_one_true_option(d, default_value=None):
if isinstance(d, dict):
for k,v in d.items():
if (isinstance(v, str) and v.lower() == 'true') or (isinstance(v, bool) and v == True):
return k
return default_value
def get_boolean(value):
if (isinstance(value, str) and value.lower() == 'true') or (isinstance(value, bool) and value == True):
return True
else:
return False
def recommenderStartProfiler(self,modelFeatures):
try:
self.log.info('----------> FillNA:0')
self.data = self.data.fillna(value=0)
self.log.info('Status:- !... Missing value treatment done')
self.log.info('----------> Remove Empty Row')
self.data = self.data.dropna(axis=0,how='all')
self.log.info('Status:- !... Empty feature treatment done')
userId,itemId,rating = modelFeatures.split(',')
self.data[itemId] = self.data[itemId].astype(np.int32)
self.data[userId] = self.data[userId].astype(np.int32)
self.data[rating] = self.data[rating].astype(np.float32)
return self.data
except Exception as inst:
self.log.info("Error: dataProfiler failed "+str(inst))
return(self.data)
def folderPreprocessing(self,folderlocation,folderdetails,deployLocation):
try:
dataset_directory = Path(folderlocation)
dataset_csv_file = dataset_directory/folderdetails['label_csv_file_name']
tfrecord_directory = Path(deployLocation)/'Video_TFRecord'
from savp import PreprocessSAVP
import csv
csvfile = open(dataset_csv_file, newline='')
csv_reader = csv.DictReader(csvfile)
PreprocessSAVP(dataset_directory,csv_reader,tfrecord_directory)
dataColumns = list(self.data.columns)
VideoProcessing = True
return dataColumns,VideoProcessing,tfrecord_directory
except Exception as inst:
self.log.info("Error: dataProfiler failed "+str(inst))
def textSimilarityStartProfiler(self, doc_col_1, doc_col_2):
import os
try:
features = [doc_col_1, doc_col_2]
pipe = None
dataColumns = list(self.data.columns)
self.numofCols = self.data.shape[1]
self.numOfRows = self.data.shape[0]
from transformations.textProfiler import textProfiler
self.log.info('-------> Execute Fill NA With Empty String')
self.data = self.data.fillna(value=" ")
self.log.info('Status:- |... Missing value treatment done')
self.data[doc_col_1] = textProfiler().textCleaning(self.data[doc_col_1])
self.data[doc_col_2] = textProfiler().textCleaning(self.data[doc_col_2])
self.log.info('-------> Concatenate: ' + doc_col_1 + ' ' + doc_col_2)
self.data['text'] = self.data[[doc_col_1, doc_col_2]].apply(lambda row: ' '.join(row.values.astype(str)), axis=1)
from tensorflow.keras.preprocessing.text import Tokenizer
pipe = Tokenizer()
pipe.fit_on_texts(self.data['text'].values)
self.log.info('-------> Tokenizer: Fit on Concatenate Field')
self.log.info('Status:- |... Tokenizer the text')
self.data[doc_col_1] = self.data[doc_col_1].astype(str)
self.data[doc_col_1] = self.data[doc_col_1].astype(str)
return (self.data, pipe, self.target_name, features)
except Exception as inst:
self.log.info("StartProfiler failed " + str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1]
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
def set_features(features,profiler=None):
if profiler:
features = [x for x in features if x not in profiler.added_features]
return features + profiler.text_feature
return features |
aionfls.py | # -*- coding: utf-8 -*-
import pandas as pd
import numpy as np
import logging
import os
import sys
from flwr.common.logger import log
from logging import INFO
from flserver import flserver
class aionfls:
def __init__(self):
self.confdata=None
def configLoad(self,jsonfile):
import json
jsonfile=str(jsonfile)
with open(jsonfile, 'r') as file:
self.confdata = json.load(file)
return self.confdata
def dataload(self,datapath):
df = pd.read_csv(datapath) #chunk_size=50000
## Data preprocess in test dataset, In aion, aion profiler will handle it.
df =df[~df.isin([np.nan, np.inf, -np.inf]).any(axis=1)]
df=df.reset_index(drop=True)
return df
# Start Flower server for n rounds of federated learning
if __name__ == "__main__":
classobj=aionfls()
json_file=sys.argv[1]
confdata = classobj.configLoad(json_file)
data_location = confdata["data_location"]
# deploy_location=confdata['deploy_location']
cwd = os.path.abspath(os.path.dirname(__file__))
model_name=confdata['model_name']
version=str(confdata['version'])
file_name=model_name+'_'+version+".log"
try:
fl_log=os.path.normpath(os.path.join(cwd,'logs',file_name))
except Exception as e:
classobj.log.info("Log path error. Error Msg: \n",e)
logging.basicConfig(filename=fl_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
classobj.log = logging.getLogger('AION')
print("===============")
print("flserver main function")
print("===============")
if (confdata['evaluation_required'].lower() == 'false'):
df=None
aionflobj=flserver(df,confdata)
print("flserver condition is false")
else:
## User selected option is True
print("flserver condition is true")
data_location = os.path.normpath(os.path.join(cwd, data_location))
# print("current data_location",data_location)
df=classobj.dataload(data_location)
aionflobj=flserver(df,confdata)
status=aionflobj.runFLServer()
classobj.log.info("Aion FL Server run Status: \n"+str(status))
|
dl_model.py | import tensorflow as tf
def dl_regression_model(input_shape, output_shape,
optimizer, loss_func, act_func):
inputs = tf.keras.Input(shape=(input_shape,))
x = tf.keras.layers.Dense(64,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(inputs)
x = tf.keras.layers.Dense(32,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(16,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(8,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
outputs = tf.keras.layers.Dense(output_shape,
kernel_initializer='he_normal',
bias_initializer='zeros')(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
model.compile(loss=loss_func ,
optimizer=optimizer,
metrics=["mean_absolute_error",
"mean_squared_error",
])
return model
def dl_multiClass_classification(input_shape, output_shape,
optimizer, loss_func, act_func, last_act_func):
inputs = tf.keras.Input(shape=(input_shape,))
x = tf.keras.layers.Dense(64,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(inputs)
x = tf.keras.layers.Dense(32,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(16,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(8,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
outputs = tf.keras.layers.Dense(output_shape,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=last_act_func)(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
model.compile(optimizer, loss_func, metrics=["accuracy"])
return model
def dl_binary_classification(input_shape, output_shape,
optimizer, loss_func, act_func, last_act_func):
inputs = tf.keras.Input(shape=(input_shape,))
x = tf.keras.layers.Dense(64,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(inputs)
x = tf.keras.layers.Dense(32,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(16,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(8,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
outputs = tf.keras.layers.Dense(output_shape,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=last_act_func)(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
model.compile(optimizer, loss_func,
metrics=["accuracy"])
return model
|
flserver.py | # -*- coding: utf-8 -*-
"""
Created on Wed Jun 15 14:36:11 2022
@author: @aionteam
"""
import flwr
import flwr as fl
import tensorflow as tf
from typing import Any, Callable, Dict, List, Optional, Tuple
import utils
from sklearn.metrics import log_loss
from sklearn.linear_model import LogisticRegression
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error, mean_absolute_error,r2_score
from typing import Dict
import numpy as np
import logging
import os
# import sys
from flwr.common.logger import log
from logging import INFO
import pickle as pkl
from flwr.server.client_proxy import ClientProxy
import dl_model
from sklearn.preprocessing import StandardScaler
import pandas as pd
## Below import can be used when aion specific grpc communication used.
# from aionflgrpcserver import aionflgrpcserver
# Make TensorFlow logs less verbose
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
os.environ["GRPC_VERBOSITY"] = "debug"
# """ AION Federated Learning Server. Geting weights from clients, aggregate the weights by FedAvg algorithm and update the client model."""
class flserver():
def __init__(self,df,confdata):
self.log = logging.getLogger('AION')
## Below params will be used later
self.confparams=confdata
self.df=df
self.fl_round=0
print("Inside flserver init func")
## Flower server number of rounds for fl model update (clients-server)
def fit_round(self,rnd: int) -> Dict:
"""Send round number to client."""
self.fl_round=rnd
log(INFO, "===========================")
self.log.info("federated learning round: "+str(rnd))
log(INFO, "federated learning round: %s ",str(rnd))
log(INFO, "===========================")
# print(f"federated learning round: {rnd}")
return {"rnd": rnd}
def fit_config(self,rnd: int) -> Dict:
"""Send round number to client."""
self.round_count = rnd
log(INFO, "===========================")
log(INFO, "Starting round %s ...",str(rnd))
log(INFO, "===========================")
model_hyperparams = self.confparams["model_hyperparams"]
batch_size = model_hyperparams["batch_size"]
local_epochs = model_hyperparams["epochs"]
config = {
"batch_size": int(batch_size),
# "local_epochs": 1 if rnd < 2 else 2,
"local_epochs": int(local_epochs),
"rnd": rnd,
}
return config
def evaluate_config(self, rnd: int):
model_hyperparams = self.confparams["model_hyperparams"]
val_batch_size = model_hyperparams["batch_size"]
# val_steps = 5 if rnd < 4 else 10
return {"val_batch_size": int(val_batch_size)}
## Loading configuration parameters
def configload(self,confparams):
try:
data=confparams
server_ip=str(data["server_IP"])
server_port=str(data["server_port"])
model_name=str(data["model_name"])
num_clients=int(data["min_available_clients"])
min_fit_clients=int(data["min_fit_clients"])
num_train_round=int(data["fl_round"])
data_location=str(data["data_location"])
model_params=data["model_hyperparams"]
problem_type=data["problem_type"]
server_address=f"{server_ip}:{server_port}"
# model_location=str(data["model_store"])
model_version=str(data["version"])
selected_feature=data["selected_feature"]
if (type(selected_feature) is str):
selected_feature=selected_feature.split(',')
target_feature=data["target_feature"]
evaluation_required=data["evaluation_required"]
self.log.info("Federated Learning <Server IP:Port> "+str(server_address))
except Exception as e:
log(INFO, "Reading server config file issue. Err.Msg: %s ",str(e))
return server_address,model_name,num_clients,min_fit_clients,num_train_round,data_location,model_params,problem_type,model_version,selected_feature,target_feature,evaluation_required
## Save the final model
def model_save(self,model,model_name,problem_type,version):
cwd = os.path.abspath(os.path.dirname(__file__))
model_location=os.path.join(cwd, 'models')
model_name=model_name
version=str(version)
model_name=self.confparams["model_name"]
if (model_name.lower() == "deeplearning"):
file_name = model_name + '_' +problem_type+'_'+version+ ".h5"
else:
file_name=file_name = model_name + '_' +problem_type+'_'+version+".sav"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
self.log.info("saved_model path: "+str(saved_model))
try:
with open (saved_model,'wb') as f:
pkl.dump(model,f)
return True
except Exception as e:
self.log.info("fl server model save error. Error Msg: "+str(e))
return False
## Load the model, not used now. If user want to use aion trained model for evaluation at serverside, use this fn.
def model_load(self, path):
model_name=self.confparams["model_name"]
if (model_name.lower() == "deeplearning"):
loaded_model = tf.keras.models.load_model(path)
else:
loaded_model = pkl.load(open(path, 'rb'))
return loaded_model
# Fo normal ml models, def get_eval_fn, evaluate each round results with own dataset. It is optional, without this, fed server will aggregate (fedAvg) client weights and update results to clients without evaluate.
def get_eval_fn(self,model,X,y,model_name,model_version):
"""Return an evaluation function for server-side evaluation."""
self.log.info("X_eval: \n"+str(X.shape))
self.log.info("y_eval: \n"+str(y.shape))
# scaler = StandardScaler()
# X_scaled = scaler.fit_transform(X)
# y = pd.get_dummies(y)
# y_class = None
def evaluate(server_round: int,
parameters: fl.common.NDArrays,
config: Dict[str, fl.common.Scalar],):
# self.log.info("server side fedavg weights \n "+str(parameters))
try:
problem_type=self.confparams["problem_type"]
# if (self.model_name.lower() == 'logisticregression' ):
# loss = log_loss(y, model.predict_proba(X))
# else:
# loss = log_loss(y, model.predict(X))
if (problem_type.lower() == 'classification'):
if (model_name.lower() == 'logisticregression' ):
utils.set_model_params(model, parameters)
loss = log_loss(y, model.predict_proba(X))
# loss = log_loss(y, model.predict_proba(X))
accuracy = model.score(X, y)
log(INFO, "Server evaluation FL Round: %s processed Weights. -- Loss: %s, -- Accuracy: %s ",str(self.fl_round),str(loss), str(accuracy))
self.log.info("Accuracy: "+str(accuracy))
self.log.info("model coefficients: "+str(model.coef_))
self.log.info("model intercept: "+str(model.intercept_))
problem_type=self.confparams["problem_type"]
self.model_save(model,model_name,problem_type,model_version)
return loss, {"accuracy": accuracy}
else:
if (model_name.lower() == 'linearregression' ):
print(model, type(model))
print(model.get_params)
# rmse = mean_squared_error(y, model.predict(X), square=True)
rmse = np.sqrt(mean_squared_error(y, model.predict(X)))
mae = mean_absolute_error(y, model.predict(X))
r2=r2_score(y, model.predict(X))
loss = rmse
mse=mean_squared_error(y, model.predict(X))
rmse = np.sqrt(mean_squared_error(y, model.predict(X)))
mae = mean_absolute_error(y, model.predict(X))
r2=r2_score(y, model.predict(X))
loss = rmse
results = {
"mean_absolute_error": mae,
"mean_squared_error": mse,
"root_mean_squared_error": rmse,
"r2":r2,
}
# accuracy=r2
log(INFO, "Server evaluation FL Round: %s processed Weights. -- Loss: %s, -- metrics: %s ",str(self.fl_round),str(rmse), str(results))
self.log.info("model coefficients: "+str(model.coef_))
self.log.info("model intercept: "+str(model.intercept_))
self.model_save(model,model_name,problem_type,model_version)
# return loss, len(X), results
return loss, results
except Exception as e:
log(INFO, "evaluate error msg: %s ",str(e))
return evaluate
# for deep learn models, def get_eval_fn, evaluate each round results with own dataset. It is optional, without this, fed server will aggregate (fedAvg) client weights and update results to clients without evaluate.
def get_eval_fn_dl(self, model,X,y,model_name,model_version):
try:
scaler = StandardScaler()
X_scaled = scaler.fit_transform(X)
# y = pd.get_dummies(y)
y_class = None
def evaluate(
server_round: int,
weights: fl.common.NDArrays,
config: Dict[str, fl.common.Scalar],
) -> Optional[Tuple[float, Dict[str, fl.common.Scalar]]]:
# Update model with the latest parameters
model.set_weights(weights)
problem_type = self.confparams["problem_type"]
self.model_save(model, model_name,problem_type, model_version)
if model_name == 'deeplearning':
if problem_type == 'regression':
loss, mean_absolute_error, mean_squared_error = model.evaluate(X_scaled,
y,
verbose=1)
y_pred = model.predict(X_scaled)
from sklearn import metrics
root_mean_squared_error = np.sqrt(metrics.mean_squared_error(y, y_pred))
log(INFO, "global model mean_absolute_error: %f ",mean_absolute_error)
log(INFO, "global model mean_squared_error: %f ",mean_squared_error)
log(INFO, "global model root_mean_squared_error: %f ",root_mean_squared_error)
return loss, {"mean_absolute_error": mean_absolute_error,
"mean_squared_error": mean_squared_error,
"root_mean_squared_error": root_mean_squared_error}
if problem_type == 'classification':
y_class = pd.get_dummies(y)
loss, accuracy = model.evaluate(X_scaled, y_class, verbose=1)
log(INFO, "global model accuracy: %f ",round(accuracy * 100, 2))
log(INFO, "global model loss: %f ", round(loss, 2))
return loss, {"accuracy": accuracy}
except Exception as e:
log(INFO, "get_eval_fn_dl error: %s ",str(e))
return evaluate
""" Below part is the aion specific grpc functions. To start the grpc server and client. Currently below modules are not used. """
# def callaiongrpcserver(self):
# agrpcobj = aionflgrpcserver()
# status=agrpcobj.startgrpcerver()
# print("server grpc start status: \t",status)
# return status
# def stopaiongrpcserver(self):
# agrpcobj = aionflgrpcserver()
# status=agrpcobj.shutserver()
# print("server grpc stop status: \t",status)
# return status
## This function called from aionflmain.py, and run server.
## Getting flower fl strategy
def get_strategy(self,min_fit_clients,num_clients,eval_fn,on_fit_config_fn,on_evaluate_config_fn):
strategy = fl.server.strategy.FedAvg(
min_fit_clients=min_fit_clients,
min_available_clients=num_clients,
evaluate_fn=eval_fn,
on_fit_config_fn=on_fit_config_fn,
on_evaluate_config_fn=on_evaluate_config_fn,
# initial_parameters=fl.common.weights_to_parameters(model.get_weights()),
)
return strategy
def runFLServer(self):
try:
server_address,model_name,num_clients,min_fit_clients,num_train_round,data_location,model_params,problem_type,model_version,selected_feature,target_feature,evaluation_required = self.configload(self.confparams)
df = self.df
if (evaluation_required.lower() == 'true'):
## One more check for NaN,Inf occurance in dataframe
df =df[~df.isin([np.nan, np.inf, -np.inf]).any(axis=1)]
## Remove index if passed.
df=df.reset_index(drop=True)
y=df[target_feature]
X=df[selected_feature]
if (problem_type.lower() == "classification"):
if (model_name.lower() == "logisticregression"):
#n_classes = df[target_feature].nunique()
no_classes = len(df.groupby(target_feature).count())
no_features=len(selected_feature)
self.log.info("no_classes: "+str(no_classes))
self.log.info("no_features: "+str(no_features))
modelName="logisticregression"
try:
model = LogisticRegression(**model_params, warm_start=True)
except Exception as e:
self.log.info("LR model error: \n"+str(e))
status=utils.setmodelName(modelName)
utils.set_initial_params(model,no_classes,no_features)
eval_fn=self.get_eval_fn(model,X,y,model_name,model_version)
on_fit_config_fn=self.fit_round
on_evaluate_config_fn=None
min_fit_clients=2
strategy = self.get_strategy(min_fit_clients,num_clients,eval_fn,on_fit_config_fn,on_evaluate_config_fn)
elif (model_name.lower() == "deeplearning"):
# model_hyperparams = self.confparams["model_hyperparams"]
optimizer = model_params["optimizer"]
loss_func = model_params["losses"]
act_func = model_params["activation"]
last_act_func = model_params["last_activation"]
input_shape = X.shape[1] # len(selected_feature)
output_shape = len(y.value_counts())
model = None
if output_shape == 2:
if last_act_func == "sigmoid" and loss_func == "binary_crossentropy":
model = dl_model.dl_binary_classification(input_shape, output_shape,
optimizer, loss_func,
act_func, last_act_func)
elif last_act_func == "softmax" and loss_func == "categorical_crossentropy":
model = dl_model.dl_binary_classification(input_shape, output_shape,
optimizer, loss_func,
act_func, last_act_func)
else:
model = dl_model.dl_multiClass_classification(input_shape,
output_shape, optimizer, loss_func,
act_func, last_act_func)
print(model.summary())
eval_fn=self.get_eval_fn_dl(model,X,y,model_name,model_version)
on_fit_config_fn=self.fit_config
on_evaluate_config_fn=self.evaluate_config
strategy = self.get_strategy(min_fit_clients,num_clients,eval_fn,on_fit_config_fn,on_evaluate_config_fn)
elif(problem_type.lower() == "regression"):
if (model_name.lower() == "linearregression"):
model=LinearRegression(**model_params)
status=utils.setmodelName(model_name)
utils.set_initial_params_reg(model,X.shape[0],len(selected_feature))
# utils.set_initial_params_reg(model,X.shape[0],X.shape[1])
eval_fn=self.get_eval_fn(model,X,y,model_name,model_version)
on_fit_config_fn=self.fit_round
on_evaluate_config_fn=None
min_fit_clients=2
strategy = self.get_strategy(min_fit_clients,num_clients,eval_fn,on_fit_config_fn,on_evaluate_config_fn)
elif(model_name.lower() == "deeplearning"):
input_shape = X.shape[1] # len(selected_feature)
output_shape = len(y.value_counts())
optimizer = model_params["optimizer"]
loss_func = model_params["losses"]
act_func = model_params["activation"]
model = None
model = dl_model.dl_regression_model(input_shape, 1,
optimizer, loss_func, act_func)
eval_fn=self.get_eval_fn_dl(model,X,y,model_name,model_version)
on_fit_config_fn=self.fit_config
on_evaluate_config_fn=self.evaluate_config
strategy = self.get_strategy(min_fit_clients,num_clients,eval_fn,on_fit_config_fn,on_evaluate_config_fn)
elif (evaluation_required.lower() == 'false'):
eval_fn=None
if (model_name.lower() == "deeplearning"):
# min_fit_clients =int( model_params["min_fit_clients"])
on_fit_config_fn=self.fit_config
on_evaluate_config_fn=self.evaluate_config
strategy = self.get_strategy(min_fit_clients,num_clients,eval_fn,on_fit_config_fn,on_evaluate_config_fn)
else:
min_fit_clients=0
on_fit_config_fn=self.fit_round
on_evaluate_config_fn=None
# strategy = self.get_strategy(min_fit_clients,num_clients,eval_fn,on_fit_config_fn,on_evaluate_config_fn)
strategy = fl.server.strategy.FedAvg(
min_available_clients=num_clients,
eval_fn=None,
on_fit_config_fn=self.fit_round,)
else:
log(INFO, "Please opt server evaluation as True or False in server configuration file.")
log(INFO, "Federated learning Server started at @: %s ",str(server_address))
server_rnd=1
while (1):
try:
fl.server.start_server(server_address=server_address, strategy=strategy, config=fl.server.ServerConfig(num_rounds=num_train_round))# config={"num_rounds": num_train_round})#config=fl.server.ServerConfig(num_rounds=3) #,force_final_distributed_eval=True)
except Exception as e:
log(INFO, "Server exception: %s ",str(e))
log(INFO, "AION federated learning server completed for execution cycle: %s ",str(server_rnd))
# Evaluate the final trained model
server_rnd+=1
log(INFO, "AION federated learning server execution successfully completed. Please check the log file for more information.")
return True
except Exception as e:
self.log.info("AION Federated Learning Server run error. Error Msg: "+str(e))
log(INFO, "Server not executing, err.msg: %s ",str(e))
return False
# Start Flower server for n rounds of federated learning
# if __name__ == "__main__":
# ''' Testing purpose code '''
# super_obj=flserver1()
# json_file=sys.argv[1]
# super_obj.log.info("User json_file: \n"+str(json_file))
# # configfile=None
# server_address,model_name,num_clients,num_train_round,data_location,model_version,model_version,selected_feature,target_feature = super_obj.configload(super_obj.confparams)
# df = pd.read_csv(data_location)
# # df=super_obj.df
# df =df[~df.isin([np.nan, np.inf, -np.inf]).any(1)]
# df=df.reset_index(drop=True)
# y=df[target_feature]
# X = df.drop(target_feature, axis=1)
# no_classes = len(df.groupby(target_feature).count())
# no_features=len(selected_feature.split(','))
# print("no_classes: \n",no_classes)
# print("no_features: \n",no_features)
# # num_classes = y_train.apply(pd.Series.nunique)
# if (model_name.lower() == "logisticregression"):
# modelName="logisticregression"
# model = LogisticRegression(penalty="l2",max_iter=10, warm_start=True)
# ## May be used in future (model load for server side eval)
# # model=super_obj.model_load(model_location)
# status=utils.setmodelName(modelName)
# utils.set_initial_params(model,no_classes,no_features)
# strategy = fl.server.strategy.FedAvg(
# min_available_clients=num_clients,
# eval_fn=super_obj.get_eval_fn(model,X,y),
# on_fit_config_fn=super_obj.fit_round,)
# # super_obj.log.info("Stating federated learning server.....\n")
# log(INFO, "Stating AION federated learning server.....")
# fl.server.start_server(server_address, strategy=strategy, config={"num_rounds": num_train_round})
# # super_obj.log.info("federated learning server execution completed.\n")
# log(INFO, "AION federated learning server execution completed.....")
|
utils.py | from typing import Tuple, Union, List
import numpy as np
from sklearn.linear_model import LogisticRegression
from sklearn.linear_model import LinearRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.linear_model import SGDClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from flwr.common.logger import log
from logging import INFO
XY = Tuple[np.ndarray, np.ndarray]
Dataset = Tuple[XY, XY]
LogRegParams = Union[XY, Tuple[np.ndarray]]
XYList = List[XY]
modelUsed=None
modelname=None
def setmodelName(modelselected):
try:
modelname=str(modelselected)
print("setmodelName ,given modelname: \n",modelname)
if (modelname.lower() == 'logisticregression'):
modelUsed=LogisticRegression()
return True
elif (modelname.lower() == "linearregression"):
modelUsed = LinearRegression()
return True
elif (modelname.lower() == "sgdclassifier"):
#from sklearn.linear_model import SGDClassifier
modelUsed=SGDClassifier()
return True
elif (modelname.lower() == "knn"):
modelUsed = KNeighborsClassifier()
return True
elif (modelname.lower() == "decisiontreeclassifier"):
modelUsed = DecisionTreeClassifier()
return True
else:
return False
except Exception as e:
log(INFO, "set fl model name fn issue: ",e)
def get_model_parameters(model:modelUsed) -> LogRegParams:
"""Returns the paramters of a sklearn LogisticRegression model."""
model_name=model.__class__.__name__
if model.fit_intercept:
params = (model.coef_, model.intercept_)
else:
params = (model.coef_,)
return params
def set_model_params(
model:modelUsed, params: LogRegParams
) -> modelUsed:
"""Sets the parameters of a sklean LogisticRegression model."""
model.coef_ = params[0]
model_name=model.__class__.__name__
try:
if model.fit_intercept:
model.intercept_ = params[1]
except Exception as e:
log(INFO, "set_model_params fn issue: ",e)
pass
return model
def set_initial_params_reg(model,no_vals,no_features):
"""Sets initial parameters as zeros Required since model params are
uninitialized until model.fit is called.
But server asks for initial parameters from clients at launch. Refer
to sklearn.linear_model.LogisticRegression documentation for more
information.
"""
no_vals = no_vals
n_features = no_features
# model.classes_ = np.array([i for i in range(n_classes)])
model.coef_ = np.zeros( n_features,)
model_name=model.__class__.__name__
try:
if model.fit_intercept:
# model.intercept_ = np.ones((no_vals,1))
model.intercept_ = np.zeros((no_vals,))
except Exception as e:
log(INFO, "set_initial_params fn issue: ",e)
pass
def set_initial_params(model,no_classes,no_features):
"""Sets initial parameters as zeros Required since model params are
uninitialized until model.fit is called.
But server asks for initial parameters from clients at launch. Refer
to sklearn.linear_model.LogisticRegression documentation for more
information.
"""
n_classes = no_classes
n_features = no_features
model.classes_ = np.array([i for i in range(n_classes)])
model.coef_ = np.zeros((n_classes, n_features))
model_name=model.__class__.__name__
try:
if model.fit_intercept:
model.intercept_ = np.zeros((n_classes,))
except Exception as e:
log(INFO, "set_initial_params fn issue: ",e)
pass
def shuffle(X: np.ndarray, y: np.ndarray) -> XY:
"""Shuffle X and y."""
rng = np.random.default_rng()
idx = rng.permutation(len(X))
return X[idx], y[idx]
def partition(X: np.ndarray, y: np.ndarray, num_partitions: int) -> XYList:
"""Split X and y into a number of partitions."""
return list(
zip(np.array_split(X, num_partitions), np.array_split(y, num_partitions))
)
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
incBatchLearning.py |
import sys
import os
import pickle
import json
import timeit
import warnings
import traceback
import logging
from pathlib import Path
warnings.filterwarnings("ignore")
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from pandas import json_normalize
import shutil
from word2number import w2n
from pytz import timezone
import datetime
from sklearn.model_selection import train_test_split
from sklearn.metrics import roc_auc_score, accuracy_score, r2_score,mean_absolute_error, mean_squared_error, recall_score, precision_score, f1_score
from river import stream
class incBatchLearner():
def __init__(self):
self.home = os.path.dirname(os.path.abspath(__file__))
self.configPath = os.path.join(self.home, 'production', 'Config.json')
self.configDict = {}
self.updConfigDict = None
self.incFillPath = os.path.join(self.home,'production','profiler','incFill.pkl')
self.incOutlierRemPath = os.path.join(self.home, 'production', 'profiler', 'incOutlierRem.pkl')
self.incLabelMappingPath = os.path.join(self.home,'production', 'profiler' , 'incLabelMapping.pkl')
self.incCatEncoderPath = os.path.join(self.home, 'production' , 'profiler', 'incCatEncoder.pkl')
self.incScalerPath = os.path.join(self.home, 'production', 'profiler','incScaler.pkl')
self.testPath = os.path.join(self.home, 'data', 'test.csv')
self.modelName = ''
self.incFill = None
self.incLabelMapping = None
self.incCatEncoder = None
self.incScaler = None
self.incOutlierRem = None
self.model = None
self.targetCol = None
self.numFtrs = []
self.catFtrs = []
self.allFtrs = []
self.logFileName=os.path.join(self.home,'log','model_training_logs.log')
filehandler = logging.FileHandler(self.logFileName, 'a','utf-8')
formatter = logging.Formatter('%(message)s')
filehandler.setFormatter(formatter)
self.log = logging.getLogger('eion')
self.log.propagate = False
self.log.addHandler(filehandler)
self.log.setLevel(logging.INFO)
def readData(self, data, isTest = False):
if not isTest:
self.log.info('New Data Path: '+str(data))
else:
self.log.info('Test Data Path: '+str(data))
startTime = timeit.default_timer()
if os.path.splitext(data)[1] == ".tsv":
df=pd.read_csv(data,encoding='utf-8',sep='\t')
elif os.path.splitext(data)[1] == ".csv":
df=pd.read_csv(data,encoding='utf-8')
elif os.path.splitext(data)[1] == ".dat":
df=pd.read_csv(data,encoding='utf-8')
else:
if os.path.splitext(data)[1] == ".json":
with open(data,'r',encoding='utf-8') as f:
jsonData = json.load(f)
else:
jsonData = json.loads(data)
df = json_normalize(jsonData)
dataLoadTime = timeit.default_timer() - startTime
self.log.info('\nData Load time(sec) :'+str(dataLoadTime))
self.log.info('\n First ten rows of new data')
self.log.info(df.head(10))
self.log.info('Data Frame shape: '+str(df.shape))
df.rename(columns=lambda x:x.strip(), inplace=True)
return df
def readConfig(self):
with open(self.configPath, 'r', encoding= 'utf8') as f:
self.configDict = json.load(f)
self.configDict['partialFit']+=1
self.log.info('************* Partial Fit '+str(self.configDict['partialFit'])+' *************** \n')
msg = '-------> Execution Start Time: '+ datetime.datetime.now(timezone("Asia/Kolkata")).strftime('%Y-%m-%d %H:%M:%S' + ' IST')
self.log.info(msg)
self.targetCol = self.configDict['targetCol']
if 'numFtrs' in self.configDict:
self.numFtrs = self.configDict['numFtrs']
if 'catFtrs' in self.configDict:
self.catFtrs = self.configDict['catFtrs']
if 'allNumCols' in self.configDict:
self.allNumCols = self.configDict['allNumCols']
if 'allCatCols' in self.configDict:
self.allCatCols = self.configDict['allCatCols']
if 'encCols' in self.configDict:
self.encCols = self.configDict['encCols']
if 'wordToNumCols' in self.configDict:
self.wordToNumericCols = self.configDict['wordToNumCols']
self.emptyFtrs = self.configDict['emptyFtrs']
if 'encTarget' in self.configDict:
self.encTarget = self.configDict['encTarget']
if 'noOfClasses' in self.configDict:
self.allClasses = list(range(int(self.configDict['noOfClasses'])))
self.misval_ratio = self.configDict['misval_ratio']
self.allFtrs = self.configDict['allFtrs']
self.modelName = self.configDict['modelName']
self.problemType = self.configDict['problemType']
self.modelPath = os.path.join(self.home, 'production', 'model', self.modelName+'.pkl')
self.scoreParam = self.configDict['scoreParam']
self.score = self.configDict['score']
def pickleLoad(self, file, filename):
if os.path.exists(file):
with open(file, 'rb') as f:
model = pickle.load(f)
file_size = os.path.getsize(file)
self.log.info(str(filename)+" size is :"+str(file_size)+"bytes")
return model
else:
return None
def s2n(self,value):
try:
x=eval(value)
return x
except:
try:
return w2n.word_to_num(value)
except:
return np.nan
def convertWordToNumeric(self,dataframe,feature):
try:
dataframe[feature]=dataframe[feature].apply(lambda x: self.s2n(x))
return dataframe
except Exception as inst:
self.log.info("convertWordToNumeric Failed ===>"+str(inst))
return dataframe
def pickleDump(self, model, path):
if model is not None:
with open(path, 'wb') as f:
pickle.dump(model, f)
def splitTrainTest(self,X,y):
if self.problemType.lower() == 'regression':
xtrain,xtest,ytrain,ytest=train_test_split(X,y,test_size=0.2,shuffle=True)
else:
try:
xtrain,xtest,ytrain,ytest=train_test_split(X,y,stratify=y,test_size=0.2,shuffle=True)
except:
xtrain,xtest,ytrain,ytest=train_test_split(X,y,test_size=0.2,shuffle=True)
return xtrain,xtest,ytrain,ytest
def loadSavedModels(self):
self.incFill = self.pickleLoad(self.incFillPath, 'Online Missing Value Filler')
self.incLabelMapping = self.pickleLoad(self.incLabelMappingPath, 'Online Label Encoder')
self.incCatEncoder = self.pickleLoad(self.incCatEncoderPath, 'Online Categorical Encoder')
self.incScaler = self.pickleLoad(self.incScalerPath, 'Online Scaler')
self.incOutlierRem = self.pickleLoad(self.incOutlierRemPath, 'Online Outlier Detector')
self.model = self.pickleLoad(self.modelPath, str(os.path.basename(self.modelPath))[:-4])
self.log.info('\nData Profiler and ML models loaded in Memory')
def saveModels(self):
os.makedirs(os.path.join(self.home, 'production', 'profiler'))
os.makedirs(os.path.join(self.home, 'production', 'model'))
if type(self.configDict['num_fill']) == type({}) or type(self.configDict['cat_fill']) == type({}):
self.pickleDump(self.incFill, self.incFillPath)
self.pickleDump(self.incLabelMapping, self.incLabelMappingPath)
self.pickleDump(self.incCatEncoder, self.incCatEncoderPath)
self.pickleDump(self.incScaler, self.incScalerPath)
self.pickleDump(self.incOutlierRem, self.incOutlierRemPath)
self.pickleDump(self.model, self.modelPath)
self.log.info('Models saved into production')
def saveConfig(self):
with open(self.configPath, 'w', encoding= 'utf8') as f:
json.dump(self.updConfigDict, f, ensure_ascii=False)
def apply_river_model(self, x, profModel, isTest):
if not isTest:
profModel.learn_one(x)
return pd.Series(profModel.transform_one(x))
def apply_enc(self, x, isTest):
if not isTest:
y = x[self.encTarget]
self.incCatEncoder.learn_one(x, y)
return pd.Series(self.incCatEncoder.transform_one(x))
def apply_od_pipe(self, x):
score = self.incOutlierRem.score_one(x)
is_anomaly = self.incOutlierRem.classify(score)
self.incOutlierRem.learn_one(x)
return is_anomaly
def dataFramePreProcess(self, df):
df = df.replace('-', np.NaN)
df = df.replace('?', np.NaN)
df = df.replace(r'^\s*$', np.NaN, regex=True)
columns = list(df.columns)
if self.wordToNumericCols:
for ftr in self.wordToNumericCols:
if ftr in columns:
self.log.info('Converting '+ftr+' to numeric type...')
tempDataFrame=df.copy(deep=True)
testDf = self.convertWordToNumeric(tempDataFrame,ftr)
try:
df[ftr]=testDf[ftr].astype(float)
except:
pass
columns = list(df.columns)
for empCol in self.emptyFtrs:
if empCol in columns:
df = df.drop(columns=[empCol])
columns = list(df.columns)
self.log.info( 'Detecting Missing Values')
nonNAArray=[]
numOfRows = df.shape[0]
for i in columns:
numNa=df.loc[(pd.isna(df[i])),i ].shape[0]
nonNAArray.append(tuple([i,numNa]))
self.missingCols = []
self.emptyCols = []
for item in nonNAArray:
numofMissingVals = item[1]
if(numofMissingVals !=0):
self.log.info('-------> Feature '+str(item[0]))
self.log.info('----------> Number of Empty Rows '+str(numofMissingVals))
self.missingCols.append(item[0])
if(numofMissingVals >= numOfRows * self.misval_ratio):
self.log.info('----------> Empty: Yes')
self.log.info('----------> Permitted Rows: '+str(int(numOfRows * self.misval_ratio)))
self.emptyCols.append(item[0])
if(len(self.missingCols) !=0):
self.log.info( '----------- Detecting for Missing Values End -----------\n')
else:
self.log.info( '-------> Missing Value Features :Not Any')
self.log.info( '----------- Detecting for Missing Values End -----------\n')
return df
def profiler(self, df, isTest=False):
if not isTest:
self.log.info('Starting profiling of New Training Data')
else:
self.log.info('Starting profiling of Testing Data')
startTime = timeit.default_timer()
df = self.dataFramePreProcess(df)
if 'num_fill' in self.configDict:
if self.configDict['num_fill'] == 'drop':
df = df.dropna(axis = 0, subset=self.allNumCols)
elif self.configDict['num_fill'] == 'zero':
df[self.allNumCols] = df[self.allNumCols].fillna(value = 0.0)
else:
df[self.allNumCols]= df[self.allNumCols].apply(pd.to_numeric)
df = df.astype(object).where(df.notna(), None) #river expects nan values to be None
df[self.allNumCols]= df[self.allNumCols].apply(lambda row: self.apply_river_model(row.to_dict(), self.incFill['num_fill'], isTest), axis='columns')
if not isTest:
self.updConfigDict['num_fill'] = {col:self.incFill['num_fill'].stats[col].get() for col in self.allNumCols}
if 'cat_fill' in self.configDict:
if self.configDict['cat_fill'] == 'drop':
df = df.dropna(axis = 0, subset=self.allCatCols)
elif self.configDict['cat_fill'] == 'zero':
df[self.allCatCols] = df[self.allCatCols].fillna(value = 0.0)
else:
df = df.astype(object).where(df.notna(), None)
df[self.allCatCols]= df[self.allCatCols].apply(lambda row: self.apply_river_model(row.to_dict(),self.incFill['cat_fill'], isTest), axis='columns')
if not isTest:
self.updConfigDict['cat_fill'] = {col:self.incFill['cat_fill'].stats[col].get() for col in self.allCatCols}
if not isTest:
self.log.info('Missing value profiler model updated')
if self.incLabelMapping:
uq_classes = df[self.targetCol].unique()
le_classes = list(self.incLabelMapping.classes_)
uq_classes = [type(le_classes[0])(x) for x in uq_classes]
unseen_classes = set(uq_classes) - set(le_classes)
self.log.info('Existing classes: '+str(le_classes))
if len(unseen_classes)>0:
self.log.info('New unseen classes: '+str(unseen_classes))
le_classes.extend(unseen_classes)
from sklearn.preprocessing import LabelEncoder
self.incLabelMapping = LabelEncoder()
self.incLabelMapping.fit(le_classes)
self.log.info(self.incLabelMapping.classes_)
self.log.info('Label encoder refitted with new unseen classes')
df[self.targetCol] = df[self.targetCol].apply(str)
df[self.targetCol] = self.incLabelMapping.transform(df[self.targetCol])
if not isTest:
self.log.info('Target column label encoding is done')
if self.incCatEncoder:
if self.problemType.lower() == 'regression':
from sklearn.preprocessing import StandardScaler
sc = StandardScaler()
self.encTarget = 'scaledTarget'
df['scaledTarget'] = sc.fit_transform(df[self.targetCol].to_numpy().reshape(-1,1))
transformed_data = df[self.encCols].apply(lambda row: self.apply_enc(row.to_dict(), isTest), axis='columns')
if self.targetCol in transformed_data.columns:
transformed_data.drop(self.targetCol, inplace=True, axis = 1)
df[self.catFtrs] = transformed_data
if not isTest:
self.updConfigDict['catEnc'] = []
if len(self.catFtrs) == 1:
col = self.catFtrs[0]
self.configDict['catEnc'].append({col:self.incCatEncoder['TargetAgg'].state.to_dict()})
else:
for i, col in enumerate(self.catFtrs):
if i==0:
no = ''
else:
no = str(i)
self.configDict['catEnc'].append({col:self.incCatEncoder['TransformerUnion']['TargetAgg'+no].state.to_dict()})
self.log.info('Categorical encoding is done and profiler model updated')
if self.incScaler:
if not isTest:
self.incScaler = self.incScaler.partial_fit(df[self.numFtrs])
self.log.info('Numerical features scaled and profiler model updated')
df[self.numFtrs] = self.incScaler.transform(df[self.numFtrs])
if self.incOutlierRem and not isTest:
df = df[df[self.numFtrs].apply(lambda x: False if self.apply_od_pipe(x.to_dict()) else True, axis=1)]
df.reset_index(drop=True, inplace=True)
self.log.info('Outliers removed and profiler model updated')
if not isTest:
self.log.info('Check config file in production folder for updated profiler values')
profilerTime = timeit.default_timer() - startTime
self.log.info('\nProfiling time(sec) :'+str(profilerTime))
return df
def riverTrain(self, X, Y):
trainStream = stream.iter_pandas(X, Y)
for i, (xi, yi) in enumerate(trainStream):
if yi!=None:
self.model.learn_one(xi, yi)
def riverEvaluate(self, xtest):
testStream = stream.iter_pandas(xtest)
preds = []
for xi,yi in testStream:
pred = self.model.predict_one(xi)
preds.append(pred)
return preds
def trainModel(self,df):
startTime = timeit.default_timer()
X = df[self.allFtrs]
Y = df[self.targetCol]
try:
self.riverTrain(X,Y)
trainTime = timeit.default_timer() - startTime
self.log.info('\nModel Training time(sec) :'+str(trainTime))
self.log.info(self.modelName+' model updated')
self.log.info('First fit model params are '+str(self.configDict['modelParams']))
except Exception as e:
raise e
def archiveModels(self):
source = os.path.join(self.home, 'production')
archivePath = os.path.join(self.home,'archives')
if os.path.isdir(archivePath):
NoOfArchives = sum(os.path.isdir(os.path.join(self.home,'archives',str(i))) for i in os.listdir(archivePath))
destination = os.path.join(self.home,'archives',str(NoOfArchives+1))
else:
destination = os.path.join(archivePath,'1')
if not os.path.exists(destination):
os.makedirs(destination)
allfiles = os.listdir(source)
for f in allfiles:
src_path = os.path.join(source, f)
dst_path = os.path.join(destination, f)
shutil.move(src_path, dst_path)
self.log.info('Previous production models archived')
def get_score(self,metric,actual,predict):
if 'accuracy' in str(metric).lower():
score = accuracy_score(actual,predict)
score = score*100
elif 'recall' in str(metric).lower():
score = recall_score(actual,predict,average='macro')
score = score*100
elif 'precision' in str(metric).lower():
score = precision_score(actual,predict,average='macro')
score = score*100
elif 'f1_score' in str(metric).lower():
score = f1_score(actual,predict, average='macro')
score = score*100
elif 'roc_auc' in str(metric).lower():
try:
score = roc_auc_score(actual,predict,average="macro")
except:
try:
actual = pd.get_dummies(actual)
predict = pd.get_dummies(predict)
score = roc_auc_score(actual,predict, average='weighted', multi_class='ovr')
except:
score = 0
score = score*100
elif ('mse' in str(metric).lower()) or ('neg_mean_squared_error' in str(metric).lower()):
score = mean_squared_error(actual,predict)
elif ('rmse' in str(metric).lower()) or ('neg_root_mean_squared_error' in str(metric).lower()):
score=mean_squared_error(actual,predict,squared=False)
elif ('mae' in str(metric).lower()) or ('neg_mean_absolute_error' in str(metric).lower()):
score=mean_absolute_error(actual,predict)
elif 'r2' in str(metric).lower():
score=r2_score(actual,predict)*100
return round(score,2)
def checkColumns(self, df):
self.log.info('Checking DataColumns in new data')
dfCols = list(df.columns)
allCols = self.allFtrs.copy()
allCols.append(self.targetCol)
missingCols = []
for col in allCols:
if col not in dfCols:
missingCols.append(col)
if len(missingCols)>0:
raise Exception('DataFrame is missing columns: '+str(missingCols))
else:
self.log.info('All required columns are present: '+str(list(dfCols)[:500]))
def plotMetric(self):
y = self.configDict['metricList']
fedrows = self.configDict['trainRowsList']
fig = plt.figure()
ax = fig.gca()
if self.configDict['problemType'] == 'classification':
ax.set_yticks(np.arange(0, 110, 10))
plt.ylim(ymin=0)
if self.configDict['problemType'] == 'regression':
minMet = min(y)
maxMet = max(y)
plt.ylim(minMet - 10, maxMet+10)
plt.plot(y)
plt.ylabel(self.scoreParam)
plt.xlabel('Partial Fits')
plt.title(str(self.scoreParam)+' over training rows')
if type(fedrows[0])!=type(''):
fedrows = [str(x) for x in fedrows]
x = list(range(len(fedrows)))
for i in range(len(fedrows)):
plt.annotate(fedrows[i], (x[i], y[i] + 5))
if self.configDict['problemType'] == 'classification':
plt.annotate(round(y[i],1), (x[i], y[i] - 3))
plt.grid()
plt.savefig(os.path.join(self.home, 'production','metric'))
return
def updateLearning(self,data):
try:
self.readConfig()
self.updConfigDict = self.configDict.copy()
df = self.readData(data)
self.checkColumns(df)
self.loadSavedModels()
X = df[self.allFtrs]
y = df[self.targetCol]
xtrain,xtest,ytrain,ytest = self.splitTrainTest(X,y)
dftrain = pd.concat((xtrain, ytrain), axis = 1)
dftest = pd.concat((xtest, ytest), axis = 1)
dftrain = self.profiler(dftrain)
dftest = self.profiler(dftest, isTest = True)
xtest = dftest[self.allFtrs]
ytest = dftest[self.targetCol]
self.trainModel(dftrain)
preds = self.riverEvaluate(xtest)
score = self.get_score(self.scoreParam, ytest, preds)
self.updConfigDict['score'] = score
self.log.info('Previous '+self.scoreParam+': '+str(self.configDict['score']))
self.log.info('Current '+self.scoreParam+': '+str(self.updConfigDict['score']))
self.configDict['trainRowsList'].append(self.configDict['trainRowsList'][-1]+xtrain.shape[0])
self.log.info('Number of data points trained on so far: '+str(self.configDict['trainRowsList'][-1]))
self.configDict['metricList'].append(self.updConfigDict['score'])
self.archiveModels()
self.plotMetric()
self.saveModels()
self.saveConfig()
msg = self.scoreParam+': Previous:'+str(self.configDict['score'])+' Current:'+ str(self.updConfigDict['score'])
output = {"status":"SUCCESS","Msg":msg}
self.log.info(str(output))
except Exception as e:
print(traceback.format_exc())
self.log.info('Partial Fit Failed '+str(traceback.format_exc()))
if self.updConfigDict != None:
self.saveConfig()
output = {"status":"FAIL","Msg":str(e).strip('"')}
return json.dumps(output)
if __name__ == "__main__":
incBLObj = incBatchLearner()
output = incBLObj.updateLearning(sys.argv[1])
print("aion_learner_status:",output)
|
incBatchPrediction.py |
import sys
import os
import pickle
import json
import traceback
import warnings
warnings.filterwarnings("ignore")
import numpy as np
import pandas as pd
import scipy
from pandas import json_normalize
from word2number import w2n
from river import stream
class incBatchPredictor():
def __init__(self):
self.home = os.path.dirname(os.path.abspath(__file__))
self.configPath = os.path.join(self.home, 'production', 'Config.json')
self.configDict = {}
self.incFillPath = os.path.join(self.home,'production','profiler','incFill.pkl')
self.incLabelMappingPath = os.path.join(self.home,'production', 'profiler' , 'incLabelMapping.pkl')
self.incCatEncoderPath = os.path.join(self.home, 'production' , 'profiler', 'incCatEncoder.pkl')
self.incScalerPath = os.path.join(self.home, 'production', 'profiler','incScaler.pkl')
self.incFill = None
self.incLabelMapping = None
self.incCatEncoder = None
self.incScaler = None
self.model = None
self.targetCol = None
self.modelName = ''
self.problemType = ''
self.numFtrs = []
self.catFtrs = []
def readData(self, data):
try:
if os.path.splitext(data)[1] == ".tsv":
df=pd.read_csv(data,encoding='utf-8',sep='\t')
elif os.path.splitext(data)[1] == ".csv":
df=pd.read_csv(data,encoding='utf-8')
elif os.path.splitext(data)[1] == ".dat":
df=pd.read_csv(data,encoding='utf-8')
else:
if os.path.splitext(data)[1] == ".json":
with open(data,'r',encoding='utf-8') as f:
jsonData = json.load(f)
else:
jsonData = json.loads(data)
df = json_normalize(jsonData)
df.rename(columns=lambda x:x.strip(), inplace=True)
return df
except KeyError as e:
output = {"status":"FAIL","message":str(e).strip('"')}
print(json.dumps(output))
except Exception as e:
output = {"status":"FAIL","message":str(e).strip('"')}
print(json.dumps(output))
def readConfig(self):
with open(self.configPath, 'r', encoding= 'utf8') as f:
self.configDict = json.load(f)
self.targetCol = self.configDict['targetCol']
if 'numFtrs' in self.configDict:
self.numFtrs = self.configDict['numFtrs']
if 'catFtrs' in self.configDict:
self.catFtrs = self.configDict['catFtrs']
if 'allNumCols' in self.configDict:
self.allNumCols = self.configDict['allNumCols']
if 'allCatCols' in self.configDict:
self.allCatCols = self.configDict['allCatCols']
if 'wordToNumCols' in self.configDict:
self.wordToNumericCols = self.configDict['wordToNumCols']
self.emptyFtrs = self.configDict['emptyFtrs']
self.allFtrs = self.configDict['allFtrs']
self.modelName = self.configDict['modelName']
self.problemType = self.configDict['problemType']
self.modelPath = os.path.join(self.home, 'production', 'model', self.modelName+'.pkl')
self.scoreParam = self.configDict['scoreParam']
self.score = self.configDict['score']
def pickleLoad(self, file):
if os.path.exists(file):
with open(file, 'rb') as f:
model = pickle.load(f)
return model
else:
return None
def s2n(self,value):
try:
x=eval(value)
return x
except:
try:
return w2n.word_to_num(value)
except:
return np.nan
def convertWordToNumeric(self,dataframe,feature):
try:
dataframe[feature]=dataframe[feature].apply(lambda x: self.s2n(x))
return dataframe
except Exception as inst:
self.log.info("convertWordToNumeric Failed ===>"+str(inst))
return dataframe
def loadSavedModels(self):
self.incFill = self.pickleLoad(self.incFillPath)
self.incLabelMapping = self.pickleLoad(self.incLabelMappingPath)
self.incCatEncoder = self.pickleLoad(self.incCatEncoderPath)
self.incScaler = self.pickleLoad(self.incScalerPath)
self.model = self.pickleLoad(self.modelPath)
def apply_river_model(self, x, profModel):
print(profModel.imputers)
return pd.Series(profModel.transform_one(x))
def apply_enc(self, x):
return pd.Series(self.incCatEncoder.transform_one(x))
def dataFramePreProcess(self, df):
df = df.replace(r'^\s*$', np.NaN, regex=True)
df = df.replace('-', np.nan)
df = df.replace('?', np.nan)
columns = list(df.columns)
if self.wordToNumericCols:
for ftr in self.wordToNumericCols:
if ftr in columns:
tempDataFrame=df.copy(deep=True)
testDf = self.convertWordToNumeric(tempDataFrame,ftr)
try:
df[ftr]=testDf[ftr].astype(float)
except:
pass
columns = list(df.columns)
for empCol in self.emptyFtrs:
if empCol in columns:
df = df.drop(columns=[empCol])
return df
def profiler(self, df):
df = df[self.allFtrs]
df = self.dataFramePreProcess(df)
if 'num_fill' in self.configDict:
if self.configDict['num_fill'] == 'drop':
df = df.dropna(axis = 0, subset=self.allNumCols)
elif self.configDict['num_fill'] == 'zero':
df[self.numFtrs] = df[self.numFtrs].fillna(value = 0.0)
else:
for x in self.numFtrs:
if x == self.targetCol:
continue
df[x] = df[x].fillna(value = self.configDict['num_fill'][x])
if 'cat_fill' in self.configDict:
if self.configDict['cat_fill'] == 'drop':
df = df.dropna(axis = 0, subset=self.allCatCols)
elif self.configDict['cat_fill'] == 'zero':
df[self.catFtrs] = df[self.catFtrs].fillna(value = 0.0)
else:
for x in self.catFtrs:
if x == self.targetCol:
continue
df[x] = df[x].fillna(value = self.configDict['cat_fill'][x])
if self.incCatEncoder:
transformed_data = df[self.catFtrs].apply(lambda row: self.apply_enc(row.to_dict()), axis='columns')
df[self.catFtrs] = transformed_data
if self.incScaler:
df[self.numFtrs] = self.incScaler.transform(df[self.numFtrs])
return df
def trainedModel(self,X):
testStream = stream.iter_pandas(X)
preds = []
if self.problemType.lower() == 'regression':
for xi,yi in testStream:
try:
pred = self.model.predict_proba_one(xi)
preds.append(pred)
except:
pred = self.model.predict_one(xi)
preds.append(pred)
preds = pd.DataFrame(preds)
return preds
elif self.problemType.lower() == 'classification':
for xi,yi in testStream:
try:
pred = self.model.predict_proba_one(xi)
preds.append(pred)
except:
continue
out = pd.DataFrame(preds)
return out
def apply_output_format(self,df,modeloutput):
if self.problemType.lower() == 'regression':
df['prediction'] = modeloutput[0]
df['prediction'] = df['prediction'].round(2)
elif self.problemType.lower() == 'classification':
modeloutput = round(modeloutput,2)
if modeloutput.shape[1] == 1:
df['prediction'] = modeloutput
df['prediction'] = df['prediction'].astype(int)
else:
try:
predCol = modeloutput.idxmax(axis=1)
df['prediction'] = predCol.astype(int)
df['prediction'] = self.incLabelMapping.inverse_transform(df['prediction'])
except:
df['prediction'] = modeloutput.idxmax(axis=1)
df['probability'] = modeloutput.max(axis=1).round(2)
modeloutput.columns = modeloutput.columns.astype(int)
modeloutput.columns = self.incLabelMapping.inverse_transform(list(modeloutput.columns))
df['remarks'] = modeloutput.apply(lambda x: x.to_json(), axis=1)
outputjson = df.to_json(orient='records')
outputjson = {"status":"SUCCESS","data":json.loads(outputjson)}
return(json.dumps(outputjson))
def predict(self,data):
try:
df = self.readData(data)
dfOrg = df.copy()
self.readConfig()
if len(self.configDict)!=0:
self.loadSavedModels()
df = self.profiler(df)
modeloutput = self.trainedModel(df)
dfOrg = dfOrg[self.allFtrs]
output = self.apply_output_format(dfOrg, modeloutput)
else:
pass
except Exception as e:
print(traceback.format_exc())
output = {"status":"FAIL","message":str(e).strip('"')}
return output
if __name__ == "__main__":
incBPobj = incBatchPredictor()
output = incBPobj.predict(sys.argv[1])
print("predictions:",output)
|
dl_model.py | import tensorflow as tf
def dl_regression_model(input_shape, output_shape,
optimizer, loss_func, act_func):
inputs = tf.keras.Input(shape=(input_shape,))
x = tf.keras.layers.Dense(64,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(inputs)
x = tf.keras.layers.Dense(32,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(16,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(8,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
outputs = tf.keras.layers.Dense(output_shape,
kernel_initializer='he_normal',
bias_initializer='zeros')(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
model.compile(loss=loss_func ,
optimizer=optimizer,
metrics=["mean_absolute_error",
"mean_squared_error",
])
return model
def dl_multiClass_classification(input_shape, output_shape,
optimizer, loss_func, act_func, last_act_func):
inputs = tf.keras.Input(shape=(input_shape,))
x = tf.keras.layers.Dense(64,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(inputs)
x = tf.keras.layers.Dense(32,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(16,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(8,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
outputs = tf.keras.layers.Dense(output_shape,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=last_act_func)(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
model.compile(optimizer, loss_func, metrics=["accuracy"])
return model
def dl_binary_classification(input_shape, output_shape,
optimizer, loss_func, act_func, last_act_func):
inputs = tf.keras.Input(shape=(input_shape,))
x = tf.keras.layers.Dense(64,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(inputs)
x = tf.keras.layers.Dense(32,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(16,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
x = tf.keras.layers.Dense(8,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=act_func)(x)
outputs = tf.keras.layers.Dense(output_shape,
kernel_initializer='he_normal',
bias_initializer='zeros',
activation=last_act_func)(x)
model = tf.keras.Model(inputs=inputs, outputs=outputs)
model.compile(optimizer, loss_func,
metrics=["accuracy"])
return model
|
aionflc.py | # -*- coding: utf-8 -*-
"""
Created on Wed May 25 21:16:54 2022
@author: @aionteam
"""
import tensorflow as tf
import warnings
import flwr as flower
import numpy as np
import pandas as pd
import os
from sklearn.linear_model import LogisticRegression
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error, mean_absolute_error,r2_score
from sklearn.model_selection import train_test_split
from sklearn.metrics import log_loss
import utils
import logging
from flwr.common.logger import log
from logging import INFO
import time
import pickle as pkl
import json
import sys
import random
import string
from sklearn.preprocessing import StandardScaler
import dl_model
from sklearn import metrics
## Below import can be used when aion specific grpc communication used.
# from aiongrpcclient import aiongrpcclient
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "3"
os.environ["GRPC_VERBOSITY"] = "debug"
logger = logging.getLogger('AION')
""" The below aion fl client is for sklearn process"""
class aionflc(flower.client.NumPyClient):
def __init__(self,model,num_rounds,model_name,version,wait_time,client_id,num_records,model_overwrite,problem_type,X_train, X_test, y_train, y_test):
self.count=0
self.num_rounds=round(num_rounds)
self.model_name=model_name
self.version=version
self.wait_time=int(wait_time)
self.client_id=client_id
self.num_records=num_records
self.model_overwrite=model_overwrite
self.model=model
self.problem_type=problem_type
self.X_train, self.X_test, self.y_train, self.y_test=X_train, X_test, y_train, y_test
# """ The below part not used now. In future, for our own grpc communication, this module will be used.Call this function where we want. Need to modify aiongrpcproto.proto according our requirement."""
# def callaiongrpcclient(self):
# clientins = aiongrpcclient()
# status=clientins.startgrpcclient()
# return status
#Save the final model
def model_save(self,model):
##Locate standard model dir to save model
cwd = os.path.abspath(os.path.dirname(__file__))
model_location=os.path.join(cwd, 'models')
try:
os.makedirs(model_location)
except FileExistsError as fe:
# here,model_location already exists
pass
model_name=self.model_name
## Saving model
if (self.model_overwrite.lower() == 'false'):
version=str(self.count)
if (model_name.lower() == "deeplearning"):
file_name=model_name+'_'+self.problem_type+'_'+version+".h5"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
model.save(saved_model)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
else:
file_name=model_name+'_'+self.problem_type+'_'+version+".sav"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
with open (saved_model,'wb') as f:
pkl.dump(model,f)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
elif (self.model_overwrite.lower() == 'true'):
version=str(self.version)
if (model_name.lower() == "deeplearning"):
file_name=model_name+'_'+self.problem_type+'_'+version+".h5"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
model.save(saved_model)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
else:
file_name=model_name+'_'+self.problem_type+'_'+version+".sav"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
with open (saved_model,'wb') as f:
pkl.dump(model,f)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
else:
##Write own user instruction
pass
def get_parameters(self, config):
return utils.get_model_parameters(self.model)
def get_properties(self,model,time_out):
"""Return the current client properties."""
client_info={'client_id':self.client_id}
time_out=100
return client_info,model,time_out
def fit(self, parameters, config):
utils.set_model_params(self.model, parameters)
with warnings.catch_warnings():
warnings.simplefilter("ignore")
num_partitions=round(self.num_rounds)
# num_partitions=round(5)
xtrain=np.array_split(self.X_train, num_partitions)[self.count]
ytrain=np.array_split(self.y_train, num_partitions)[self.count]
self.model.fit(xtrain, ytrain)
time.sleep(self.wait_time)
self.count+=1
print("-- Received Weights from Server.")
print(f"\n Training finished for FL round: {config['rnd']}.\n")
logger.info("-- Received Weights from Server. ")
logger.info("Training finished for FL round: "+str(config['rnd'])+" -- Received Weights from Server")
model_param=utils.get_model_parameters(self.model)
model_param=list(model_param)
return model_param, len(self.X_train),{}
# def evaluate(self, parameters, config):
# utils.set_model_params(self.model, parameters)
# print("******** Test_1 ****************** \n")
# if (self.problem_type.lower() == 'classification'):
# if (self.model_name.lower() == 'logisticregression' ):
# loss = log_loss(self.y_test, self.model.predict_proba(self.X_test))
# print("******** Test_1a ****************** \n")
# else:
# if (self.model_name.lower() == 'linearregression' ):
# print("******** Test_1b ****************** \n")
# # loss = log_loss(self.y_test, self.model.predict(self.X_test))
# rmse = np.sqrt(mean_squared_error(self.y_test, self.model.predict(self.X_test)))
# mae = mean_absolute_error(self.y_test, self.model.predict(self.X_test))
# r2=r2_score(self.y_test, self.model.predict(self.X_test))
# loss = rmse
# # accuracy=r2
# print(f"{self.client_id} Sending weights -- data processed {self.num_records}, -- Loss: {(rmse)}. -- r2: {r2}. ")
# logger.info(str(self.client_id)+" Sending weights -- data processed "+str(self.num_records)+".-- Loss: "+str(rmse)+". -- r2: "+str(r2))
# logger.info("FL Client model intercept: "+str(model.intercept_))
# logger.info("FL Client model coefficients: "+str(model.coef_))
# self.model_save(self.model)
# return loss, len(self.X_test), {"r2": r2}
# print("******** Test_1c ****************** \n")
# print("******** Test_2 ****************** \n")
# accuracy = self.model.score(self.X_test, self.y_test)
# print(f"{self.client_id} Sending weights -- data processed {self.num_records}, -- Loss: {(loss)}. -- accuracy: {accuracy}. ")
# logger.info(str(self.client_id)+" Sending weights -- data processed "+str(self.num_records)+".-- Loss: "+str(loss)+". -- accuracy: "+str(accuracy))
# logger.info("FL Client model intercept: "+str(model.intercept_))
# logger.info("FL Client model coefficients: "+str(model.coef_))
# self.model_save(self.model)
# return loss, len(self.X_test), {"accuracy": accuracy}
def evaluate(self, parameters, config):
utils.set_model_params(self.model, parameters)
if (self.problem_type.lower() == 'classification'):
if (self.model_name.lower() == 'logisticregression' ):
loss = log_loss(self.y_test, self.model.predict_proba(self.X_test))
accuracy = self.model.score(self.X_test, self.y_test)
print(f"{self.client_id} Sending weights -- data processed {self.num_records}, -- Loss: {(loss)}. -- accuracy: {accuracy}. ")
logger.info(str(self.client_id)+" Sending weights -- data processed "+str(self.num_records)+".-- Loss: "+str(loss)+". -- accuracy: "+str(accuracy))
logger.info("FL Client model intercept: "+str(model.intercept_))
logger.info("FL Client model coefficients: "+str(model.coef_))
self.model_save(self.model)
return loss, len(self.X_test), {"accuracy": accuracy}
elif (self.problem_type.lower() == 'regression'):
if (self.model_name.lower() == 'linearregression' ):
# loss = log_loss(self.y_test, self.model.predict(self.X_test))
mse=mean_squared_error(self.y_test, self.model.predict(self.X_test))
rmse = np.sqrt(mean_squared_error(self.y_test, self.model.predict(self.X_test)))
mae = mean_absolute_error(self.y_test, self.model.predict(self.X_test))
r2=r2_score(self.y_test, self.model.predict(self.X_test))
loss = rmse
results = {
"mean_absolute_error": mae,
"mean_squared_error": mse,
"root_mean_squared_error": rmse,
"r2":r2,
}
print(f"{self.client_id} Sending weights -- data processed {self.num_records}, -- Loss: {(rmse)}. -- metrics: {results}. ")
logger.info(str(self.client_id)+" Sending weights -- data processed "+str(self.num_records)+".-- Loss: "+str(rmse)+". -- metrics: "+str(results))
logger.info("FL Client model intercept: "+str(self.model.intercept_))
logger.info("FL Client model coefficients: "+str(self.model.coef_))
self.model_save(self.model)
return loss, len(self.X_test), results
""" The below aion fl client is for deep learning process.
Why different client for sklearn and deeplearn ?: Because, flower calling the client object and process all functions (get_parameters,fit and evaluate) internally. So, user space we cannot combine both (sklearn n dl) using if..else. """
class aionflc_dl(flower.client.NumPyClient):
def __init__(self,model,num_rounds,model_name,version,wait_time,client_id,num_records,model_overwrite,problem_type,X_train, X_test, y_train, y_test,model_params):
self.count=0
self.num_rounds=round(num_rounds)
self.model_name=model_name
self.version=version
self.wait_time=int(wait_time)
self.client_id=client_id
self.num_records=num_records
self.model_overwrite=model_overwrite
self.model=model
self.problem_type=problem_type
self.X_train, self.X_test, self.y_train, self.y_test=X_train, X_test, y_train, y_test
self.model_params=model_params
# """ The below part not used now. In future, for our own grpc communication, this module will be used.Call this function where we want. Need to modify aiongrpcproto.proto according our requirement."""
# def callaiongrpcclient(self):
# clientins = aiongrpcclient()
# status=clientins.startgrpcclient()
# return status
#Save the final model
def model_save(self,model):
##Locate standard model dir to save model
cwd = os.path.abspath(os.path.dirname(__file__))
model_location=os.path.join(cwd, 'models')
try:
os.makedirs(model_location)
except FileExistsError as fe:
# here,model_location already exists
pass
model_name=self.model_name
# version=self.version
## Saving model
if (self.model_overwrite.lower() == 'false'):
version=str(self.count)
if (model_name.lower() == "deeplearning"):
file_name=model_name+'_'+self.problem_type+'_'+version+".h5"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
model.save(saved_model)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
else:
file_name=model_name+'_'+self.problem_type+'_'+version+".sav"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
with open (saved_model,'wb') as f:
pkl.dump(model,f)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
elif (self.model_overwrite.lower() == 'true'):
version=str(self.version)
if (model_name.lower() == "deeplearning"):
file_name=model_name+'_'+self.problem_type+'_'+version+".h5"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
model.save(saved_model)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
else:
file_name=model_name+'_'+self.problem_type+'_'+version+".sav"
saved_model=os.path.normpath(os.path.join(model_location,file_name))
log(INFO, "flclient saved_model path: %s ",str(saved_model))
try:
with open (saved_model,'wb') as f:
pkl.dump(model,f)
return True
except Exception as e:
logger.info("model save error. Err.Msg: "+str(e))
return False
else:
##Write own user instruction
pass
def get_parameters(self, config):
"""Get parameters of the local model."""
return self.model.get_weights()
def get_properties(self,model,time_out):
"""Return the current client properties."""
client_info={'client_id':self.client_id}
time_out=100
return client_info,model,time_out
def fit(self, parameters, config):
"""Train parameters on the locally held training set."""
# Update local model parameters
self.model.set_weights(parameters)
num_partitions=(self.num_rounds)
# num_partitions=round(5)
xtrain=np.array_split(self.X_train, num_partitions)[self.count]
ytrain=np.array_split(self.y_train, num_partitions)[self.count]
# y_train = np_utils.to_categorical(y_train, num_classes)
# y_test = np_utils.to_categorical(y_test, num_classes)
# Get hyperparameters for this round
batch_size: int = int(self.model_params["batch_size"])
epochs: int = int(self.model_params["epochs"])
# round: int = config["rnd"]
# self.round_id = round
log(INFO, "===========================")
log(INFO, "Start training model on local client %s round %i",
self.client_id, config['rnd'])
time.sleep(self.wait_time)
self.count+=1
# Train the model using hyperparameters from config
history = self.model.fit(
xtrain,
ytrain,
batch_size,
epochs,
shuffle=False,
# validation_split=0.1,
validation_data=(self.X_test, self.y_test),
verbose=1
)
# Return updated model parameters and results
parameters_prime = self.model.get_weights()
num_examples_train = len(self.X_train)
model_name = self.model_name
problem_type = self.problem_type
if model_name == "deeplearning":
if problem_type == "classification":
acc = self.model.history.history['val_accuracy']
log(INFO, "Validated accuracy at the end of current round of client %s : %.2f %%",
self.client_id, acc[-1]*100)
log(INFO, "Finished training model on local client %s", self.client_id)
results = {
"loss": history.history["loss"][0],
"accuracy": history.history["accuracy"][0],
"val_loss": history.history["val_loss"][0],
"val_accuracy": history.history["val_accuracy"][0],
}
if problem_type == "regression":
mean_absolute_error = history.history['mean_absolute_error'][0]
mean_squared_error = history.history['mean_squared_error'][0]
y_pred = self.model.predict(self.X_test)
from sklearn import metrics
root_mean_squared_error = np.sqrt(metrics.mean_squared_error(self.y_test, y_pred))
log(INFO, "Mean Absolute Error at the end of current round of client %s : %f",
self.client_id, mean_absolute_error)
log(INFO, "Mean Squared Error at the end of current round of client %s : %f",
self.client_id, mean_squared_error)
log(INFO, "Root Mean Squared Error at the end of current round of client %s : %f",
self.client_id, root_mean_squared_error)
log(INFO, "Finished training model on local client %s", self.client_id)
results = {
"mean_absolute_error": mean_absolute_error,
"mean_squared_error": mean_squared_error,
"root_mean_squared_error": root_mean_squared_error,
}
return parameters_prime, num_examples_train, results
def evaluate(self, parameters, config):
"""Evaluate parameters on the locally held test set."""
# Update local model with global parameters
self.model.set_weights(parameters)
num_partitions=(self.num_rounds)
# Get config values
# batch_size: int = config["val_batch_size"]
batch_size: int = int(self.model_params["batch_size"])
steps: int = np.ceil(len(self.X_test)/batch_size)
num_examples_test = len(self.X_test)
log(INFO, "Run for only %i steps", steps)
# Evaluate global model parameters on the local test data and return results
model_name = self.model_name
problem_type = self.problem_type
self.model_save(self.model)
if model_name == "deeplearning":
if problem_type == "classification":
loss, accuracy = self.model.evaluate(self.X_test, self.y_test,verbose=0)
log(INFO, "Client %s : Accuracy %.2f %%", self.client_id, accuracy*100)
log(INFO, "Client %s : Loss %.4f ", self.client_id, loss)
return loss, num_examples_test, {"accuracy": accuracy}
if problem_type == "regression":
loss, mean_absolute_error, mean_squared_error = self.model.evaluate(self.X_test, self.y_test,
steps=steps,verbose=1)
y_pred = self.model.predict(self.X_test)
root_mean_squared_error = np.sqrt(metrics.mean_squared_error(self.y_test, y_pred))
log(INFO, "Client %s : mean_absolute_error %f ",
self.client_id, mean_absolute_error)
log(INFO, "Client %s : mean_squared_error %f ",
self.client_id, mean_squared_error)
log(INFO, "Client %s : root_mean_squared_error %f ",
self.client_id, root_mean_squared_error)
return loss, num_examples_test, {"mean_absolute_error": mean_absolute_error,
"mean_squared_error": mean_squared_error,
"root_mean_squared_error": root_mean_squared_error}
def randclientid(s,c):
c=string.ascii_uppercase + string.digits
return ''.join(random.choice(c) for x in range(s))
## Loading input data
def dataLoad(jsonfile):
with open(jsonfile, 'r') as file:
data = json.load(file)
server_ip=str(data["server_IP"])
server_port=str(data["server_port"])
model_name=str(data["model_name"])
problem_type=str(data["problem_type"])
data_location=str(data["data_location"])
# deploy_location=str(data["deploy_location"])
model_params=data["model_hyperparams"]
train_size=int(data["train_size"])
model_version=str(data["version"])
selected_feature=data["selected_feature"]
if (type(selected_feature) is str):
selected_feature=selected_feature.split(',')
model_overwrite=data['model_overwrite']
target_feature=data["target_feature"]
num_records=int(data['num_records_per_round'])
wait_time=data['wait_time']
server_address=server_ip+':'+server_port
# server_address=f"{server_ip}:{server_port}"
return server_address,model_name,problem_type,data_location,model_params,model_version,selected_feature,target_feature,train_size,num_records,wait_time,model_overwrite
# def getfilepath()
""" Main aion federated learning client function call. """
if __name__ == "__main__":
##Client random id gen.
rand_id=randclientid(9, "ABC1234567890")
client_id='flclient-'+str(rand_id)
try:
json_file=sys.argv[1]
except Exception as e:
# sys.stdout.write("Please provide input configuration file. example: < python.exe 'fedclient\aionflc.py' 'fedclient\config.json' > ")
log(INFO, "Please provide input configuration file. example: <python.exe 'fedclient\aionflc.py' 'fedclient\config.json'> \n")
server_address,model_name,problem_type,data_location,model_params,model_version,selected_feature,target_feature,train_size,num_records,wait_time,model_overwrite = dataLoad(json_file)
file_name=model_name+'_'+model_version+".log"
cwd = os.path.abspath(os.path.dirname(__file__))
log_location = os.path.join(cwd, 'logs')
try:
os.makedirs(log_location)
except FileExistsError as fe:
# here,log_location already exists
pass
try:
logobj = logging.getLogger('AION')
fl_log=os.path.normpath(os.path.join(log_location,file_name))
log(INFO, "flclient log file path: %s ",str(fl_log))
logging.basicConfig(filename=fl_log,format='%(asctime)s %(message)s', filemode='w',level=logging.DEBUG)
except Exception as e:
log(INFO, "logging error. Error Msg: %s ",str(e))
pass
## default data location ~data\inputfile.csv
data_location = os.path.normpath(os.path.join(cwd, data_location))
df = pd.read_csv(data_location)
df =df[~df.isin([np.nan, np.inf, -np.inf]).any(axis=1)]
df=df.reset_index(drop=True)
y=df[target_feature]
# X = df.drop(target_feature, axis=1)
#
# print("selected_feature: \n",selected_feature)
X=df[selected_feature]
input_shape = X.shape[1] # len(selected_feature)
output_shape = len(y.value_counts())
test_size=(100-train_size)/100
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=test_size)
no_classes = len(df.groupby(target_feature).count())
no_features=len(selected_feature)
## Pass the train data.
(X_train, y_train) = utils.partition(X_train, y_train, 1)[0]
scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)
# y_train = pd.get_dummies(y_train)
# y_test = pd.get_dummies(y_test)
y_train_dl = pd.get_dummies(y_train, sparse=True)
y_test_dl = pd.get_dummies(y_test, sparse=True)
if (problem_type.lower() == "classification"):
if (model_name.lower() == "logisticregression"):
#n_classes = df[target_feature].nunique()
no_classes = len(df.groupby(target_feature).count())
no_features=len(selected_feature)
logger.info("no_classes: "+str(no_classes))
logger.info("no_features: "+str(no_features))
modelName="logisticregression"
model = None
model = LogisticRegression(**model_params, warm_start=True)
try:
status=utils.setmodelName(model_name)
utils.set_initial_params(model,no_classes,no_features)
except Exception as e:
print("util error: \n",e)
num_rounds=round(len(df)/num_records)
log(INFO, "Federated learning Client connecting to Server @: %s ",str(server_address))
try:
flower.client.start_numpy_client(server_address=server_address, client=aionflc(model,num_rounds,model_name,model_version,wait_time,client_id,num_records,model_overwrite,problem_type,X_train_scaled, X_test_scaled, y_train, y_test))
except Exception as e:
logger.info("AION FL Client instance error: \n"+str(e))
log(INFO, "AION federated learning Client %s execution completed.",str(client_id))
elif (model_name.lower() == "deeplearning"):
optimizer = model_params["optimizer"]
loss_func = model_params["losses"]
act_func = model_params["activation"]
last_act_func = model_params["last_activation"]
input_shape = X.shape[1] # len(selected_feature)
output_shape = len(y.value_counts())
print(f"input_shape:{input_shape}, output_shape:{output_shape}.")
model = None
if output_shape == 2:
if last_act_func == "sigmoid" and loss_func == "binary_crossentropy":
model = dl_model.dl_binary_classification(input_shape, output_shape,
optimizer, loss_func,
act_func, last_act_func)
elif last_act_func == "softmax" and loss_func == "categorical_crossentropy":
model = dl_model.dl_binary_classification(input_shape, output_shape,
optimizer, loss_func,
act_func, last_act_func)
else:
model = dl_model.dl_multiClass_classification(input_shape,
output_shape, optimizer, loss_func,
act_func, last_act_func)
print(model.summary())
# status=utils.setmodelName(modelName)
# utils.set_initial_params(model,no_classes,no_features)
num_rounds=round(len(df)/num_records)
log(INFO, "Federated learning Client connecting to Server @: %s ",str(server_address))
try:
flower.client.start_numpy_client(server_address=server_address,
client=aionflc_dl(model,num_rounds,model_name,model_version,wait_time,client_id,num_records,model_overwrite,problem_type,X_train_scaled, X_test_scaled, y_train_dl, y_test_dl,model_params))
except Exception as e:
logger.info("AION FL Client instance error: \n"+str(e))
log(INFO, "AION federated learning Client %s execution completed.",str(client_id))
logger.info("AION federated learning Client execution completed."+str(client_id))
elif(problem_type.lower() == "regression"):
if (model_name.lower() == "linearregression"):
# model=LinearRegression(**model_params,warm_start=True)
if model_params['fit_intercept'] == 'True':
model_params['fit_intercept'] = True
else:
model_params['fit_intercept'] = False
if model_params['copy_X'] == 'True':
model_params['copy_X'] = True
else:
model_params['copy_X'] = False
if model_params['positive'] == 'True':
model_params['positive'] = True
else:
model_params['positive'] = False
model=LinearRegression(**model_params)
status=utils.setmodelName(model_name)
utils.set_initial_params_reg(model,X_train.shape[0],X_train.shape[1])
num_rounds=round(len(df)/num_records)
log(INFO, "Federated learning Client connecting to Server @: %s ",str(server_address))
try:
flower.client.start_numpy_client(server_address=server_address, client=aionflc(model,num_rounds,model_name,model_version,wait_time,client_id,num_records,model_overwrite,problem_type,X_train_scaled, X_test_scaled, y_train, y_test))
except Exception as e:
logger.info("AION FL Client instance error: \n"+str(e))
log(INFO, "AION federated learning Client %s execution completed.",str(client_id))
elif(model_name.lower() == "deeplearning"):
input_shape = X.shape[1] # len(selected_feature)
output_shape = len(y.value_counts())
optimizer = model_params["optimizer"]
loss_func = model_params["losses"]
act_func = model_params["activation"]
model = None
model = dl_model.dl_regression_model(input_shape, 1,
optimizer, loss_func, act_func)
num_rounds=round(len(df)/num_records)
log(INFO, "Federated learning Client connecting to Server @: %s ",str(server_address))
try:
flower.client.start_numpy_client(server_address=server_address, client=aionflc_dl(model,num_rounds,model_name,model_version,wait_time,client_id,num_records,model_overwrite,problem_type,X_train_scaled, X_test_scaled, y_train, y_test,model_params))
except Exception as e:
logger.info("AION FL Client instance error: \n"+str(e))
log(INFO, "AION federated learning Client %s execution completed.",str(client_id))
|
utils.py | from typing import Tuple, Union, List
import numpy as np
from sklearn.linear_model import LogisticRegression
from sklearn.linear_model import LinearRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.linear_model import SGDClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from flwr.common.logger import log
from logging import INFO
XY = Tuple[np.ndarray, np.ndarray]
Dataset = Tuple[XY, XY]
LogRegParams = Union[XY, Tuple[np.ndarray]]
XYList = List[XY]
modelUsed=None
modelname=None
def setmodelName(modelselected):
try:
modelname=str(modelselected)
print("setmodelName ,given modelname: \n",modelname)
if (modelname.lower() == 'logisticregression'):
modelUsed=LogisticRegression()
return True
elif (modelname.lower() == "linearregression"):
modelUsed = LinearRegression()
return True
elif (modelname.lower() == "sgdclassifier"):
#from sklearn.linear_model import SGDClassifier
modelUsed=SGDClassifier()
return True
elif (modelname.lower() == "knn"):
modelUsed = KNeighborsClassifier()
return True
elif (modelname.lower() == "decisiontreeclassifier"):
modelUsed = DecisionTreeClassifier()
return True
else:
return False
except Exception as e:
log(INFO, "set fl model name fn issue: ",e)
def get_model_parameters(model:modelUsed) -> LogRegParams:
"""Returns the paramters of a sklearn LogisticRegression model."""
model_name=model.__class__.__name__
if model.fit_intercept:
params = (model.coef_, model.intercept_)
else:
params = (model.coef_,)
return params
def set_model_params(
model:modelUsed, params: LogRegParams
) -> modelUsed:
"""Sets the parameters of a sklean LogisticRegression model."""
model.coef_ = params[0]
model_name=model.__class__.__name__
try:
if model.fit_intercept:
model.intercept_ = params[1]
except Exception as e:
log(INFO, "set_model_params fn issue: ",e)
pass
return model
def set_initial_params_reg(model,no_vals,no_features):
"""Sets initial parameters as zeros Required since model params are
uninitialized until model.fit is called.
But server asks for initial parameters from clients at launch. Refer
to sklearn.linear_model.LogisticRegression documentation for more
information.
"""
no_vals = no_vals
n_features = no_features
# model.classes_ = np.array([i for i in range(n_classes)])
model.coef_ = np.zeros( n_features,)
model_name=model.__class__.__name__
try:
if model.fit_intercept:
# model.intercept_ = np.ones((no_vals,1))
model.intercept_ = np.zeros((no_vals,))
except Exception as e:
log(INFO, "set_initial_params fn issue: ",e)
pass
def set_initial_params(model,no_classes,no_features):
"""Sets initial parameters as zeros Required since model params are
uninitialized until model.fit is called.
But server asks for initial parameters from clients at launch. Refer
to sklearn.linear_model.LogisticRegression documentation for more
information.
"""
n_classes = no_classes
n_features = no_features
model.classes_ = np.array([i for i in range(n_classes)])
model.coef_ = np.zeros((n_classes, n_features))
model_name=model.__class__.__name__
try:
if model.fit_intercept:
model.intercept_ = np.zeros((n_classes,))
except Exception as e:
log(INFO, "set_initial_params fn issue: ",e)
pass
def shuffle(X: np.ndarray, y: np.ndarray) -> XY:
"""Shuffle X and y."""
rng = np.random.default_rng()
idx = rng.permutation(len(X))
return X[idx], y[idx]
def partition(X: np.ndarray, y: np.ndarray, num_partitions: int) -> XYList:
"""Split X and y into a number of partitions."""
return list(
zip(np.array_split(X, num_partitions), np.array_split(y, num_partitions))
)
|
parameters.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
from hyperopt import fmin, tpe, hp, STATUS_OK, Trials
import numpy as np
import logging
import sys
import os
class parametersDefine():
def __init__(self):
self.paramDict = None
self.log = logging.getLogger('eion')
def getParamSpaceSize(self,paramDict):
size=1
if(len(paramDict)==0):
return 0
for keys in paramDict.keys():
size=size*len(paramDict[keys])
return size
def paramDefine(self, paramSpace, method):
paramDict = {}
for j in list(paramSpace.keys()):
inp = paramSpace[j]
try:
isLog = False
isLin = False
isRan = False
isList = False
isString = False
try:
# check if functions are given as input and reassign paramspace
v = paramSpace[j]
if 'logspace' in paramSpace[j]:
paramSpace[j] = v[v.find("(") + 1:v.find(")")].replace(" ", "")
isLog = True
elif 'linspace' in paramSpace[j]:
paramSpace[j] = v[v.find("(") + 1:v.find(")")].replace(" ", "")
isLin = True
elif 'range' in paramSpace[j]:
paramSpace[j] = v[v.find("(") + 1:v.find(")")].replace(" ", "")
isRan = True
elif 'list' in paramSpace[j]:
paramSpace[j] = v[v.find("(") + 1:v.find(")")].replace(" ", "")
isList = True
elif '[' and ']' in paramSpace[j]:
paramSpace[j] = v.split('[')[1].split(']')[0].replace(" ", "")
isList = True
x = paramSpace[j].split(',')
except Exception as e:
if isinstance(paramSpace[j], (int, float)):
paramSpace[j] = str(paramSpace[j])
x = []
x.append(paramSpace[j])
str_arg = paramSpace[j]
# check if arguments are string
try:
test = eval(x[0])
except:
isString = True
if isString:
paramDict.update({j: hp.choice(j, x)} if method == 'bayesopt' else {j: x})
else:
res = eval(str_arg)
if isLin:
y = eval('np.linspace' + str(res))
paramDict.update({j: hp.uniform(j, eval(x[0]), eval(x[1]))} if method == 'bayesopt' else {j: y})
elif isLog:
y = eval('np.logspace' + str(res))
paramDict.update(
{j: hp.uniform(j, 10 ** eval(x[0]), 10 ** eval(x[1]))} if method == 'bayesopt' else {j: y})
elif isRan:
y = eval('np.arange' + str(res))
paramDict.update({j: hp.choice(j, y)} if method == 'bayesopt' else {j: y})
# check datatype of argument
elif isinstance(eval(x[0]), bool):
y = list(map(lambda i: eval(i), x))
paramDict.update({j: hp.choice(j, eval(str(y)))} if method == 'bayesopt' else {j: y})
elif isinstance(eval(x[0]), float):
res = eval(str_arg)
if len(str_arg.split(',')) == 3 and not isList:
y = eval('np.linspace' + str(res))
#print(y)
paramDict.update({j: hp.uniform(j, eval(x[0]), eval(x[1]))} if method == 'bayesopt' else {j: y})
else:
y = list(res) if isinstance(res, tuple) else [res]
paramDict.update({j: hp.choice(j, y)} if method == 'bayesopt' else {j: y})
else:
res = eval(str_arg)
if len(str_arg.split(',')) == 3 and not isList:
y = eval('np.linspace' +str(res)) if eval(x[2]) >= eval(x[1]) else eval('np.arange'+str(res))
else:
y = list(res) if isinstance(res, tuple) else [res]
paramDict.update({j: hp.choice(j, y)} if method == 'bayesopt' else {j: y})
except Exception as inst:
self.log.info('\n-----> Parameter parsing failed!!!.' + str(inst))
self.log.info("The entered parameter is invalid: {"+ j +':'+ inp+'}')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
raise
return paramDict
|
machinelearning.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import warnings
warnings.filterwarnings('ignore')
import logging
import sklearn
from sklearn.neighbors import NearestNeighbors
from sklearn.cluster import KMeans
from sklearn.cluster import DBSCAN
from random import sample
from numpy.random import uniform
import numpy as np
import math
import pickle
import os
from math import isnan
from sklearn.preprocessing import binarize
from sklearn.preprocessing import LabelEncoder
from sklearn.metrics import davies_bouldin_score
from utils.file_ops import save_csv_compressed
from sklearn.metrics import silhouette_score
try:
from sklearn.metrics import calinski_harabasz_score as calinski_harabaz_score
except:
from sklearn.metrics import calinski_harabaz_score
import pandas as pd
from sklearn.metrics import confusion_matrix
from sklearn.metrics import classification_report
from sklearn.metrics import roc_curve, auc
from sklearn.metrics import roc_auc_score
from sklearn.metrics import matthews_corrcoef
from sklearn.metrics import brier_score_loss
from sklearn.preprocessing import LabelBinarizer
from sklearn.model_selection import train_test_split
from sklearn.decomposition import LatentDirichletAllocation
from learner.classificationModel import ClassifierModel
from learner.regressionModel import RegressionModel
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error,make_scorer
from sklearn.metrics import mean_squared_error
from sklearn.metrics import RocCurveDisplay, auc, roc_curve
import matplotlib.pyplot as plt
#print("1")
#from alibi.explainers import ALE,plot_ale
#pd.set_option('display.max_columns', 10)
#pd.set_option('display.width', None)
def get_prediction( model, loaded_model, xtrain, xtest=None):
train_prob = None
test_prob = None
predictedData = []
if xtest.empty:
is_xtest = False
else:
is_xtest = True
if model.lower() == 'lda':
if is_xtest:
predictedData = loaded_model.transform(xtest).argmax(axis=1)
trainPredictedData = loaded_model.transform(xtrain)
elif model.lower() == 'dbscan':
if is_xtest:
predictedData = loaded_model.fit_predict(xtest)
predictedData = loaded_model.labels_
trainPredictedData = loaded_model.fit_predict(xtrain)
trainPredictedData = loaded_model.labels_
elif model == 'Neural Architecture Search':
train_prob = estimator.predict(xtrain)
if train_prob.shape[1] == 1:
train_prob = np.hstack(( 1-train_prob, train_prob))
trainPredictedData = np.argmax(train_prob, axis=1)
if is_xtest:
test_prob = estimator.predict(xtest)
if test_prob.shape[1] == 1:
test_prob = np.hstack(( 1-test_prob, test_prob))
predictedData = np.argmax(test_prob, axis=1)
elif model in ['Deep Q Network','Dueling Deep Q Network']:
from tf_agents.trajectories import time_step
from tensorflow import constant
q, _ = loaded_model(np.array(xtrain), step_type=constant([time_step.StepType.FIRST] * np.array(xtrain).shape[0]), training=False)
train_prob = q.numpy()
if train_prob.shape[1] == 1:
train_prob = np.hstack(( 1-train_prob, train_prob))
trainPredictedData = np.argmax(train_prob, axis=1)
predictedData = np.argmax(test_prob, axis=1)
if is_xtest:
q,_ = loaded_model(np.array(xtest), step_type=constant([time_step.StepType.FIRST] * np.array(xtest).shape[0]), training=False)
test_prob = q.numpy()
if test_prob.shape[1] == 1:
test_prob = np.hstack(( 1-test_prob, test_prob))
predictedData = np.argmax(test_prob, axis=1)
else:
if is_xtest:
predictedData = loaded_model.predict(xtest)
trainPredictedData = loaded_model.predict(xtrain)
if hasattr(loaded_model, 'predict_proba'):
train_prob = loaded_model.predict_proba(xtrain)
if is_xtest:
test_prob = loaded_model.predict_proba(xtest)
return trainPredictedData, predictedData, train_prob, test_prob
class machinelearning(object):
def __init__(self):
self.features=[]
self.log = logging.getLogger('eion')
self.plots = []
def cluster_tendency(self,featureData):
self.log.info("\n------------- Cluster Tendency Check -------------")
d = featureData.shape[1]
n = len(featureData)
m = int(0.1 * n)
nbrs = NearestNeighbors(n_neighbors=1).fit(featureData.values)
rand_X = sample(range(0, n, 1), m)
ujd = []
wjd = []
for j in range(0, m):
u_dist, _ = nbrs.kneighbors(uniform(np.amin(featureData,axis=0),np.amax(featureData,axis=0),d).reshape(1, -1), 2, return_distance=True)
ujd.append(u_dist[0][1])
if isinstance(featureData.iloc[rand_X[j]].values, pd.core.arrays.sparse.array.SparseArray):
featureData_reshaped = np.asarray(featureData.iloc[rand_X[j]].values).reshape(1, -1)
else:
featureData_reshaped = featureData.iloc[rand_X[j]].values.reshape(1, -1)
w_dist, _ = nbrs.kneighbors(featureData_reshaped, 2, return_distance=True)
wjd.append(w_dist[0][1])
try:
clusetTendency = sum(ujd) / (sum(ujd) + sum(wjd))
except:
clusetTendency = 0
if isnan(clusetTendency):
clusetTendency = 0
self.log.info("-------> Cluster Tendency value using Hopkins Statistic: "+str(clusetTendency))
self.log.info("------------- Cluster Tendency Check End-------------\n")
return (clusetTendency)
def calculateNumberofCluster(self,featureData):
self.log.info("\n------------- Calculate Number of Cluster -------------")
Sum_of_squared_distances = []
K = range(1,15)
for k in K:
km = KMeans(n_clusters=k)
km = km.fit(featureData)
Sum_of_squared_distances.append(km.inertia_)
x1, y1 = 1, Sum_of_squared_distances[0]
x2, y2 = 15, Sum_of_squared_distances[len(Sum_of_squared_distances)-1]
distances = []
for inertia in range(len(Sum_of_squared_distances)):
x0 = inertia+2
y0 = Sum_of_squared_distances[inertia]
numerator = abs((y2-y1)*x0 - (x2-x1)*y0 + x2*y1 - y2*x1)
denominator = math.sqrt((y2 - y1)**2 + (x2 - x1)**2)
distances.append(numerator/denominator)
n_clusters=distances.index(max(distances)) + 2
self.log.info("-------> n_clusters: "+str(n_clusters-1))
self.log.info("------------- Calculate Number of Cluster End-------------\n")
return(n_clusters-1)
def getclusterMatrix(self,featureData,targetData):
silhouetteAvg = silhouette_score(featureData,targetData)
self.log.info("-------> SilHouette_Avg: "+str(silhouetteAvg))
daviesBouldinScore=davies_bouldin_score(featureData, targetData)
self.log.info("-------> DaviesBouldinScore: "+str(daviesBouldinScore))
calinskiHarabazScore=calinski_harabaz_score(featureData,targetData)
self.log.info("-------> CalinskiHarabazScore: "+str(calinskiHarabazScore))
matrix = '"SilHouette_Avg":'+str(silhouetteAvg)+',"DaviesBouldinScore":'+str(daviesBouldinScore)+',"CalinskiHarabazScore":'+str(calinskiHarabazScore)
return(matrix)
def get_regression_matrix(self,targetData,predictedData):
try:
r2score=r2_score(targetData, predictedData)
self.log.info('-------> R2_score :'+str(r2score))
except Exception as e:
self.log.info('\n--------- r2_score ',str(e))
r2score = 0
try:
meanabsoluteerror=(mean_absolute_error(targetData, predictedData))
self.log.info('-------> MAE :'+str(meanabsoluteerror))
except Exception as e:
self.log.info('\n---------Error: meanabsoluteerror ',str(e))
meanabsoluteerror = 0
try:
meanssquatederror=mean_squared_error(targetData, predictedData)
self.log.info('-------> MSE :'+str(meanssquatederror))
except Exception as e:
self.log.info('\n---------Error: meanssquatederror ',str(e))
meanssquatederror = 0
try:
rootmeanssquatederror=mean_squared_error(targetData, predictedData,squared=False)
self.log.info('-------> RMSE :'+str(rootmeanssquatederror))
except Exception as e:
self.log.info('\n---------Error: rootmeanssquatederror ',str(e))
rootmeanssquatederror = 0
try:
normalised_rmse_percentage = (rootmeanssquatederror/ ( np.max(targetData) - np.min(targetData) )) * 100
self.log.info('-------> Normalised RMSE percentage :'+str(normalised_rmse_percentage))
except Exception as e:
self.log.info('\n---------Error: Normalised RMSE percentage ',str(e))
normalised_rmse_percentage = -1
try:
targetArray, predictedArray = np.array(targetData), np.array(predictedData)
try:
EPSILON = 1e-10
meanpercentageerror=np.mean(np.abs((targetArray - predictedArray) / (targetArray+EPSILON)))*100
except ZeroDivisionError:
meanpercentageerror = 0
self.log.info('-------> MAPE :'+str(meanpercentageerror))
except Exception as e:
self.log.info('\n---------Error: meanpercentageerror ',str(e))
meanpercentageerror = 0
matrix = '"MAE":'+str(round(meanabsoluteerror,2))+',"R2Score":'+str(round(r2score,2))+',"MSE":'+str(round(meanssquatederror,2))+',"MAPE":'+str(round(meanpercentageerror,2))+',"RMSE":'+str(round(rootmeanssquatederror,2))+',"Normalised RMSE(%)":'+str(round(normalised_rmse_percentage,2))
return matrix
def getClassificationPerformaceMatrix(self,le_trainY,predictedData,prob,labelMaps):
setOfyTrue = set(le_trainY)
unqClassLst = list(setOfyTrue)
if len(unqClassLst) <= 20:
if str(labelMaps) != '{}':
inv_mapping_dict = {v: k for k, v in labelMaps.items()}
unqClassLst2 = (pd.Series(unqClassLst)).map(inv_mapping_dict)
unqClassLst2 = list(unqClassLst2)
else:
unqClassLst2 = unqClassLst
indexName = []
columnName = []
targetnames=[]
for item in unqClassLst2:
indexName.append("act:"+str(item))
columnName.append("pre:"+str(item))
targetnames.append(str(item))
matrixconfusion = pd.DataFrame(confusion_matrix(le_trainY,predictedData, labels = unqClassLst),index = indexName, columns = columnName)
pd.set_option('display.max_columns',len(targetnames)+2)
self.log.info('-------> Confusion Matrix: ')
self.log.info(matrixconfusion)
pd.reset_option('display.max_columns')
classificationreport = pd.DataFrame(classification_report(le_trainY, predictedData, labels = unqClassLst,target_names=targetnames,output_dict=True)).transpose()
self.log.info('-------> Classification Report: ')
self.log.info(classificationreport)
matrixconfusion = matrixconfusion.to_json(orient='index')
classificationreport = classificationreport.to_json(orient='index')
else: #bugid: 14540
self.log.info('-------> As the number of class is more than 20, skipping the creation of confusion_matrix and classification Report')
return ""
lb = LabelBinarizer()
lb.fit(le_trainY)
transformTarget= lb.transform(le_trainY)
if transformTarget.shape[-1] == 1:
transformTarget = le_trainY
prob = np.delete( prob, 0, 1)
rocaucscore = roc_auc_score(transformTarget,prob,average="macro")
brier_score = None
mcc_score = matthews_corrcoef(le_trainY,predictedData)
if len(unqClassLst) > 2:
brier_score = np.mean(np.sum(np.square(prob - transformTarget), axis=1))
else:
brier_score = brier_score_loss(transformTarget,prob)
self.log.info('-------> ROC AUC SCORE :'+str(rocaucscore))
self.log.info(f'-------> Matthews correlation coefficient SCORE : {mcc_score}')
self.log.info(f'-------> BRIER SCORE : {brier_score}')
matrix = f'"ConfusionMatrix": {matrixconfusion},"ClassificationReport": {classificationreport},"ROC_AUC_SCORE": {rocaucscore},"MCC_SCORE": {mcc_score},"BRIER_SCORE": {brier_score}'
return(matrix)
def split_into_train_test_data(self,featureData,targetData,testPercentage,modelType='classification'):
'''
if cvSplit == None:
'''
self.log.info('\n-------------- Test Train Split ----------------')
if testPercentage == 0:
xtrain=featureData
ytrain=targetData
xtest=featureData
ytest=targetData
else:
testSize=testPercentage/100
if modelType == 'regression':
self.log.info('-------> Split Type: Random Split')
xtrain,xtest,ytrain,ytest=train_test_split(featureData,targetData,test_size=testSize,shuffle=True)
else:
try:
self.log.info('-------> Split Type: Stratify Split')
xtrain,xtest,ytrain,ytest=train_test_split(featureData,targetData,stratify=targetData,test_size=testSize,shuffle=True)
except:
self.log.info('-------> Split Type: Random Split')
xtrain,xtest,ytrain,ytest=train_test_split(featureData,targetData,test_size=testSize,shuffle=True)
self.log.info('Status:- !... Train / test split done: '+str(100-testPercentage)+'% train,'+str(testPercentage)+'% test')
self.log.info('-------> Train Data Shape: '+str(xtrain.shape)+' ---------->')
self.log.info('-------> Test Data Shape: '+str(xtest.shape)+' ---------->')
self.log.info('-------------- Test Train Split End ----------------\n')
'''
else:
xtrain=featureData
ytrain=targetData
xtest=featureData
ytest=targetData
'''
return(xtrain,ytrain,xtest,ytest)
def checkForClassBalancing(self,targetData):
imbalancedCount=0
valueCount=targetData.value_counts()
self.log.info("---------- Checking for Class Imbalance on Train Data---------")
self.log.info("-------> Categories and Count:")
self.log.info(valueCount)
categoryList=valueCount.keys().tolist()
categoryCountList=valueCount.tolist()
for i in range(0,len(categoryCountList)):
if float(categoryCountList[i])<=float(0.5*max(categoryCountList)):
self.log.info("-------> Found Imbalanced class: '"+str(categoryList[i])+"' Count: "+str(categoryCountList[i]))
imbalancedCount=imbalancedCount+1
if imbalancedCount == 0:
self.log.info("-------> Status: Balanced")
self.log.info('Status:- |... Check for Data balancing done: Balanced')
else:
self.log.info("-------> Status: Unbalanced")
self.log.info('Status:- |... Check for Data balancing done: Unbalanced')
self.log.info("---------- Checking for Class Imbalance on Train Data End---------")
return(imbalancedCount)
def ExecuteClassBalancing(self,featureData,targetData,balancingMethod):
from imblearn.over_sampling import SMOTE
from imblearn.under_sampling import TomekLinks
from collections import Counter
self.log.info('\n------------ Balancing Start --------------')
if balancingMethod.lower() == "oversample":
self.log.info("-------> Method: SMOTE OverSampling Technique")
k=1
seed=100
try:
oversample = SMOTE(sampling_strategy='auto', k_neighbors=k, random_state=seed)
balfeatureData, baltargetData = oversample.fit_resample(featureData, targetData)
self.log.info(baltargetData.value_counts())
except Exception as inst:
self.log.info("\n!!!!!!!!! OverSampling Fails "+str(inst)+" !!!!!!!!!!!!!!\n")
balfeatureData = featureData
baltargetData = targetData
elif balancingMethod.lower() == "undersample":
self.log.info("-------> Method: Tomelinks UnderSampling Technique")
tLinks = TomekLinks()
balfeatureData, baltargetData= tLinks.fit_resample(featureData, targetData)
#Added for checking balancing act by the algorithm.
counter = Counter(baltargetData)
self.log.info("Class counter:\t"+str(baltargetData.value_counts()))
max_class = max(counter,key=counter.get)
max_value = max(counter.values())
self.log.info("Max samples: "+str(max_value)+ " in the class: "+str(max_class))
for k,v in counter.items():
if v < (max_value*98/100):
self.log.info("Undersampling is not able to do perfect data balancing.")
self.log.info("The method is used to identify the desired samples of data from the majority class that is having the lowest Euclidean distance with the minority class data. Downsampling may not balance the class after applying this method.\n")
self.log.info(baltargetData.value_counts())
else:
balfeatureData = featureData
baltargetData = targetData
self.log.info("-------> Method: Balancing Not Applied")
self.log.info('-------> Memory Usage by Training DataFrame After Class Balancing '+str(featureData.memory_usage(deep=True).sum()))
self.log.info('Status:- |... Data balancing done: '+str(balancingMethod))
self.log.info('------------ Balancing End --------------\n')
return(balfeatureData,baltargetData)
def combine_text_features(self,dataFrame,dataColumns):
column_merge_flag = False
merge_columns = []
if(len(dataColumns) > 1):
dataFrame['combined'] = dataFrame[dataColumns].apply(lambda row: ' '.join(row.values.astype(str)), axis=1)
merge_columns = dataColumns
features = ['combined']
column_merge_flag = True
self.log.info("After Text Concatenation")
self.log.info(dataFrame['combined'].head(10))
self.log.info("List of Combined Columns ---> "+ str(dataColumns) +"\n")
else:
features = dataColumns
return(dataFrame,features,column_merge_flag,merge_columns)
'''
def create_preprocessing_pipeline(self,X):
textDataProfilerObj=textDataProfiler()
tfidfVector = TfidfVectorizer(tokenizer = textDataProfilerObj.textTokenizer)
pipe = Pipeline([("cleaner", TextCleaner()),('vectorizer', tfidfVector)])
vectors=pipe.fit(X)
transformedVector=pipe.transform(X)
return(pipe,transformedVector)
'''
def get_topics(self, model, feature_names, no_top_words):
topicDict = {}
for topic_idx, topic in enumerate(model.components_):
wordDict = {}
topicProb = [(feature_names[i],topic[i]/topic.sum()) for i in topic.argsort()[:-no_top_words - 1:-1]]
for word, prob in topicProb:
if word.endswith('_vect'):
word = word[:-len('_vect')]
wordDict[word] = prob
topicDict[ topic_idx] = wordDict
return topicDict
def transform_target_feature(self,dataFrame,targetColumn):
targetDataType=dataFrame[targetColumn].dtypes
pandasNumericDtypes=['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
labelMapping= {}
if targetDataType not in pandasNumericDtypes:
le = LabelEncoder()
le.fit(dataFrame[targetColumn])
le_trainY = le.transform(dataFrame[targetColumn])
labelMapping = dict(zip(le.classes_, le.transform(le.classes_)))
self.log.info(" \n encoded Values of predicator column ===>"+str(labelMapping))
else:
le_trainY = dataFrame[targetColumn]
return le_trainY,labelMapping
def setScoreParams(self,scoreParam,modelType,categoryCountList):
if modelType == 'classification' or modelType == 'TextClassification':
allowedmatrix = ['accuracy','recall','f1_score','precision','roc_auc']
if(scoreParam.lower() not in allowedmatrix):
scoreParam = 'accuracy'
elif scoreParam.lower() == 'none':
scoreParam = 'accuracy'
elif scoreParam.lower() == "recall":
if len(categoryCountList) > 2:
scoreParam = make_scorer(sklearn.metrics.recall_score, average = 'weighted')
else:
scoreParam = make_scorer(sklearn.metrics.recall_score)
elif scoreParam.lower() == "precision" :
if len(categoryCountList) > 2:
scoreParam = make_scorer(sklearn.metrics.precision_score, average = 'weighted')
else:
scoreParam = make_scorer(sklearn.metrics.precision_score)
elif scoreParam.lower() == "f1_score" :
if len(categoryCountList) > 2:
scoreParam = make_scorer(sklearn.metrics.f1_score, average = 'weighted')
else:
scoreParam = make_scorer(sklearn.metrics.f1_score)
elif scoreParam.lower() == "roc_auc" :
if len(categoryCountList) > 2:
scoreParam = make_scorer(sklearn.metrics.roc_auc_score,needs_proba=True,multi_class='ovr',average='weighted')
else:
scoreParam = make_scorer(sklearn.metrics.roc_auc_score)
else:
scoreParam = scoreParam
else:
allowedmatrix = ['mse','r2','rmse','mae']
if(scoreParam.lower() not in allowedmatrix):
scoreParam = 'neg_mean_squared_error'
elif scoreParam.lower() == 'none':
scoreParam = 'neg_mean_squared_error'
elif scoreParam.lower() == 'mse':
scoreParam = 'neg_mean_squared_error'
elif scoreParam.lower() == 'rmse':
#scoreParam = make_scorer(sklearn.metrics.mean_squared_error, squared = False)
scoreParam='neg_root_mean_squared_error'
elif scoreParam.lower() == 'mae':
scoreParam = 'neg_mean_absolute_error'
elif scoreParam.lower() == 'r2':
scoreParam = 'r2'
else:
scoreParam = scoreParam
#self.log.info('Status:- !... Scoring parameters selected')
self.log.info("-------> Scoring parameter: "+str(scoreParam))
return(scoreParam)
def getbestfeatureModel(self,modelType,scoreParam,score1,score2,model1,model2,threshold1,pscore1,rscore1,threshold2,pscore2,rscore2,featuresset1,featureset2):
best_feature_model = featuresset1
self.log.info('\n ---------- ML Summary ------------')
if modelType.lower() == "classification":
if(threshold1 == -1 and threshold2 == -1):
if score1> score2:
self.log.info('-------> Best Features: '+str(featuresset1))
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featuresset1
else:
self.log.info('-------> Best Features:'+str(featureset2))
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featureset2
elif(threshold1 == -1):
self.log.info('-------> Best Features: '+str(featureset2))
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featureset2
elif(threshold1 == -2):
self.log.info('-------> Best Features: '+str(featuresset1))
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model =featuresset1
else:
if pscore1 == pscore2:
if rscore1 > rscore2:
self.log.info('-------> Best Features: '+str(featuresset1))
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featuresset1
else:
self.log.info('-------> Best Features: '+str(featureset2))
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featureset2
elif rscore1 == rscore2:
if pscore1 > pscore2:
self.log.info('-------> Best Features: '+str(featuresset1))
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featuresset1
else:
self.log.info('-------> Best Features: '+str(featureset2))
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featureset2
elif modelType.lower() == "regression":
if scoreParam == "r2" or scoreParam == "explained_variance":
if score1> score2 :
self.log.info('-------> Best Features: '+str(featuresset1))
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featuresset1
else:
self.log.info('-------> Best Features: '+str(featureset2))
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featureset2
else:
if score1< score2 :
self.log.info('-------> Best Features: '+str(featuresset1))
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featuresset1
else:
self.log.info('-------> Best Features: '+str(featureset2))
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = featureset2
self.log.info('---------- ML Summary End ------------\n')
return(best_feature_model)
def startLearning(self,mlconfig,modelType,modelParams,modelList,scoreParam,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,topFeatures,modelFeatures,allFeatures,targetType,deployLocation,iterName,iterVersion,trained_data_file,predicted_data_file,labelMaps,featuresBasedOn,code_configure,featureEngineeringSelector,modelEvaluationConfig,imageFolderLocation):
model = 'None'
params = 'None'
score = 0xFFFF
estimator = None
model_tried = ''
threshold = -1
pscore = -1
rscore = -1
topics = {}
if(targetColumn != ''):
targetData = dataFrame[targetColumn]
datacolumns=list(dataFrame.columns)
if targetColumn in datacolumns:
datacolumns.remove(targetColumn)
if(modelType != 'clustering') and (modelType != 'TopicModelling'):
scoreParam = self.setScoreParams(scoreParam,modelType,categoryCountList)
if len(topFeatures) > 0:
self.log.info('\n-------------- Training ML: Top/StatisticalBased Features Start --------------')
modelbasedon = 'StatisticalBased'
if featureEngineeringSelector.lower() == 'true':
self.log.info('Status:- |... Algorithm analysis based on feature engineering based feature selection started')
modelbasedon = 'DimensionalityReduction'
else:
self.log.info('Status:- |... Algorithm analysis based on statistical based feature selection started')
model_type1,model1,params1, score1, estimator1,model_tried1,xtrain1,ytrain1,xtest1,ytest1,threshold1,pscore1,rscore1,method,topics=self.startLearnerModule(mlconfig,modelType,modelParams,modelList,scoreParam,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,topFeatures,targetType,deployLocation,iterName,iterVersion,trained_data_file,labelMaps,featuresBasedOn, modelbasedon,code_configure,modelEvaluationConfig)
if model_tried != '':
model_tried += ','
model_tried += model_tried1
topFeaturesStatus = True
if featureEngineeringSelector.lower() == 'true':
self.log.info('Status:- |... Algorithm analysis based on feature engineering based feature selection completed')
else:
self.log.info('Status:- |... Algorithm analysis for statistical based feature completed')
self.log.info('-------------- Training ML: Top/StatisticalBased Features End --------------\n')
else:
topFeaturesStatus = False
if len(modelFeatures) > 0:
self.log.info('\n-------------- Training ML: Models Based Selected Features Start --------------')
self.log.info('Status:- |... Algorithm analysis based on model based feature selection started')
model_type2,model2,params2, score2, estimator2,model_tried2,xtrain2,ytrain2,xtest2,ytest2,threshold2,pscore2,rscore2,method,topics=self.startLearnerModule(mlconfig,modelType,modelParams,modelList,scoreParam,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,modelFeatures,targetType,deployLocation,iterName,iterVersion,trained_data_file,labelMaps,featuresBasedOn, "ModelBased",code_configure,modelEvaluationConfig)
#model_tried2['Features'] = 'ModelBased'
if model_tried != '':
model_tried += ','
model_tried += model_tried2
modelFeaturesStatus = True
self.log.info('Status:- |... Algorithm analysis for model based selected features completed')
self.log.info('-------------- Training ML: Models Based Selected Features End --------------\n')
else:
modelFeaturesStatus = False
if len(allFeatures) > 0:
self.log.info('Status:- |... Algorithm analysis based on all features Start')
model_type3,model3,params3, score3, estimator3,model_tried3,xtrain3,ytrain3,xtest3,ytest3,threshold3,pscore3,rscore3,method,topics=self.startLearnerModule(mlconfig,modelType,modelParams,modelList,scoreParam,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,allFeatures,targetType,deployLocation,iterName,iterVersion,trained_data_file,labelMaps,featuresBasedOn, "AllFeatures",code_configure,modelEvaluationConfig)
#model_tried3['Features'] = 'AllFeatures'
allFeaturesStatus = True
if model_tried != '':
model_tried += ','
model_tried += model_tried3
self.log.info('Status:- |... Algorithm analysis based all features completed')
else:
allFeaturesStatus = False
#print(topFeaturesStatus,modelFeaturesStatus,allFeaturesStatus)
if topFeaturesStatus:
if modelFeaturesStatus:
best_feature_model = self.getbestfeatureModel(modelType,scoreParam,score1,score2,model1,model2,threshold1,pscore1,rscore1,threshold2,pscore2,rscore2,'StatisticalBased','ModelBased')
if best_feature_model == 'StatisticalBased' and allFeaturesStatus:
best_feature_model = self.getbestfeatureModel(modelType,scoreParam,score1,score3,model1,model3,threshold1,pscore1,rscore1,threshold3,pscore3,rscore3,'StatisticalBased','AllFeatures')
if best_feature_model == 'ModelBased' and allFeaturesStatus:
best_feature_model = self.getbestfeatureModel(modelType,scoreParam,score2,score3,model2,model3,threshold2,pscore2,rscore2,threshold3,pscore3,rscore3,'ModelBased','AllFeatures')
elif allFeaturesStatus:
best_feature_model = self.getbestfeatureModel(modelType,scoreParam,score1,score3,model1,model3,threshold1,pscore1,rscore1,threshold3,pscore3,rscore3,'StatisticalBased','AllFeatures')
else:
best_feature_model = 'StatisticalBased'
if featureEngineeringSelector.lower() == 'true':
best_feature_model = 'DimensionalityReduction'
else:
if modelFeaturesStatus and allFeaturesStatus:
best_feature_model = self.getbestfeatureModel(modelType,scoreParam,score2,score3,model2,model3,threshold2,pscore2,rscore2,threshold3,pscore3,rscore3,'ModelBased','AllFeatures')
elif modelFeaturesStatus:
best_feature_model = 'ModelBased'
elif allFeaturesStatus:
best_feature_model = 'AllFeatures'
if (best_feature_model == 'StatisticalBased' or best_feature_model == 'DimensionalityReduction'):
model_type = model_type1
model = model1
params = params1
score = score1
estimator = estimator1
#model_tried = model_tried1
xtrain = xtrain1
ytrain = ytrain1
xtest = xtest1
ytest = ytest1
features = topFeatures
threshold = threshold1
pscore = pscore1
rscore = rscore1
elif (best_feature_model == 'AllFeatures'):
model_type = model_type3
model = model3
params = params3
score = score3
estimator = estimator3
#model_tried = model_tried3
xtrain = xtrain3
ytrain = ytrain3
xtest = xtest3
ytest = ytest3
features = allFeatures
threshold = threshold3
pscore = pscore3
rscore = rscore3
else:
model_type = model_type2
model = model2
params = params2
score = score2
estimator = estimator2
#model_tried = model_tried2
xtrain = xtrain2
ytrain = ytrain2
xtest = xtest2
ytest = ytest2
threshold = threshold2
pscore = pscore2
rscore = rscore2
features = modelFeatures
if score != 'NA':
self.log.info('Status:- |... Final Best Algorithm selected: '+model+' having score='+str(round(score,2))+' based on '+best_feature_model+' feature selection')
filename = os.path.join(deployLocation,'model',iterName+'_'+iterVersion+'.sav')
saved_model = iterName+'_'+iterVersion+'.sav'
if model == 'Neural Architecture Search':
loaded_model = estimator
try:
estimator.save(filename, save_format="tf")
except Exception:
filename = os.path.join(deployLocation,'model','autoKerasModel.h5')
estimator.save(filename)
saved_model = 'autoKerasModel.h5'
else:
pickle.dump(estimator, open(filename, 'wb'))
loaded_model = pickle.load(open(filename, 'rb'))
if not xtest.empty:
df_test = xtest.copy()
else:
df_test = xtrain.copy()
if threshold == -1:
if model.lower() == 'lda':
predictedData = loaded_model.transform(xtest).argmax(axis=1)
trainPredictedData = loaded_model.transform(xtrain)
elif model.lower() == 'dbscan':
predictedData = loaded_model.fit_predict(xtest)
predictedData = loaded_model.labels_
trainPredictedData = loaded_model.fit_predict(xtrain)
trainPredictedData = loaded_model.labels_
elif model == 'Neural Architecture Search':
test_prob = estimator.predict(xtest)
train_prob = estimator.predict(xtrain)
if train_prob.shape[1] == 1:
train_prob = np.hstack(( 1-train_prob, train_prob))
test_prob = np.hstack(( 1-test_prob, test_prob))
predictedData = np.argmax(test_prob, axis=1)
trainPredictedData = np.argmax(train_prob, axis=1)
elif model in ['Deep Q Network','Dueling Deep Q Network']:
from tf_agents.trajectories import time_step
from tensorflow import constant
from sklearn.preprocessing import MinMaxScaler
q, _ = loaded_model(np.array(xtest), step_type=constant([time_step.StepType.FIRST] * np.array(xtest).shape[0]), training=False)
test_prob = MinMaxScaler().fit_transform( q.numpy())
q, _ = loaded_model(np.array(xtrain), step_type=constant([time_step.StepType.FIRST] * np.array(xtrain).shape[0]), training=False)
train_prob = MinMaxScaler().fit_transform( q.numpy())
predictedData = np.argmax(test_prob, axis=1)
trainPredictedData = np.argmax(train_prob, axis=1)
elif modelType == 'clustering':
if not xtest.empty:
predictedData = loaded_model.predict(xtest)
trainPredictedData = loaded_model.predict(xtrain)
else:
if not xtest.empty:
predictedData = loaded_model.predict(xtest)
trainPredictedData = loaded_model.predict(xtrain)
if hasattr(loaded_model, 'predict_proba'):
train_prob = loaded_model.predict_proba(xtrain)
if not xtest.empty:
test_prob = loaded_model.predict_proba(xtest)
else:
self.log.info("-------> Threshold :"+str(threshold))
if not xtest.empty:
#bug 12437
if 'predict_proba' in dir(loaded_model):
test_prob = loaded_model.predict_proba(xtest)
predictedData = binarize(test_prob[:,1].reshape(-1, 1),threshold=threshold)
else:
raise Exception('--------- Loaded model does not support predict_proba ---------\n')
train_prob = loaded_model.predict_proba(xtrain)
trainPredictedData = binarize(train_prob[:,1].reshape(-1, 1),threshold=threshold)
matrix = ''
try:
if(model_type == 'Classification'):
self.log.info('\n--------- Performance Matrix with Train Data ---------')
train_matrix = self.getClassificationPerformaceMatrix(ytrain,trainPredictedData,train_prob,labelMaps)
self.log.info('--------- Performance Matrix with Train Data End ---------\n')
if not xtest.empty:
self.log.info('\n--------- Performance Matrix with Test Data ---------')
performancematrix = self.getClassificationPerformaceMatrix(ytest,predictedData,test_prob,labelMaps)
df_test['actual'] = ytest
df_test['predict'] = predictedData
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
matrix = performancematrix
if hasattr( loaded_model, 'predict_proba'):
predictedData_fit = loaded_model.predict_proba(xtest)
elif model == 'Neural Architecture Search':
predictedData_fit = estimator.predict(xtest)
elif model in ['Deep Q Network','Dueling Deep Q Network']:
from tf_agents.trajectories import time_step
from tensorflow import constant
q, _ = loaded_model(np.array(xtest), step_type=constant([time_step.StepType.FIRST] * np.array(xtest).shape[0]), training=False)
predictedData_fit = q.numpy()
else:
predictedData_fit = loaded_model.predict(xtest)
if predictedData_fit.shape[1] == 1:
predictedData_fit = np.hstack((1 - predictedData_fit, predictedData_fit))
self.auc_roccurve(ytest,predictedData_fit,labelMaps,imageFolderLocation)
else:
df_test['actual'] = ytrain
df_test['predict'] = trainPredictedData
elif(model_type == 'Regression'):
self.log.info('\n--------- Performance Matrix with Train Data ---------')
train_matrix = self.get_regression_matrix(ytrain, trainPredictedData)
self.log.info('--------- Performance Matrix with Train Data End ---------\n')
if not xtest.empty:
self.log.info('\n--------- Performance Matrix with Test Data ---------')
matrix = self.get_regression_matrix(ytest, predictedData)
df_test['actual'] = ytest
df_test['predict'] = predictedData
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
else:
df_test['actual'] = ytrain
df_test['predict'] = trainPredictedData
elif(model_type == 'Clustering'):
self.log.info('\n--------- Performance Matrix with Train Data ---------')
train_matrix = self.getclusterMatrix(xtrain,trainPredictedData)
self.log.info('--------- Performance Matrix with Train Data End ---------\n')
self.log.info('\n--------- Performance Matrix with Test Data ---------')
performacematrix = self.getclusterMatrix(xtest,predictedData)
df_test['predict'] = predictedData
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
matrix = performacematrix
elif(model_type.lower() == 'topicmodelling'):
self.log.info('\n--------- Performance Matrix with Train Data ---------')
train_matrix = ""
self.log.info('--------- Performance Matrix with Train Data End ---------\n')
self.log.info('\n--------- Performance Matrix with Test Data ---------')
performacematrix = ""
df_test['predict'] = predictedData
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
matrix = performacematrix
except Exception as Inst:
self.log.info('--------- Error Performance Matrix ---------\n')
self.log.info(str(Inst))
df_test['predict'] = predictedData
matrix = ""
train_matrix = ""
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
save_csv_compressed(df_test, predicted_data_file, encoding='utf-8')
return 'Success',model_type,model,saved_model,matrix,train_matrix,xtrain.shape,model_tried,score,filename,features,threshold,pscore,rscore,method,estimator,xtrain,ytrain,xtest,ytest,topics,params
def auc_roccurve(self,y_true,y_score,classee,imageFolderLocation):
from keras.utils import to_categorical
from sklearn.preprocessing import label_binarize
import re
n_classes = len(classee)
y_true = to_categorical(y_true,num_classes = n_classes)
fpr ={}
tpr={}
roc_auc={}
class_names = list(classee.keys())
typeofclass = list(classee.values())
n_class = len(typeofclass)
for i in range(n_classes):
fpr[i],tpr[i],_ = roc_curve(y_true[:,i], y_score[:,i])
roc_auc[i]= auc(fpr[i],tpr[i])
plt.figure()
plt.plot(fpr[i],tpr[i],label=f'{class_names[i]} (AUC = {roc_auc[i]:.2f})')
plt.plot([0,1],[0,1], linestyle='--')
plt.xlabel('False positive rate')
plt.ylabel('True positive rate')
plt.title(f'{class_names[i]} ROC Curve')
plt.legend()
img_location = os.path.join(imageFolderLocation,str(i)+'_roc.png') #15092
plt.savefig(img_location)
def startLearnerModule(self,mlconfig,modelType,modelParams,modelList,scoreParam,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,topFeatures,targetType,deployLocation,iterName,iterVersion,trained_data_file,labelMaps,featuresBasedOn, modelFeatureBased,code_configure,modelEvaluationConfig):
matrix = ''
threshold = -1
pscore = -1
rscore = -1
method = mlconfig['optimizationMethod']
method = method.lower()
geneticParam = ''
topics = {}
optimizationHyperParameter = mlconfig['optimizationHyperParameter']
cvSplit = optimizationHyperParameter['trainTestCVSplit']
nIter = int(optimizationHyperParameter['iterations'])
if(method.lower() == 'genetic'):
geneticParam = optimizationHyperParameter['geneticparams']
scoreParam = scoreParam
if 'thresholdTunning' in mlconfig:
thresholdTunning = mlconfig['thresholdTunning']
else:
thresholdTunning = 'NA'
if len(topFeatures) !=0:
self.features=topFeatures
else:
datacolumns=list(xtrain.columns)
if targetColumn in datacolumns:
datacolumns.remove(targetColumn)
self.features =datacolumns
self.log.info(f'-------> Number of Features Used For Training the Model: {len(self.features)}')
features_names = str(self.features)
if len(features_names) > 500:
features_names = ','.join(self.features[:2]) + ', ..... ,' + ','.join(self.features[-2:])
self.log.info(f'-------> Features Used For Training the Model: {features_names}')
xtrain = xtrain[self.features]
if not xtest.empty:
xtest = xtest[self.features]
if cvSplit == "":
cvSplit =None
else:
cvSplit =int(cvSplit)
if modelType == 'classification':
model_type = "Classification"
MakeFP0 = False
MakeFN0 = False
if(len(categoryCountList) == 2):
self.log.info("\n -------------- Check for FP or FN -------------- ")
self.log.info("-------> Binary Classification")
if(thresholdTunning.lower() == 'fp0'):
self.log.info("-------> Threshold Tuning: False Positive")
MakeFP0 = True
elif(thresholdTunning.lower() == 'fn0'):
self.log.info("-------> Threshold Tuning: False Negative")
MakeFN0 = True
if MakeFP0 == False and MakeFN0 == False:
self.log.info("-------> Threshold Tuning: Not Any")
self.log.info("-------------- Check for FP or FN End-------------- \n")
elif(len(categoryCountList) > 2): #bug 12438
self.log.info("\n -------------- Check for FP or FN -------------- ")
self.log.info("-------> Multiclass Classification")
if(thresholdTunning.lower() == 'fp0' or thresholdTunning.lower() == 'fn0'):
self.log.info("-------> Threshold Tuning: Not supported")
else:
self.log.info("-------> Threshold Tuning: Not Any")
self.log.info("-------------- Check for FP or FN End-------------- \n")
objClf = ClassifierModel(modelList, modelParams, scoreParam, cvSplit, nIter,geneticParam, xtrain,ytrain,xtest,ytest,method,modelType,MakeFP0,MakeFN0,deployLocation)
model, params, score, estimator,model_tried,threshold,pscore,rscore = objClf.classModelling( modelFeatureBased,code_configure)
elif modelType == 'regression':
model_type = "Regression"
objClf = RegressionModel(modelList, modelParams, scoreParam, cvSplit, nIter,geneticParam, xtrain,ytrain,xtest,ytest,method,deployLocation)
model,params,score,estimator,model_tried = objClf.regressionModelling(modelFeatureBased,code_configure)
elif modelType =='clustering':
model_type = 'Clustering'
print(modelList)
if 'KMeans' in modelList:
clustendency = self.cluster_tendency(xtrain)
model='KMeans'
model_tried = '{"Model":"KMeans","Score":"NA"}'
kmeanmodelparams=modelParams['KMeans']
n_clusters = kmeanmodelparams['n_clusters']
if n_clusters == None or n_clusters == 0 or n_clusters == '':
n_clusters = self.calculateNumberofCluster(xtrain)
kmeanmodelparams['n_clusters'] = n_clusters
kmeans=KMeans(n_clusters=n_clusters)
targetData=kmeans.fit_predict(xtrain)
self.log.info('Status:- |... ML Algorithm applied: KMeans')
self.log.info('\n------------ Centers Points Start------------')
values = kmeans.cluster_centers_.squeeze()
#print(values)
centers = pd.DataFrame(kmeans.cluster_centers_,columns= xtrain.columns)
filename = os.path.join(deployLocation,'centers.csv')
centers.to_csv(filename)
labels = kmeans.labels_
i=0
for value_row in values:
j=0
self.log.info('------->Label: '+str(i))
for value in value_row:
self.log.info('---------->Feature: "'+str(self.features[j])+'" Center Point: '+str(value))
j = j+1
i = i+1
self.log.info('------------ Centers Points Start------------\n')
score='NA'
scoreParam=None
params=kmeanmodelparams
estimator=kmeans
if 'DBSCAN' in modelList:
DBSCAN_ModelParams=modelParams['DBSCAN']
db = DBSCAN(eps=DBSCAN_ModelParams['eps'],min_samples = DBSCAN_ModelParams['min_samples']).fit(xtrain)
#targetData=db.fit_predict(xtrain)
self.log.info('Status:- |... ML Algorithm applied: DBSCAN')
labels = db.labels_
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
n_noise_ = list(labels).count(-1)
self.log.info('------->Labels: '+str(labels))
self.log.info('------->No Of Cluster: '+str(n_clusters_))
self.log.info('------->No Of Noise Point: '+str(n_noise_))
score='NA'
scoreParam=None
params=''
estimator=db
model='DBSCAN'
model_tried = '{"Model":"DBSCAN","Score":"NA"}'
elif modelType == 'topicmodelling':
model_type = 'TopicModelling'
model='LDA'
model_tried = '{"Model":"LDA","Score":"NA"}'
LDAmodelparams=modelParams['LDA']
n_topics = LDAmodelparams['n_topics']
n_words_per_topic = LDAmodelparams['n_words_per_topic']
if n_topics == None or n_topics == 0 or n_topics == '':
n_topics = 10
LDAmodelparams['n_topics'] = n_topics
if n_words_per_topic == None or n_words_per_topic == 0 or n_words_per_topic == '':
n_words_per_topic = 10
LDAmodelparams['n_words_per_topic'] = n_words_per_topic
lda = LatentDirichletAllocation(n_components=n_topics,random_state=0)
self.log.info('Status:- |... ML Algorithm applied: LDA')
targetData=lda.fit_transform(xtrain)
topics = self.get_topics(lda, topFeatures, n_words_per_topic)
self.log.info(topics)
score='NA'
scoreParam=None
params=LDAmodelparams
estimator=lda
return model_type,model,params, score, estimator,model_tried,xtrain,ytrain,xtest,ytest,threshold,pscore,rscore,method, topics |
anomalyDetectionAE.py | # -*- coding: utf-8 -*-
#Py Libraries import
import numpy as np
import pandas as pd
import tensorflow as tf
import matplotlib
import matplotlib.pyplot as plt
from sklearn.metrics import accuracy_score
from tensorflow.keras.optimizers import Adam,SGD
# from sklearn.preprocessing import MinMaxScaler
from tensorflow.keras import Model, Sequential
from tensorflow.keras.layers import Dense, Dropout
from sklearn.model_selection import train_test_split
from tensorflow.keras.losses import MeanSquaredLogarithmicError
import os
import json
import keras
from keras.layers import Input, Dense
from keras.callbacks import ModelCheckpoint, EarlyStopping
from sklearn.cluster import DBSCAN
from sklearn.model_selection import RandomizedSearchCV
from sklearn import metrics
import traceback
import seaborn as sns
import warnings
warnings.filterwarnings("ignore")
import logging
import joblib
from sklearn import preprocessing
'''
Py Engine: >= Python 3.8>
anomalyDetectionAE class purpose: To findout anomalies in the user data using autoencoder mechanism. The file contains base class anomalyDetectionAE which act as entry point .
mainAnomalyDetectionfn() method decide which algorithm (autoencoder or dbscan to be used based on user input).
Two different approaches to find out anomalies,
1.LSTM approach (anomalyDetectionAE.aetsmodel_lstm() fn)
2. Autoencoder approach (AeDetector class)
By default, autoencoder approach used (User can select in basic config file.)
One more limitation is, time series data will be handled by autoencoder,lstm algs.
DBScan algorithm is not supporting time series. Suggestion here is if time series data feature received, drop it.
AION dataprofiler automatically remove time feature for DBScan. But for autoencoder we are passing time series.
Parameter information:
df: Input dataframe from aion base class
paramSpace: Default parameter from basic config (user config settings)
deployLocation: Deployment location. Detailed anomalies data information are stored in ../target/output loc.
Target: only for supervised problems.
anomalyMethod: Algorithm to be used (autoEncoder or DBScan) received from paramSpace.
testSize: For supervised problems, unsupervised problems we are passing whole input data.
datetimeFeature: data time feature for autoencoder
mv_featurebased_ad_status: If <True> univariate feature for autoencoder enabled. We findout anomalies for each features selected by user.
'''
##For autoencoder (encoder-decoder) based base class, keras Model class act as meta class. Config params received from AION config file (GUI). The below class instance have keras subclassing call to run encoder and decoder.
class AeDetector(Model):
def __init__(self,train,test,units,latent_units,activation):
super(AeDetector, self).__init__()
#Because anomaly detection ,we r using 'sigmoid' activation for all problems
last_layer_activation='sigmoid'
self.encoder = tf.keras.Sequential([
Dense(units, activation=activation),
Dense((units/2), activation=activation),
Dense(latent_units, activation=activation)
])
self.decoder = tf.keras.Sequential([
Dense((units/2), activation=activation),
Dense(units, activation=activation),
Dense(train.shape[1], activation=last_layer_activation)
])
## Using keras subclassing api
def call(self, x):
encoded = self.encoder(x)
decoded = self.decoder(encoded)
return decoded
##This below function create get_datetime class python file in target->scripts folder
'''This aion_gettimegranularity class is used to retrive the time pattern (Granularity) of given datetime feature.'''
class aion_gettimegranularity:
cls_name="datetimeinformation"
def __init__(self,dataframe, datetimefeature):
self.df=dataframe
self.datetimefeature=datetimefeature
# self.log=logging.getLogger('AION')
self.log = logging.getLogger('eion')
self.log.info("To retrive the granularity of given datetime feature by aion.")
def get_dfinfo(self,df):
from io import StringIO
buf = StringIO()
df.info(buf=buf)
#self.log.info(buf.getvalue())
return buf.getvalue()
##Main time granularity find function
def get_granularity(self):
try:
##get local df
df_t=self.df
buf_info=self.get_dfinfo(df_t)
self.log.info(buf_info)
df_t.drop(df_t.filter(regex='Unname'),axis=1,inplace=True)
try:
df_t[self.datetimefeature] = pd.to_datetime(df_t[self.datetimefeature])
except Exception as e:
self.log.info("Datetime feature to python datetime format convertion error.\n"+str(e))
df_t['time_diff']=df_t[self.datetimefeature].diff().shift(-1)
datetime_mean=df_t['time_diff'].mean()
totsec = datetime_mean.total_seconds()
## Dict variable to store datetime details.Initialized all date param as False.
status_time={"h":False,"m":False,"s":False,
"us":False,"ns":False,"Y":False,"M":False,"D":False}
if (datetime_mean.days == 0):
if (totsec/3600 > 1):
## hour
status_time['h']=True
else:
if (totsec/60 >1):
## minute
status_time['m']=True
else:
if (totsec <= 1e-06 and totsec > 1e-09):
## microsecond
status_time['us']=True
elif (totsec<= 1e-09 and totsec >=1e-012):
## nanosecond
status_time['ns']=True
else:
## second
status_time['s']=True
else:
days=datetime_mean.days
if (days/365>1):
## year
status_time['Y']=True
else:
if (days>30):
## month
status_time['M']=True
else:
## day
status_time['D']=True
time_pattern=None
for k,v in status_time.items():
if (v == True):
time_pattern=k
self.log.info("<----- DateTime feature pattern (year/month/day/hour/minute/second/millisecond/microsecond/nanosecond) is: \t"+str(time_pattern))
try:
try:
df_t[self.datetimefeature] = pd.to_datetime(df_t[self.datetimefeature])
except Exception as e:
pass
df_t['Time_diff'] = ((df_t[self.datetimefeature])).diff(-1).dt.floor('T').dt.total_seconds().div(60).abs()
time_threshold=1
df_t['anomalyType'] = (np.where((df_t['Time_diff'] != 1),"Point","Sequence"))
df_t.drop("Time_diff",axis=1,inplace=True)
except Exception as e:
self.log.info("time_diff err message: "+str(e))
except Exception as e:
print("time_diff err message: ",str(e))
return df_t
## AION Anomaly detection autoencoder main class. It receives input params from anomalyDetector class
class anomalyDetectionAE:
def __init__(self,df,paramSpace,deployLocation,target,anomalyMethod,testSize,datetimeFeature,mv_featurebased_ad_status):
self.mc=None
self.es=None
#aion gui inputs
self.df=df
self.paramSpace=paramSpace
self.deployLocation=deployLocation
self.target=target
self.anomalyMethod=anomalyMethod
self.testSize=round(testSize,1)
self.datetimeFeature=datetimeFeature
self.log = logging.getLogger('eion')
self.mv_featurebased_ad_status=mv_featurebased_ad_status
""" Uncomment below for debug purpose. """
# self.log.info("anomalyDetectionAE constructor: df head: \n"+str(df.head()))
# self.log.info("anomalyDetectionAE constructor: df type: \n"+str(type(df)))
# self.log.info("anomalyDetectionAE constructor: df len: \n"+str(len(df)))
# self.log.info("anomalyDetectionAE constructor: self.datetimeFeature: \n"+str((self.datetimeFeature)))
def targetnumClass(self,data):
count_classes = pd.value_counts(data, sort = True)
num_of_classes= len(count_classes)
return num_of_classes
def configload(self):
cwd = os.path.abspath(os.path.dirname(__file__))
file_name='config.json'
try:
config_file=os.path.normpath(os.path.join(cwd,'config',file_name))
except Exception as e:
self.log.info("<---- config path error. Error Msg: ---->"+str(e))
with open(config_file, 'r') as file:
data = json.load(file)
datapath=str(data["data"])
target=str(data["target"])
anomaly_algorithm=str(data["anomalydetection_algorithm"])
ae_hyperparameter=data["autoEncoder"]
dbscan_hyperparameter=data["DBScan"]
return datapath,target,ae_hyperparameter,anomaly_algorithm,dbscan_hyperparameter
## model summary
def summary(self,model):
return model.summary()
##To load ae model
def model_load(self, path):
cwd = os.path.abspath(os.path.dirname(__file__))
file_name=path
try:
model_location=os.path.normpath(os.path.join(cwd,'model',file_name))
except Exception as e:
self.log.info("<---- Model path error. Error Msg: ---->"+str(e))
loaded_model = joblib.load(model_location)
return loaded_model
## Load dataset
def dataload(self,datapath):
cwd = os.path.abspath(os.path.dirname(__file__))
file_name=datapath
try:
data_file=os.path.normpath(os.path.join(cwd,'data',file_name))
except Exception as e:
self.log.info("<---- data path error. Error Msg:: ---->"+str(e))
df = pd.read_csv(data_file)
return df
## Create dataframe with time sequence data, if not time series, sequence length always 1.
def create_dataset(self,X, y, time_steps=1):
Xs, ys = [], []
for i in range(len(X) - time_steps):
v = X.iloc[i:(i + time_steps)].values
Xs.append(v)
ys.append(y.iloc[i + time_steps])
return np.array(Xs), np.array(ys)
## model for time series based AE encoder, decoder fn
def aetsmodel_lstm(self,n_dims, n_timesteps, n_bottleneck,units,activation,df):
# inputs = Input(shape = (n_timesteps, n_dims))
inputs = Input(shape = (df.shape[1], df.shape[2]))
e = keras.layers.LSTM(units, activation = activation, return_sequences = True)(inputs)
## code layer or compressed form of data produced by the autoencoder, bottleneck layer
latent_space = keras.layers.LSTM(n_bottleneck, activation = activation,
return_sequences = False,
name = 'bottleneck_layer')(e)
e = keras.layers.RepeatVector(n_timesteps)(latent_space)
decoder = keras.layers.LSTM(n_bottleneck, activation = activation,
return_sequences = True)(e)
decoder = keras.layers.LSTM(units, activation = activation, return_sequences = True)(decoder)
outputs = keras.layers.TimeDistributed(Dense(n_dims))(decoder)
model = Model(inputs = inputs, outputs = outputs)
return model
## adding some model checkpoints to ensure the best values will be saved and early stopping to prevent the model from running unnecessary.
def callbacks(self, **kwargs):
self.mc = ModelCheckpoint(filepath = kwargs.get("filename"),
save_best_only = True, verbose = 0)
self.es = EarlyStopping(monitor = kwargs.get("monitor"),
patience = kwargs.get("patience"))
return self.es,self.mc
##This below function create get_datetime class python file in target->scripts folder
'''This aion_gettimegranularity class is used to retrive the time pattern (for getting time granularity) of given datetime feature.'''
def create_datetime_pyfile(self):
try:
datetimepattern_code=r"""##
import pandas as pd
import numpy as np
class aion_gettimegranularity:
cls_name="datetimeinformation"
def __init__(self,dataframe, datetimefeature):
self.df=dataframe
self.datetimefeature=datetimefeature
def get_dfinfo(self,df):
from io import StringIO
buf = StringIO()
df.info(buf=buf)
#print(buf.getvalue())
return buf.getvalue()
def get_granularity(self):
try:
##get local df
df_t=self.df
buf_info=self.get_dfinfo(df_t)
df_t.drop(df_t.filter(regex='Unname'),axis=1,inplace=True)
try:
df_t[self.datetimefeature] = pd.to_datetime(df_t[self.datetimefeature])
except Exception as e:
pass
# print("Datetime feature to python datetime format convertion error.\n",e)
df_t['time_diff']=df_t[self.datetimefeature].diff().shift(-1)
datetime_mean=df_t['time_diff'].mean()
totsec = datetime_mean.total_seconds()
## Dict variable to store datetime details.Initialized all date param as False.
status_time={"h":False,"m":False,"s":False,"us":False,"ns":False,"Y":False,"M":False,"D":False}
if (datetime_mean.days == 0):
if (totsec/3600 > 1):
## hour
status_time['h']=True
else:
if (totsec/60 >1):
## minute
status_time['m']=True
else:
if (totsec <= 1e-06 and totsec > 1e-09):
## microsecond
status_time['us']=True
elif (totsec<= 1e-09 and totsec >=1e-012):
## nanosecond
status_time['ns']=True
else:
## second
status_time['s']=True
else:
days=datetime_mean.days
if (days/365>1):
## year
status_time['Y']=True
else:
if (days>30):
## month
status_time['M']=True
else:
## day
status_time['D']=True
time_pattern=None
for k,v in status_time.items():
if (v == True):
time_pattern=k
#print("<----- DateTime feature pattern (year/month/day/hour/minute/second/millisecond/microsecond/nanosecond) is: \t",(time_pattern))
try:
try:
df_t[self.datetimefeature] = pd.to_datetime(df_t[self.datetimefeature])
except Exception as e:
pass
df_t['Time_diff'] = ((df_t[self.datetimefeature])).diff(-1).dt.floor('T').dt.total_seconds().div(60).abs()
time_threshold=1
df_t['anomalyType'] = np.where((df_t['Time_diff'] != 1),"Point","Sequence")
df_t.drop("Time_diff",axis=1,inplace=True)
except Exception as e:
print("time_diff err message: ",str(e))
except Exception as e:
pass
# print("get_granularity err msg: ",(e))
return df_t
"""
cwd=self.deployLocation
file_name='aion_granularity'+'.py'
try:
data_file=os.path.normpath(os.path.join(cwd,'script',file_name))
with open(data_file,'w') as file:
file.write(datetimepattern_code)
except Exception as error:
self.log.info("<---- datetimepattern_code write Error.: ---->"+str(error))
self.log.info("datetimepattern source code created at target folder...\n")
except Exception as error:
self.log.info("<---- datetimepattern_code function Error.: ---->"+str(error))
## Simple mlp based autoencoder model, not used now.
# def aetsmodel_lstm(self,X_train):
# model = keras.Sequential()
# # autoencoder encoder
# model.add(keras.layers.LSTM(
# units=64,
# input_shape=(X_train.shape[1], X_train.shape[2])
# ))
# model.add(keras.layers.Dropout(rate=0.2))
# model.add(keras.layers.RepeatVector(n=X_train.shape[1]))
# # autoencoder decoder
# model.add(keras.layers.LSTM(units=64, return_sequences=True))
# model.add(keras.layers.Dropout(rate=0.2))
# model.add(
# keras.layers.TimeDistributed(
# keras.layers.Dense(units=X_train.shape[2])
# )
# )
# return model
## To find optimal anomaly threshold value
def find_threshold(self,model, x_train_scaled):
reconstructions = model.predict(x_train_scaled)
# provides losses of individual instances msle
reconstruction_errors = tf.keras.losses.mae(reconstructions, x_train_scaled)
# threshold for anomaly scores
threshold = np.mean(reconstruction_errors.numpy())+ 2*np.std(reconstruction_errors.numpy())
return threshold
## compiling the model with adam optimizer and mean squared error loss
def model_compile(self, model,lr, loss, opt):
if opt == "adam":
opt = Adam(learning_rate = lr)
else:
opt = SGD(learning_rate = lr)
model.compile(loss = loss, optimizer = opt)
## save anomaly points in aion target folder
def save_anomalyvalues(self,df,file_name):
# cwd = os.path.abspath(os.path.dirname(__file__))
cwd=self.deployLocation
file_name=file_name+'.csv'
try:
out_path=os.path.normpath(os.path.join(cwd,'output'))
if not os.path.isdir(out_path):
os.makedirs(out_path)
data_file=os.path.normpath(os.path.join(cwd,'output',file_name))
except Exception as error:
self.log.info("<---- autoencoder artifact_dir path. Error Msg: ---->"+str(error))
try:
df.to_csv(data_file,index=False)
except Exception as e:
self.log.info("<---- Saving log data frame error. Error Msg: ---->"+str(e))
## model summary
def summary(self,model):
return model.summary()
##Method to find subsequence and point anomalies aion_gettimegranularity
def find_point_subsequence_anomalies(self,datetime_column,dataframe=None):
try:
dataframe.reset_index(level=0, inplace=True)
try:
dataframe[datetime_column] = pd.to_datetime(dataframe[datetime_column])
except Exception as e:
self.log.info("Dataframe contains no datetime feature.Err.Msg: \n"+str(e))
pass
try:
##Below commented part using normalize with time delta, find point anomalies.But not used,just for reference.
##get day to check difference
#date_f = dataframe[datetime_column].dt.normalize()
##compare successive rows and identify group size
#dataframe['anomaly_value'] = np.where(dataframe[datetime_column].groupby(date_f.ne(date_f.shift()).cumsum()).transform('size').gt(1),'subsequence_anomaly', 'Point_anomaly')
##Using get_timepattern method
aion_gettimegranularity_obj=aion_gettimegranularity(dataframe,datetime_column)
anomaly_info_df=aion_gettimegranularity_obj.get_granularity()
except Exception as e:
self.log.info("find_point_subsequence_anomalies,: aion_gettimegranularity err msg:: \n"+str(e))
self.log.info("find_point_subsequence_anomalies,: anomaly_info_df: \n"+str(anomaly_info_df))
except Exception as e:
self.log.info("find_point_subsequence_anomalies,: err msg:: \n"+str(e))
return anomaly_info_df
## Auto encoder time series function call
## dataframe info() not working for py logging, so workaround we can get information in buffer and log it.
def get_df_info(self,df):
from io import StringIO
buf = StringIO()
df.info(buf=buf)
#self.log.info(buf.getvalue())
return buf.getvalue()
## Method to detect time series based anomalies in user data. Using both lstm and dense based autoencoder approaches.
def aionAEAnomalyTS(self,df,test_size_perc,target,time_steps,dropout,mv_unique_feature_ad):
ae_hyperparameter=self.paramSpace
anomaly_algorithm=self.anomalyMethod
# test_size=float(self.testSize)
test_size=0.0
# train_size=1-test_size
train_size=1-test_size
# train_size_perc=train_size*100
train_size=int(len(df) * train_size)
try:
timeseries_layers=ae_hyperparameter['timeseries_layers']
## Here we are checking whether to use only LSTM layers for dnn or dense layers. Dense layers better for predicting point as well sequence anomalies in time series.
if (timeseries_layers.lower() == 'lstm'):
try:
## Need to get normalized data for threshold calculation.
data_mean=df.mean(axis=0)
data_std=df.std(axis=0)
data=(df-data_mean)/data_std
# train, test = df[:train_size], df[train_size:]
train, test = data[:train_size], data[train_size:]
test=train
test1=test ## Need to copy test data
train_index=train.index
test_index=test.index
cols = df.columns
# train, test = train_test_split(df, test_size=test_size,random_state=42)
X_train, y_train = self.create_dataset(
train,
train,
time_steps
)
X_test, y_test = self.create_dataset(
test,
test,
time_steps )
n_dims=X_train.shape[2]
n_timesteps=X_train.shape[1]
opt=ae_hyperparameter['optimizer']
loss_fn=ae_hyperparameter["loss"]
epochs=int(ae_hyperparameter['epochs'])
batch_size=int(ae_hyperparameter['batch_size'])
learning_rate=float(ae_hyperparameter['learning_rate'])
n_bottleneck=int(ae_hyperparameter['latentspace_size'])
units=int(ae_hyperparameter['hidden_units'])
activation=ae_hyperparameter['activation']
##For task 20731
minimum_threshold_user = str(ae_hyperparameter['min_threshold'])
maximum_threshold_user = str(ae_hyperparameter['max_threshold'])
autoencoder=self.aetsmodel_lstm(n_dims, n_timesteps, n_bottleneck,units,activation,X_train)
##To save file
# cwd = os.path.abspath(os.path.dirname(__file__))
cwd=self.deployLocation
try:
artifact_dir=os.path.normpath(os.path.join(cwd,'output','autoenc_artifact_dir'))
if not os.path.isdir(artifact_dir):
os.makedirs(artifact_dir)
except Exception as e:
self.log.info("<---- Autoencoder artifact_dir path error. Error Msg: ---->"+str(e))
#dl callback fn to get best loss fn, early stopping & model checkpoint call backs
es,mc=self.callbacks(filename = artifact_dir, patience = 5, monitor = "val_loss")
self.model_compile(autoencoder,learning_rate, loss_fn, opt)
X_train = np.reshape(X_train,(X_train.shape[0],X_train.shape[1],X_train.shape[2]))
X_test = X_test.reshape((X_test.shape[0], X_test.shape[1],n_dims))
# y_test = y_test.reshape((y_test.shape[0], y_test.shape[1], n_dims))
model_hist = autoencoder.fit(
X_train, X_train,
epochs=epochs,
batch_size=batch_size,
validation_split=0.1,
shuffle=False,callbacks = [mc, es]
)
model_info=self.summary(autoencoder)
X_train_pred = autoencoder.predict(X_train)
train_mae_loss = np.mean(np.abs(X_train_pred - X_train), axis=1)
## Task 20731
if ((minimum_threshold_user and minimum_threshold_user.strip()) and (maximum_threshold_user and maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = float(minimum_threshold_user)
elif ((minimum_threshold_user.strip()) and (not maximum_threshold_user.strip())):
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = float(minimum_threshold_user)
elif ((not minimum_threshold_user.strip()) and (maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
else:
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
# threshold = np.mean(train_mae_loss) + np.std(train_mae_loss)
self.log.info("Anomaly threshold max value based on loss fn (MAE): "+str(threshold))
self.log.info("Anomaly threshold min value based on loss fn (MAE): "+str(min_threshold))
X_test_pred = autoencoder.predict(X_test)
test_mae_loss = np.mean(np.abs(X_test_pred - X_test), axis=1)
test_score_df = pd.DataFrame(index=test_index[time_steps:])
if (n_dims >1):
columns = [f'loss_{num}' for num in range(n_dims)]
# test_score_df = pd.DataFrame(test_mae_loss, columns=columns, index=test_index[time_steps:])
test_score_df['loss'] = test_mae_loss.mean(axis=1)
else:
test_score_df['loss'] = test_mae_loss
test_score_df['max_threshold'] = threshold
test_score_df['min_threshold'] = min_threshold
test_score_df['anomaly_value'] = (test_score_df.loss > test_score_df.max_threshold)
test_score_df['anomaly_value'] = (test_score_df.loss < test_score_df.min_threshold)
## Newly added for lstm issue
## if coming dataframe have datetime index , copy it before concat (different indexed dfs)
import pandas.api.types as ptypes
# if (isinstance(test_score_df, pd.DatetimeIndex) and isinstance(df, pd.DatetimeIndex)):
test_cp_index=None
if (ptypes.is_datetime64_dtype(test_score_df.index) and ptypes.is_datetime64_dtype(df.index)):
# self.log.info("test_score_df and df have datetime index cols")
test_cp_index=test_score_df.index
df_cp_index=df.index
test_score_df=test_score_df.reset_index()
df=df.reset_index() ##self.datetimeFeature
test_score_df.dropna()
try:
test_score_df[self.datetimeFeature]=pd.to_datetime(test_score_df[self.datetimeFeature])
df[self.datetimeFeature]=pd.to_datetime(df[self.datetimeFeature])
except:
pass
try:
final_df=pd.DataFrame()
cols_to_use = df.columns.difference(test_score_df.columns)
final_df = pd.merge(test_score_df, df[cols_to_use], left_index=True, right_index=True, how='inner')
except Exception as e:
self.log.info("final_df creation err msg: \n: "+str(e))
else:
test_index=test_score_df.reset_index(drop=True)
test_cp_index=test_index.index
df_index=df.reset_index(drop=True)
final_df=pd.DataFrame()
final_df = test_score_df.join(df)
final_df.dropna()
##Again set datetime index to dataframes,drop datetime feature column and set it as index.
try:
final_df.set_index(self.datetimeFeature,inplace=True)
df.set_index(self.datetimeFeature,inplace=True)
df.drop(self.datetimeFeature,axis=1,inplace=True)
final_df.drop(self.datetimeFeature,axis=1,inplace=True)
except:
pass
## Below commented code used to print df.info() in log file (using get_df_info() methos).
# self.log.info("anomaly final_df info: \n")
# buf_info=self.get_df_info(final_df)
# self.log.info(buf_info)
# final_df=pd.DataFrame()
##Getback the datetime index back
final_df.index=test_cp_index
normal_prediction_df=test_score_df.loc[test_score_df['anomaly_value']==False]
anomaly_prediction_df=test_score_df.loc[test_score_df['anomaly_value']==True]
## Newly added for lstm issue
anomaly_prediction_df=pd.merge(anomaly_prediction_df, final_df, on=['loss', 'max_threshold','min_threshold', 'anomaly_value'], how="left")
# anomaly_prediction_df.fillna(anomaly_prediction_df.mean(), inplace=True)
anomaly_prediction_df['anomaly_value'] = anomaly_prediction_df['anomaly_value'].replace([np.inf, -np.inf], np.nan)
# anomaly_prediction_df['anomaly_value'] = anomaly_prediction_df['anomaly_value'].replace([np.inf, -np.inf], np.nan)
final_df['anomaly_value'] = final_df['anomaly_value'].replace([np.inf, -np.inf], np.nan)
anomaly_prediction_df['anomaly_value'] = anomaly_prediction_df['anomaly_value'].replace({True: 1, False: 0})
final_df['anomaly_value'] = final_df['anomaly_value'].replace({True:1, False: 0})
#make sure no nan values after dataframe operations
anomaly_prediction_df.dropna()
final_df.dropna()
# anomal_loss_threshold=anomaly_prediction_df #use if we want to save loss and threshold as dataframe info.
self.log.info("Anomaly data with loss and threshold informations: \n"+str(anomaly_prediction_df))
""" Saving anomaly plots in target->output->anomaly_plot folder """
## Goto if cond for multivariate whole dataset anomaly prediction, else goto else part for feature based ad prediction.
if (mv_unique_feature_ad.lower()=='false'):
for col in df.columns:
df_subset = anomaly_prediction_df[col]
fig, ax = plt.subplots()
df[col].plot(legend=False, ax=ax)
df_subset.plot(legend=False, ax=ax, color="r")
plot_name=col
ax.set_title(plot_name+"_Anomaly Data Plot")
ax.set_xlabel("DateTime")
ax.set_ylabel("Values")
plot_name=plot_name+'_'+'anomalyplot.png'
try:
plot_dir=os.path.normpath(os.path.join(cwd,'output','anomaly_plot'))
if not os.path.isdir(plot_dir):
os.makedirs(plot_dir)
plotpath=str(plot_dir)+'/'+plot_name
except Exception as e:
self.log.info("<---- plot_dir path error. Error Msg: ---->"+str(e))
if os.path.exists(plotpath):
os.remove(plotpath)
plt.savefig(plotpath)
# plt.savefig(str(plot_dir)+'/'+plot_name)
plt.clf()
plt.cla()
plt.close()
else:
df_subset = anomaly_prediction_df
fig, ax = plt.subplots()
df.plot(legend=False, ax=ax)
ax.set_title("Anomaly Data Plot")
ax.set_xlabel("X values")
ax.set_ylabel("Y Values")
df_subset.plot(legend=False, ax=ax, color="r")
plot_name=df.columns[0]
ax.set_title(plot_name+"_Anomaly Data Plot")
# ax.set_xlabel("DateTime")
# ax.set_ylabel("Values")
# plot_name=df.columns[0]
plot_name=plot_name+'_'+'anomalyplot.png'
try:
plot_dir=os.path.normpath(os.path.join(cwd,'output','anomaly_plot'))
if not os.path.isdir(plot_dir):
os.makedirs(plot_dir)
plotpath=str(plot_dir)+'/'+plot_name
except Exception as e:
self.log.info("<---- plot_dir path error. Error Msg: ---->"+str(e))
if os.path.exists(plotpath):
os.remove(plotpath)
plt.savefig(plotpath)
# plt.savefig(str(plot_dir)+'/'+plot_name)
plt.clf()
plt.cla()
plt.close()
#process dt feature and save anomalies.
datetime_column=str(self.datetimeFeature)
try:
anomaly_prediction_df=self.find_point_subsequence_anomalies(datetime_column,anomaly_prediction_df)
# normal_prediction_df=self.find_point_subsequence_anomalies(datetime_column,normal_prediction_df)
except:
##If any issue in time series point anomaly detection, skip it.
self.log.info("Detecting point anomalies have some issue,check datetime feature.")
pass
combined_df=pd.concat([anomaly_prediction_df,normal_prediction_df],ignore_index=True)
combined_df['anomaly_value']=combined_df['anomaly_value'].fillna('Normal_Data')
## If categorical features in original df, then inverse transform the values.
anomaly_prediction_df['anomaly_value'] = anomaly_prediction_df['anomaly_value'].replace({1: "Anomaly", 0: "Normal"})
final_df['anomaly_value'] = final_df['anomaly_value'].replace({1: "Anomaly", 0: "Normal"})
##Now we are storing anomaly log (as dataframe) based on two options: 1. Anomalies based on all features, 2. Anomalies based on each individual feature.
if (mv_unique_feature_ad.lower()=='true'):
## Multivariate and saving individual feature based anomalies
self.save_anomalyvalues(anomaly_prediction_df,(str(feature_name)+'_ts_anomaly_dataframe'))
# self.save_anomalyvalues(combined_df,(str(feature_name)+'_ts_overall_dataframe'))
try:
final_df=self.merge_pre_post_dfs(final_df)
except Exception as e:
self.log.info("Anomaly Detection Merge df exception:\n"+str(e))
#If merge fails, just out!
pass
self.save_anomalyvalues(final_df,(str(feature_name)+'_ts_overall_dataframe'))
## If we want anomaly dataframe with loss and threshold for each values (rows),please uncomment below line
# self.save_anomalyvalues(anomal_loss_threshold,'ts_anomaly_dataframe_lt'))
## Save actual test data test_score_df
#self.save_anomalyvalues(test_score_df,(str(feature_name)+'_testdata'))
else:
self.save_anomalyvalues(anomaly_prediction_df,'ts_anomaly_dataframe')
# self.save_anomalyvalues(combined_df,'ts_normal_anomaly_dataframe')
try:
final_df=self.merge_pre_post_dfs(final_df)
except Exception as e:
self.log.info("Anomaly Detection Merge df exception:\n"+str(e))
#If merge fails, just out!
pass
self.save_anomalyvalues(final_df,'ts_overall_dataframe')
## If we want anomaly dataframe with loss and threshold for each values (rows),please uncomment below line
# self.save_anomalyvalues(anomal_loss_threshold,'ts_anomaly_dataframe_lt'))
## Save actual test data test_score_df
#self.save_anomalyvalues(test_score_df,'testdata')
anomaly_info_df=final_df
self.log.info("<---- Autoencoder time series data anomalies: ---->"+str(anomaly_prediction_df))
self.log.info("<---- Autoencoder time series:Number of anomalies in data:: ---->"+str(len(anomaly_prediction_df)))
# return model
except Exception as e:
self.log.info("AD lstm traceback error: \n"+str(traceback.format_exc()))
## Dense layer based time series AD, most real world usecases, it is working best compared to lstm based..
elif (timeseries_layers.lower() == 'dense'):
try:
feature_name=df.columns
feature_name = ' '.join(map(str, feature_name))
try:
#Passing whole data,so test size set as zero.
test_size=0.0
# train_size=1-test_size
train_size=1-test_size
# train_size_perc=train_size*100
train_size=int(len(df) * train_size)
train_data,test_data = df[:train_size], df[train_size:]
test_data=train_data
except:
#If any error comes,us sklearn train test split
train_data,test_data = train_test_split(df,test_size=test_size,random_state=42)
pass
test_index=test_data.index ## to get datetime index
units=int(ae_hyperparameter['hidden_units'])
latent_units=int(ae_hyperparameter['latentspace_size'])
activation=ae_hyperparameter['activation']
##For task 20731
minimum_threshold_user = str(ae_hyperparameter['min_threshold'])
maximum_threshold_user = str(ae_hyperparameter['max_threshold'])
train_data=train_data.values
test_data=test_data.values
## tss is time series flag, true or false
autoencoder = AeDetector(train_data,test_data,units,latent_units,activation)
opt=ae_hyperparameter['optimizer']
loss_fn=ae_hyperparameter["loss"]
epochs=int(ae_hyperparameter['epochs'])
batch_size=int(ae_hyperparameter['batch_size'])
learning_rate=float(ae_hyperparameter['learning_rate'])
cwd=self.deployLocation
try:
artifact_dir=os.path.normpath(os.path.join(cwd,'output','autoenc_artifact_dir'))
if not os.path.isdir(artifact_dir):
os.makedirs(artifact_dir)
except Exception as e:
self.log.info("<---- artifact_dir path error. Error Msg: ---->"+str(e))
es,mc=self.callbacks(filename = artifact_dir, patience = 5, monitor = "val_loss")
self.model_compile(autoencoder,learning_rate, loss_fn, opt)
# autoencoder.compile(optimizer='adam', loss='mae')
autoencoder.fit(train_data, train_data, epochs = epochs, batch_size=batch_size, validation_data=(test_data, test_data),callbacks = [mc, es])
# reconstructed = autoencoder(train_data)
reconstructed = autoencoder.predict(train_data)
train_mae_loss = tf.keras.losses.mae(reconstructed, train_data)
## Task 20731
if ((minimum_threshold_user and minimum_threshold_user.strip()) and (maximum_threshold_user and maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = float(minimum_threshold_user)
elif ((minimum_threshold_user.strip()) and (not maximum_threshold_user.strip())):
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = float(minimum_threshold_user)
elif ((not minimum_threshold_user.strip()) and (maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
else:
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
# threshold = np.mean(train_mae_loss) + np.std(train_mae_loss)
self.log.info("Anomaly threshold max value based on loss fn (MAE): "+str(threshold))
self.log.info("Anomaly threshold min value based on loss fn (MAE): "+str(min_threshold))
test_labels=None
if (len(self.datetimeFeature) >= 1):
time_series_data="True"
else:
time_series_data="False"
pred,test_score_df,actual_data,anomaly_info_df = self.prediction(autoencoder, test_data,min_threshold, threshold,test_labels,time_series_data,time_steps,test_index)
# normal_prediction_df=(anomaly_info_df[~anomaly_info_df['anomaly_value']])
normal_prediction_df=anomaly_info_df.loc[anomaly_info_df['anomaly_value']==False]
anomaly_prediction_df=anomaly_info_df.loc[anomaly_info_df['anomaly_value']==True]
#Below ts_dataframe_anomaly not for production, just for testing purpose. If uncommented, comment it.
#self.save_anomalyvalues(anomaly_info_df,'ts_dataframe_normal')
# anomal_loss_threshold=anomaly_prediction_df #use if we want to save loss and threshold as dataframe info.
self.log.info("Anomaly data with loss and threshold informations: \n"+str(anomaly_prediction_df))
# anomaly_prediction_df_plot=anomaly_prediction_df
""" Saving anomaly plots in target->output->anomaly_plot folder """
## Only for multivariate (all features) based anomaly data plot
## Use of the below part if anomaly df columns came as numerical columns.
# if not (df.columns.equals(anomaly_prediction_df.columns)):
# num_cols = []
# try:
# num_cols=[num_cols.append(float(col)) for col in anomaly_prediction_df.columns.values]
# except ValueError:
# pass
# #Dense layer scaler conversion makes column names as int values, so here find the int cols and rename to original names.
# if (num_cols):
# anomaly_prediction_df=anomaly_prediction_df[num_cols]
# anomaly_prediction_df.columns=df.columns
# normal_prediction_df=normal_prediction_df[num_cols]
# normal_prediction_df.columns=df.columns
## Goto if cond for multivariate whole dataset anomaly prediction, else goto else part for feature based ad prediction.
if (mv_unique_feature_ad.lower()=='false'):
# for col in df.columns:
for col in actual_data.columns:
df_subset = anomaly_prediction_df[col]
fig, ax = plt.subplots()
df[col].plot(legend=False, ax=ax)
df_subset.plot(legend=False, ax=ax, color="r")
plot_name=col
ax.set_title(plot_name+"_Anomaly Data Plot")
ax.set_xlabel("DateTime")
ax.set_ylabel("Values")
plot_name=plot_name+'_'+'anomalyplot.png'
try:
plot_dir=os.path.normpath(os.path.join(cwd,'output','anomaly_plot'))
if not os.path.isdir(plot_dir):
os.makedirs(plot_dir)
plotpath=str(plot_dir)+'/'+plot_name
except Exception as e:
self.log.info("<---- plot_dir path error. Error Msg: ---->"+str(e))
if os.path.exists(plotpath):
os.remove(plotpath)
plt.savefig(plotpath)
# plt.savefig(str(plot_dir)+'/'+plot_name)
plt.clf()
plt.cla()
plt.close()
else:
df_subset = anomaly_prediction_df
fig, ax = plt.subplots()
df.plot(legend=False, ax=ax)
ax.set_title("Anomaly Data Plot")
ax.set_xlabel("DateTime")
ax.set_ylabel("Values")
df_subset.plot(legend=False, ax=ax, color="r")
plot_name=df.columns[0]
ax.set_title(plot_name+"_Anomaly Data Plot")
# ax.set_xlabel("DateTime")
# ax.set_ylabel("Values")
# plot_name=df.columns[0]
plot_name=plot_name+'_'+'anomalyplot.png'
try:
plot_dir=os.path.normpath(os.path.join(cwd,'output','anomaly_plot'))
if not os.path.isdir(plot_dir):
os.makedirs(plot_dir)
plotpath=str(plot_dir)+'/'+plot_name
except Exception as e:
self.log.info("<---- plot_dir path error. Error Msg: ---->"+str(e))
if os.path.exists(plotpath):
os.remove(plotpath)
plt.savefig(plotpath)
# plt.savefig(str(plot_dir)+'/'+plot_name)
plt.clf()
plt.cla()
plt.close()
datetime_column=str(self.datetimeFeature)
# anomaly_prediction_df=self.find_point_subsequence_anomalies(datetime_column,anomaly_prediction_df)
# normal_prediction_df=self.find_point_subsequence_anomalies(datetime_column,normal_prediction_df)
try:
anomaly_prediction_df=self.find_point_subsequence_anomalies(datetime_column,anomaly_prediction_df)
# normal_prediction_df=self.find_point_subsequence_anomalies(datetime_column,normal_prediction_df)
except:
self.log.info("Detecting point anomalies have some issue,check datetime feature.")
##Just pass if datetime column provides issue, use without datetime column info
pass
combined_df=pd.concat([anomaly_prediction_df,normal_prediction_df],ignore_index=True)
combined_df['anomaly_value']=combined_df['anomaly_value'].fillna('Normal_Data')
## If categorical features in original df, then inverse transform the values.
try:
# anomaly_info_df['anomaly_value']=anomaly_info_df['anomaly_value'].astype(str).replace(replace_values_F,'NormalDataPoint', regex=True)
self.naming_anomalyvalues(anomaly_info_df)
except Exception as e:
self.log.info("anomaly_info_df exception err msg: \n"+str(e))
##Now we are storing anomaly log (as dataframe) based on two options: 1. Anomalies based on all features, 2. Anomalies based on each individual feature.
if (mv_unique_feature_ad.lower()=='true'):
## Multivariate and saving individual feature based anomalies
self.save_anomalyvalues(anomaly_prediction_df,(str(feature_name)+'_ts_anomaly_dataframe'))
try:
anomaly_info_df=self.merge_pre_post_dfs(anomaly_info_df)
except Exception as e:
#If merge fails, just out!.
self.log.info("Anomaly Detection Merge df exception :\n"+str(e))
finally:
#check merging done or not, to be imp.
pass
self.save_anomalyvalues(anomaly_info_df,(str(feature_name)+'_ts_overall_dataframe'))
'''For overall ordered output,uncomment the below.'''
# self.save_anomalyvalues(combined_df,(str(feature_name)+'_ts_overall_dataframe_ordered'))
## If we want anomaly dataframe with loss and threshold for each values (rows),please uncomment below line
# self.save_anomalyvalues(anomal_loss_threshold,'ts_anomaly_dataframe_lt')
## Save actual test data actual_data
#self.save_anomalyvalues(actual_data,(str(feature_name)+'_testdata'))
else:
self.save_anomalyvalues(anomaly_prediction_df,'ts_anomaly_dataframe')
try:
anomaly_info_df=self.merge_pre_post_dfs(anomaly_info_df)
except Exception as e:
#If merge fails, just out!.
self.log.info("Anomaly Detection Merge df exception :\n"+str(e))
finally:
#check merging done or not, to be imp.
pass
self.save_anomalyvalues(anomaly_info_df,'ts_overall_dataframe')
#Ordered data
# self.save_anomalyvalues(combined_df,'ts_overall_dataframe_ordered')
## If we want anomaly dataframe with loss and threshold for each values (rows),please uncomment below line
# self.save_anomalyvalues(anomal_loss_threshold,'ts_anomaly_dataframe_lt'),
## Save actual test data test_score_df
#self.save_anomalyvalues(actual_data,'testdata')
self.log.info("<---- Autoencoder time series anomalies : ---->"+str(anomaly_prediction_df))
self.log.info("<---- Autoencoder time series, Number of anomalies in data: ---->"+str(len(anomaly_prediction_df)))
# self.save_anomalyvalues(anomaly_prediction_df,'ts_anomaly_dataframe')
except Exception as e:
self.log.info("dense layer anomaly error: \n"+str(traceback.format_exc()))
else:
self.log.info("Only LSTM and Dense layers supported for time series.")
except Exception as e:
self.log.info("<---- time series error msg: ---->"+str(e))
self.log.info("<---- time series error msg (detailed): ---->"+str(traceback.format_exc()))
return autoencoder,anomaly_prediction_df,anomaly_info_df
## To normalize data,use when necessary
def normalize_data(train_data,test_data):
min_val=tf.reduce_min(train_data)
max_val=tf.reduce_max(train_data)
train_data = (train_data - min_val)/(max_val - min_val)
test_data = (test_data - min_val)/(max_val - min_val)
#converte the data into float
train_data = tf.cast(train_data, dtype=tf.float32)
test_data = tf.cast(test_data, dtype=tf.float32)
return train_data,test_data
## Scaling data ,Not used because of our aion preprocessing data profiler option. use when necessary.
def getScaledData(method='standard', train_df=None, test_df=None, feature_col='feature'):
from sklearn.preprocessing import StandardScaler
if method == 'standard':
scaler = StandardScaler()
else:
scaler = MinMaxScaler()
scaler = scaler.fit(train_df[[feature_col]])
train_df['scaled_'+feature_col] = scaler.transform(train_df[[feature_col]])
test_df['scaled_'+feature_col] = scaler.transform(test_df[[feature_col]])
return train_df, test_df, scaler
## prediction fn
def prediction(self,model, data,min_threshold, threshold,test_labels,time_series_status,time_steps,test_index):
# data1=scaler.inverse_transform(data)
try:
df_new=self.df.drop(self.datetimeFeature,axis=1,inplace=False)
except:
df_new=self.df
try:
actual_data=pd.DataFrame(self.df,columns=df_new.columns)
except Exception as e:
actual_data=pd.DataFrame(self.df)
pass
n_features=data.shape[1]
self.log.info("prediction: number of features: \n"+str(n_features))
predicted_data = model.predict(data)
loss = tf.keras.losses.mae(predicted_data, data)
if (time_series_status.lower() == 'true'):
test_score_df = pd.DataFrame(index=test_index)
actual_data = actual_data.set_index(test_index)
anomaly_info_df=pd.DataFrame()
test_score_df['loss'] = loss
test_score_df['max_threshold'] = threshold
test_score_df['min_threshold'] = min_threshold
## Task 20731
#test_score_df['anomaly_value'] = test_score_df.apply(lambda x: x.loss > x.max_threshold or x.loss <= x.min_threshold, axis=1)
test_score_df['anomaly_value'] = np.where((test_score_df["loss"] > test_score_df["max_threshold"]) | (test_score_df["loss"] <= test_score_df["min_threshold"]), True, False)
anomaly_info_df = pd.concat([actual_data, test_score_df], axis=1)
else:
test_score_df = pd.DataFrame()
anomaly_info_df=pd.DataFrame()
test_score_df['loss'] = loss
#test_score_df['threshold'] = threshold
test_score_df['max_threshold'] = threshold
test_score_df['min_threshold'] = min_threshold
## Task 20731
#test_score_df['anomaly_value'] = (test_score_df.loss >= test_score_df.max_threshold)
#test_score_df['anomaly_value'] = (test_score_df.loss < test_score_df.min_threshold)
test_score_df['anomaly_value'] = np.where((test_score_df["loss"] > test_score_df["max_threshold"]) | (test_score_df["loss"] <= test_score_df["min_threshold"]), True, False)
anomaly_info_df = pd.concat([actual_data, test_score_df], axis=1)
return tf.math.less(loss, threshold),test_score_df,actual_data,anomaly_info_df
##Not used now, for data ploting purpose
# def plot(self,autoencoder,data, n):
# enc_img = autoencoder.encoder(data)
# dec_img = autoencoder.decoder(enc_img)
# plt.plot(data[n], 'b')
# plt.plot(dec_img[n], 'r')
# plt.fill_between(np.arange(data.shape[1]), data[n], dec_img[n], color = 'lightcoral')
# plt.legend(labels=['Input', 'Reconstruction', 'Error'])
# plt.show()
## autoencoder fn for non timeseries data
def ae_nontimeseriesmodelfn(self,df,target):
autoencoder=None
mv_unique_feature_ad=self.mv_featurebased_ad_status
#For supervised non time series problems, we need to remove datetime feature. This will help scaler algs process the numeric data only.
try:
if (target == ''):
try:
test_size=0.0
# train_size=1-test_size
train_size=1-test_size
# train_size_perc=train_size*100
train_size=int(len(df) * train_size)
train_data,test_data = df[:train_size], df[train_size:]
test_data=train_data
except:
test_size=float(self.testSize)
train_data,test_data = train_test_split(df,test_size=test_size,random_state=42)
pass
ae_hyperparameter=self.paramSpace
units=int(ae_hyperparameter['hidden_units'])
latent_units=int(ae_hyperparameter['latentspace_size'])
activation=ae_hyperparameter['activation']
##For task 20731
minimum_threshold_user = str(ae_hyperparameter['min_threshold'])
maximum_threshold_user = str(ae_hyperparameter['max_threshold'])
train_data=train_data.values
test_data=test_data.values
autoencoder = AeDetector(train_data,test_data,units,latent_units,activation)
opt=ae_hyperparameter['optimizer']
loss_fn=ae_hyperparameter["loss"]
# loss_fn='binary_crossentropy'
epochs=int(ae_hyperparameter['epochs'])
batch_size=int(ae_hyperparameter['batch_size'])
learning_rate=float(ae_hyperparameter['learning_rate'])
# autoencoder.save('../output/autoenc',save_format='tf')
# cwd = os.path.abspath(os.path.dirname(__file__))
cwd=self.deployLocation
try:
artifact_dir=os.path.normpath(os.path.join(cwd,'output','autoenc_artifact_dir'))
if not os.path.isdir(artifact_dir):
os.makedirs(artifact_dir)
except Exception as e:
self.log.info("<---- artifact_dir path error. Error Msg: ---->"+str(e))
es,mc=self.callbacks(filename = artifact_dir, patience = 5, monitor = "val_loss")
# es,mc=self.callbacks(filename = "../output/autoenc.sav", patience = 3, monitor = "val_loss")
self.model_compile(autoencoder,learning_rate, loss_fn, opt)
# autoencoder.compile(optimizer='adam', loss='mae')
autoencoder.fit(train_data, train_data, epochs = epochs, batch_size=batch_size, validation_data=(test_data, test_data),callbacks = [mc, es])
reconstructed = autoencoder(train_data)
train_mae_loss = tf.keras.losses.mae(reconstructed, train_data)
#threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
#min_threshold = np.mean(train_mae_loss)- 2*np.std(train_mae_loss)
## Task 20731
if ((minimum_threshold_user and minimum_threshold_user.strip()) and (maximum_threshold_user and maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = float(minimum_threshold_user)
elif ((minimum_threshold_user.strip()) and (not maximum_threshold_user.strip())):
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = float(minimum_threshold_user)
elif ((not minimum_threshold_user.strip()) and (maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
else:
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
# threshold = np.mean(train_mae_loss) + np.std(train_mae_loss)
self.log.info("Anomaly Upper Threshold value based on loss fn (MAE): "+str(threshold))
self.log.info("Anomaly lower_threshold value based on loss fn (MAE): "+str(min_threshold))
test_labels=None ## No test labels passed
pred,test_score_df,actual_data,anomaly_info_df = self.prediction(autoencoder, test_data, min_threshold,threshold,test_labels,'False',None,None)
# normal_prediction_df=(anomaly_info_df[~anomaly_info_df['anomaly_value']])
normal_prediction_df=anomaly_info_df.loc[anomaly_info_df['anomaly_value']==False]
anomaly_prediction_df=anomaly_info_df.loc[anomaly_info_df['anomaly_value']==True]
self.log.info("<---- Autoencoder (non timeseries) based anomaly detection, anomalies in data: ---->"+str(anomaly_prediction_df))
self.log.info("<---- Number of anomalies in data: ---->"+str(len(anomaly_prediction_df)))
self.save_anomalyvalues(anomaly_prediction_df,'anomaly_dataframe')
# combined_df=pd.concat([anomaly_prediction_df,normal_prediction_df],ignore_index=True)
self.log.info("<---- Autoencoder (non timeseries) based anomaly detection, overall data (both anomaly and non-anomaly ) in data: ---->"+str(anomaly_info_df))
# self.save_anomalyvalues(combined_df,'overall_dataframe')
## If categorical features in original df, then inverse transform the values.
try:
##anomaly_info_df,total dataframe.
self.naming_anomalyvalues(anomaly_info_df)
except Exception as e:
self.log.info("anomaly_info_df exception err msg: \n"+str(e))
##Now we are storing anomaly log (as dataframe) based on two options: 1. Anomalies based on all features, 2. Anomalies based on each individual feature.
if (mv_unique_feature_ad.lower()=='true'):
## Multivariate and saving individual feature based anomalies
self.save_anomalyvalues(anomaly_prediction_df,(str(feature_name)+'_anomaly_dataframe'))
try:
anomaly_info_df=self.merge_pre_post_dfs(anomaly_info_df)
except Exception as e:
#If merge fails, just out!.
self.log.info("Anomaly Detection Merge df exception :\n"+str(e))
finally:
#check merging done or not, to be imp.
pass
self.save_anomalyvalues(anomaly_info_df,(str(feature_name)+'_overall_dataframe'))
## Save actual test data actual_data
#self.save_anomalyvalues(actual_data,(str(feature_name)+'_testdata'))
else:
self.save_anomalyvalues(anomaly_prediction_df,'anomaly_dataframe')
try:
anomaly_info_df=self.merge_pre_post_dfs(anomaly_info_df)
except Exception as e:
#If merge fails, just out!.
self.log.info("Anomaly Detection Merge df exception :\n"+str(e))
finally:
#check merging done or not, to be imp.
pass
self.save_anomalyvalues(anomaly_info_df,'overall_dataframe')
#Ordered data
# self.save_anomalyvalues(combined_df,'ts_overall_dataframe_ordered')
## If we want anomaly dataframe with loss and threshold for each values (rows),please uncomment below line
# self.save_anomalyvalues(anomal_loss_threshold,'ts_anomaly_dataframe_lt'),
## Save actual test data test_score_df
#self.save_anomalyvalues(actual_data,'testdata')
self.log.info("<---- Autoencoder non time series / supervised problem anomalies : ---->"+str(anomaly_prediction_df))
#ploting
df_subset = anomaly_prediction_df
fig, ax = plt.subplots()
df.plot(legend=False, ax=ax)
df_subset.plot(legend=False, ax=ax, color="r")
ax.set_title("Anomaly Data Plot")
ax.set_xlabel("DateTime")
ax.set_ylabel("Values")
plot_name='anomalyplot.png'
try:
plot_dir=os.path.normpath(os.path.join(cwd,'output','anomaly_plot'))
if not os.path.isdir(plot_dir):
os.makedirs(plot_dir)
plotpath=str(plot_dir)+'/'+plot_name
except Exception as e:
self.log.info("<---- plot_dir path error. Error Msg: ---->"+str(e))
if os.path.exists(plotpath):
os.remove(plotpath)
plt.savefig(plotpath)
# plt.savefig(str(plot_dir)+'/'+plot_name)
plt.clf()
plt.cla()
plt.close()
else:
y=df[target]
X=df.drop(target, axis=1)
train_data,test_data,train_labels,test_labels=train_test_split(X,y,test_size=0.2,random_state=42)
count_classes = pd.value_counts(df[target], sort = True)
num_of_classes= len(count_classes)
self.log.info("train_data info: \n"+str(train_data.info()))
if (num_of_classes >= 2):
# scaler = StandardScaler()
# train_data = scaler.fit_transform(train_data)
# test_data = scaler.fit_transform(test_data)
# self.saveScaler(scaler)
train_labels = train_labels.astype(bool)
test_labels = test_labels.astype(bool)
n_train_data = train_data[train_labels]
n_test_data = test_data[test_labels]
# data1=scaler.inverse_transform(n_test_data)
n_test_data_actual=pd.DataFrame(n_test_data)
##anomaly data
an_train_data = train_data[~train_labels]
an_test_data = test_data[~test_labels]
n_train_data = train_data[train_labels]
n_test_data = test_data[test_labels]
ae_hyperparameter=self.paramSpace
# autoencoder = AeDetector(n_train_data,n_test_data)
activation=ae_hyperparameter['activation']
units=int(ae_hyperparameter['hidden_units'])
latent_units=int(ae_hyperparameter['latentspace_size'])
##For task 20731
minimum_threshold_user = str(ae_hyperparameter['min_threshold'])
maximum_threshold_user = str(ae_hyperparameter['max_threshold'])
autoencoder = AeDetector(n_train_data,n_test_data,units,latent_units,activation)
opt=ae_hyperparameter['optimizer']
loss_fn=ae_hyperparameter["loss"]
batch_size=int(ae_hyperparameter['batch_size'])
# loss_fn='binary_crossentropy'
epochs=int(ae_hyperparameter['epochs'])
learning_rate=float(ae_hyperparameter['learning_rate'])
cwd=self.deployLocation
try:
artifact_dir=os.path.normpath(os.path.join(cwd,'output','autoenc_artifact_dir'))
if not os.path.isdir(artifact_dir):
os.makedirs(artifact_dir)
except Exception as e:
self.log.info("<---- artifact_dir path error. Error Msg: ---->"+str(e))
es,mc=self.callbacks(filename = artifact_dir, patience = 5, monitor = "val_loss")
self.model_compile(autoencoder,learning_rate, loss_fn, opt)
# autoencoder.compile(optimizer='adam', loss='mae')
autoencoder.fit(n_train_data, n_train_data, epochs = epochs, batch_size=batch_size, validation_data=(n_test_data, n_test_data),callbacks = [mc, es])
model_info=self.summary(autoencoder)
self.log.info("<---- Auto encoder anomaly detection model information: ---->"+str(model_info))
# reconstructed = autoencoder(n_train_data)
reconstructed = autoencoder.predict(n_train_data)
#threshold = self.find_threshold(autoencoder, n_train_data)
train_mae_loss = tf.keras.losses.mae(reconstructed, n_train_data)
pred=tf.math.less(train_mae_loss, threshold)
## Task 20731
if ((minimum_threshold_user and minimum_threshold_user.strip()) and (maximum_threshold_user and maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = float(minimum_threshold_user)
elif ((minimum_threshold_user.strip()) and (not maximum_threshold_user.strip())):
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = float(minimum_threshold_user)
elif ((not minimum_threshold_user.strip()) and (maximum_threshold_user.strip())):
threshold = float(maximum_threshold_user)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
else:
threshold = np.mean(train_mae_loss) + 2*np.std(train_mae_loss)
min_threshold = np.mean(train_mae_loss) - 2*np.std(train_mae_loss)
self.log.info("Anomaly threshold max value based on loss fn (MAE): "+str(threshold))
self.log.info("Anomaly threshold min value based on loss fn (MAE): "+str(min_threshold))
pred,test_score_df,actual_data,anomaly_info_df = self.prediction(autoencoder, n_test_data, min_threshold,threshold,test_labels,'False',None,None)
# normal_prediction_df=(anomaly_info_df[~anomaly_info_df['anomaly_value']])
normal_prediction_df=anomaly_info_df.loc[anomaly_info_df['anomaly_value']==False]
# normal_prediction_df.to_csv('normal_prediction_df.csv')
# anomaly_prediction_df=(anomaly_info_df[anomaly_info_df['anomaly_value']])
anomaly_prediction_df=anomaly_info_df.loc[anomaly_info_df['anomaly_value']==True]
self.log.info("<---- Autoencoder (non timeseries) based anomaly detection, overall data (both anomaly and non-anomaly ) in data: ---->"+str(anomaly_info_df))
# self.save_anomalyvalues(combined_df,'overall_dataframe')
## If categorical features in original df, then inverse transform the values.
try:
##anomaly_info_df,total dataframe.
self.naming_anomalyvalues(anomaly_info_df)
except Exception as e:
self.log.info("anomaly_info_df exception err msg: \n"+str(e))
##Now we are storing anomaly log (as dataframe) based on two options: 1. Anomalies based on all features, 2. Anomalies based on each individual feature.
if (mv_unique_feature_ad.lower()=='true'):
## Multivariate and saving individual feature based anomalies
self.save_anomalyvalues(anomaly_prediction_df,(str(feature_name)+'_anomaly_dataframe'))
try:
anomaly_info_df=self.merge_pre_post_dfs(anomaly_info_df)
except Exception as e:
#If merge fails, just out!.
self.log.info("Anomaly Detection Merge df exception :\n"+str(e))
finally:
#check merging done or not, to be imp.
pass
self.save_anomalyvalues(anomaly_info_df,(str(feature_name)+'_overall_dataframe'))
## Save actual test data actual_data
#self.save_anomalyvalues(actual_data,(str(feature_name)+'_testdata'))
else:
self.save_anomalyvalues(anomaly_prediction_df,'anomaly_dataframe')
try:
anomaly_info_df=self.merge_pre_post_dfs(anomaly_info_df)
except Exception as e:
#If merge fails, just out!.
self.log.info("Anomaly Detection Merge df exception :\n"+str(e))
finally:
#check merging done or not, to be imp.
pass
self.save_anomalyvalues(anomaly_info_df,'overall_dataframe')
## Save actual test data test_score_df
#self.save_anomalyvalues(actual_data,'testdata')
self.log.info("<----Number of anomalies in data: ---->"+str(len(anomaly_prediction_df)))
""" Plot to show case anomalies, now commented, for testing purpose uncomment and check visually anomalies. """
#ploting
df_subset = anomaly_prediction_df
fig, ax = plt.subplots()
df.plot(legend=False, ax=ax)
df_subset.plot(legend=False, ax=ax, color="r")
# plt.show()
ax.set_title("Anomaly Data Plot")
ax.set_xlabel("DateTime")
ax.set_ylabel("Values")
plot_name='anomalyplot.png'
try:
plot_dir=os.path.normpath(os.path.join(cwd,'output','anomaly_plot'))
if not os.path.isdir(plot_dir):
os.makedirs(plot_dir)
plotpath=str(plot_dir)+'/'+plot_name
except Exception as e:
self.log.info("<---- plot_dir path error. Error Msg: ---->"+str(e))
if os.path.exists(plotpath):
os.remove(plotpath)
plt.savefig(plotpath)
# plt.savefig(str(plot_dir)+'/'+plot_name)
plt.clf()
plt.cla()
plt.close()
else:
self.log.info("<---- Check dataset and basic configurations. ---->")
except Exception as e:
self.log.info("<---- Non time series anomaly detection error msg: ---->"+str(e))
self.log.info("<---- Non time series anomaly detection error msg (detailed): ---->"+str(traceback.format_exc()))
return autoencoder,anomaly_prediction_df,anomaly_info_df
## Hyperparameter tuning autoencoders, not implemented
def hyperparamtuning_AE(self):
try:
self.log.info ("autoencoder hyperparam tuning.not implemented.")
except Exception as e:
self.log.info("autoencoder hyperparam tuning error: "+str(e))
pass
## randomsearch for dbscan
def hyperparamtuning_dbscan(self,model,tuner,Parameter_Trials,data):
params=model.get_params().keys()
try:
labels = model.labels_
#df = pd.DataFrame(labels)
try:
scorer = metrics.silhouette_score(data, labels)
except:
pass
if (tuner.lower() == 'randomsearch'):
# Parameters to try
cluster_labels = model.labels_
Random_Search = RandomizedSearchCV(model, Parameter_Trials, n_iter=50,cv=5, scoring='adjusted_rand_score', refit=True, n_jobs=1, verbose=5)
RandomSearchResults=Random_Search.fit(data)
# Fetching the best hyperparameters
best_params=RandomSearchResults.best_params_
# All the parameter combinations tried by RandomizedSearchCV
RandomSearchResults.cv_results_['params']
except Exception as e:
self.log.info("<---- dbscan hpt error msg: ---->"+str(e))
self.log.info("<---- dbscan hpt error msg (detailed): ---->"+str(traceback.format_exc()))
return best_params
## Reading aion postprocess data from target->AION_usecaseNo->data->postprocess data
def read_inputdata(self):
cwd=self.deployLocation
try:
in_path=os.path.normpath(os.path.join(cwd,'data'))
if not os.path.isdir(in_path):
self.log.info("<---- Anomaly detection target data folder not available.--->\n")
postprocesseddata=os.path.normpath(os.path.join(cwd,'data','postprocesseddata.csv'))
postprocessed_df=pd.read_csv(postprocesseddata)
except Exception as e:
self.log.info("<---- Anomaly detection target data folder not available, Reading postprocess csv file issue. Error Msg: ---->"+str(e))
return postprocessed_df
## Get original dataframe values using preprocess pipe after output data created.
##get_label_dict fn not used now. Use if preprocess_pipe based transform needed.
def get_label_dict(self, pipe):
label_dict = {}
dict_pipe={}
for (comp_name, component) in pipe.transformer_list:
if 'labelencoding' in comp_name:
i=1
for step in component.steps:
key='val'+'_'+str(i)
ordinalencoder=step[1]
dict_pipe[f'val_{i}']=ordinalencoder
# dict_pipe[key].append(ordinalencoder)
label_dict.update(dict_pipe)
i=i+1
return label_dict
else:
continue
return label_dict
## Decode label features using aion preprocessed_pipe model,not used now. If we need to use preprocess pipe for inverse transform,use below block.
def decoder_labeled_features(self,df):
import joblib
try:
cwd=self.deployLocation
# in_path=os.path.normpath(os.path.join(cwd,'data'))
if not os.path.isdir(in_path):
self.log.info("<---- Anomaly detection target model folder not available.--->\n")
preprocessed_pipe=os.path.normpath(os.path.join(cwd,'model','preprocess_pipe.pkl'))
model = joblib.load(preprocessed_pipe)
label_dict = get_label_dict(model)
encoder=label_dict.get('val_4')
num_cols = orig_data.select_dtypes(include=np.number).columns.tolist()
cat_cols = orig_data.select_dtypes(exclude=np.number).columns.tolist()
cat_col_actual=[]
for col in cat_cols:
try:
df1=encoder.inverse_transform(df[col])
cat_col_actual.append(col)
except:
pass
df1=pd.DataFrame(data=df1)
df1.columns=cat_cols
df2=df[num_cols]
df_anomalyinfo_col=df['anomaly_value']
df_list = [df2, df1, df_anomalyinfo_col] # List of your dataframes
combined_df = pd.concat(df_list, join='outer', axis=1).fillna(0)
except:
combined_df=None
pass
return combined_df
## save predicted data and actual data columns. For get back user original data features
#
def merge_pre_post_dfs(self,out_df=None):
cwd=self.deployLocation
anomaly_algorithm=str(self.anomalyMethod)
try:
in_path=os.path.normpath(os.path.join(cwd,'data'))
if not os.path.isdir(in_path):
self.log.info("<---- Anomaly detection target data folder not available.--->\n")
preprocessed_file=os.path.normpath(os.path.join(cwd,'data','preprocesseddata.csv'))
preprocessed_df=pd.read_csv(preprocessed_file)
## cat_cols will get categorical col from preprocessed, cat_diff_cols will get common cat col between output df and preprocessed.
cat_cols=preprocessed_df.select_dtypes(exclude=np.number).columns.tolist()
num_cols = preprocessed_df.select_dtypes(include=np.number).columns.tolist()
cat_diff_cols=list(set(cat_cols).intersection(out_df.columns.tolist()))
diff_cols=list(set(preprocessed_df.columns).difference(out_df.columns))
if (cat_diff_cols):
if (len(preprocessed_df) == len(out_df)):
#Drop each categorical col of original data from output df (which have numerical converted values). So, in merging can be done on perfect columns
try:
## get common categorical col name between actual and output df
for col in cat_diff_cols :
out_df.drop(col,axis=1,inplace=True)
except:
self.log.info("drop col not possible, pass the step.")
#Just continue
pass
diff_cols=list(set(preprocessed_df.columns).difference(out_df.columns))
try:
## Check any datetime column in output df and preprocesseddata
import pandas.api.types as ptypes
outdf_dt_index_check=ptypes.is_datetime64_dtype(out_df.index)
#Is output df have datetime col
if (outdf_dt_index_check):
if ((self.datetimeFeature.lower() !='na' and self.datetimeFeature)):
try:
preprocessed_df[self.datetimeFeature] = pd.to_datetime(preprocessed_df[self.datetimeFeature])
preprocessed_df.set_index(self.datetimeFeature, inplace=True)
except Exception as e:
self.log.info("Given data not contain datetime specified."+str(traceback.format_exc()))
## Below step ,making datetime index to date time column. for merging and droping purpose.
preprocessed_df.reset_index(inplace=True)
preprocessed_df.rename(columns={"index":self.datetimeFeature},inplace=True)
out_df.reset_index(inplace=True)
out_df.rename(columns={"index":self.datetimeFeature},inplace=True)
else:
## If no datetime column, we need to keep both dataframe index columns as unique. so making them as int index.
preprocessed_df.reset_index(inplace=True, drop=True)
out_df.reset_index(inplace=True, drop=True)
pass
## below part is to get status of index columns type (datetime,int or str), commented now. If needed for debug,pls use.
# dt_index_check=ptypes.is_datetime64_dtype(out_df.index)
# int_index_check=ptypes.is_numeric_dtype(out_df.index)
# str_index_check=ptypes.is_string_dtype(out_df.index)
## Get common column between preprocess and output df
try:
if (anomaly_algorithm.lower() == 'autoencoder'):
common_cols=out_df.drop(['loss','max_threshold','min_threshold','anomaly_value'],axis=1)
common_cols.drop(common_cols.filter(regex="Unname"),axis=1, inplace=True)
merge_on_cols=common_cols.columns.tolist()
combined_df = preprocessed_df.merge(out_df, on=merge_on_cols,how='inner')
## Drop duplicate based on columns except time
# drop_duplicate_on=merge_on_cols.extend(cat_diff_cols)
drop_duplicate_on=merge_on_cols+cat_diff_cols
combined_df = combined_df.drop_duplicates(drop_duplicate_on, keep=False)
else:
## otherwise, it is dbscan algorithm
common_cols=out_df.drop(['cluster','anomaly_value'],axis=1)
common_cols.drop(common_cols.filter(regex="Unname"),axis=1, inplace=True)
merge_on_cols=common_cols.columns.tolist()
combined_df = preprocessed_df.merge(out_df, on=merge_on_cols,how='inner')
## Drop duplicate based on columns except time
# drop_duplicate_on=merge_on_cols+cat_diff_cols
combined_df = combined_df.drop_duplicates(merge_on_cols, keep='last')
except:
combined_df=out_df
pass
## Just for reference, in future if you want different merge/join option
# combined_df = pd.merge(preprocessed_df[diff_cols],out_df, left_index=True, right_index=True, how='inner')
except Exception as e:
self.log.info("<---- merge error msg : ---->"+str(e))
self.log.info("<---- merge error msg (detailed): ---->"+str(traceback.format_exc()))
pass
## if both data frame have different columns (preprocessed and outdf)
else:
self.log.info("User data is preprocessed and data cleaning happened.So, actual data and processed data length mismatch. So,data records range may vary.")
try:
# combined_df=self.decoder_labeled_features(out_df)
combined_df = preprocessed_df.merge(out_df,on=num_cols,how='inner')
# combined_df = combined_df.drop_duplicates(cat_cols, keep='last')
combined_df = combined_df.drop_duplicates(num_cols, keep='last')
except:
## If nothing in merge works,then make outdf as final dataframe.
try:
## If above merge fails, change drop_duplicate hyperparam keep='last' last appearance of key occurance.
combined_df = preprocessed_df.merge(out_df,on=num_cols,how='inner')
combined_df = combined_df.drop_duplicates(cat_cols, keep=False)
except:
#If nothing is works, just keep out_df as combined df
combined_df=out_df
## if no common categorical col found between preprocessed and outdf.
else:
## If merge not works,then make outdf as final dataframe.
if (len(cat_cols) > 0):
try:
combined_df = preprocessed_df.merge(out_df,on=num_cols,how='inner')
combined_df = combined_df.drop_duplicates(cat_cols, keep='last')
except:
#make safe for return
combined_df=out_df
else:
##If no categorical features available
combined_df = preprocessed_df.merge(out_df,on=num_cols,how='inner')
combined_df = combined_df.drop_duplicates(num_cols, keep=False)
except Exception as e:
self.log.info("<---- Anomaly detection target data folder not available, dataframe merging issue. Error Msg: ---->"+str(e))
self.log.info("making output df as final merged data, no categorical column found in output anomaly data. It is user responsibility to check the anomaly data.")
#make safe for return
combined_df=out_df
return combined_df
## for module reusability, this below naming fn created.
def naming_anomalyvalues(self,df):
replace_values_T='|'.join(['TRUE','True','true'])
replace_values_F='|'.join(['FALSE','False','false'])
try:
df['anomaly_value']=df['anomaly_value'].astype(str).replace(replace_values_T,'AnomalyDataPoint', regex=True)
except:
df['anomaly_value']=df['anomaly_value'].replace(replace_values_T,'AnomalyDataPoint', regex=True)
df['anomaly_value']=df['anomaly_value'].astype(str).replace(replace_values_F,'NormalDataPoint', regex=True)
return df
## DBScan based anomaly detection
def dbscan_ad(self,data,eps,min_samples,cols):
try:
tuner='randomsearch'
Parameter_Trials={'eps':eps,
'min_samples':min_samples}
model = DBSCAN(algorithm='auto')
hist = model.fit(data)
pred = model.fit_predict(data)
best_params = self.hyperparamtuning_dbscan(model,tuner,Parameter_Trials,data)
self.log.info("<---- Best hyper parameters for dbscan: ---->"+str(best_params))
best_eps=best_params['eps']
best_min_samples=best_params['min_samples']
if (best_min_samples < len(cols)):
min_samples=len(cols)+1
if (best_eps < 0.2):
best_eps=0.2
self.log.info("best_eps: \n"+str(best_eps))
self.log.info("best_min_samples: \n"+str(best_min_samples))
best_model=DBSCAN(algorithm='auto',eps = best_eps, min_samples = best_min_samples)
hist = best_model.fit(data)
pred = best_model.fit_predict(data)
best_labels=best_model.labels_
cluster_name = ["Cluster"+str(i) for i in set(best_labels)]
# outliers = data[best_model.labels_ == -1]
outlier_df = data.copy()
outlier_df.loc[:,'cluster'] = best_model.labels_
outliers_final=outlier_df[outlier_df['cluster']==-1]
outliers_final['anomaly_value']=outliers_final['cluster']==-1
normaldata= outlier_df[outlier_df['cluster']!=-1]
self.log.info("<---- DBScan: Anomalies in data: ---->"+str(outliers_final))
self.log.info("<---- DBScan: Number of anomalies in data: ---->"+str(len(outliers_final)))
# num_cat_features=len(self.cat_cols)
try:
self.save_anomalyvalues(outliers_final,'dbscan_anomaly_dataframe')
self.save_anomalyvalues(normaldata,'dbscan_normaldata_dataframe')
outlier_df['anomaly_value']=outlier_df['cluster']==-1
outlier_df=self.naming_anomalyvalues(outlier_df)
##Convert results to original input data form for end user ease of understanding
try:
outlier_df=self.merge_pre_post_dfs(outlier_df)
except Exception as e:
self.log.info("Anomaly Detection Merge df exception:\n"+str(e))
#If merge fails, just out!
pass
self.save_anomalyvalues(outlier_df,'dbscan_overall_dataframe')
except Exception as e:
self.log.info("DBScan inversetransform err. msg: \n"+str(e))
no_clusters = len(set(best_labels)) - (1 if -1 in best_labels else 0)
self.log.info("<---- DBScan: No of clusters: ---->"+str(no_clusters))
n_noise_ = list(best_labels).count(-1)
## Ploting the dbscan clusters
plot_name='dbscan_anomalyplot.png'
fig, ax = plt.subplots()
ax.set_title("DBScan Clusters")
ax.scatter(data.iloc[:, 0], data.iloc[:, 1], c=best_labels)
outliers_plot = data[best_model.labels_ == -1]
ax.scatter(outliers_plot.iloc[:, 0], outliers_plot.iloc[:, 1], c='red')
cwd=self.deployLocation
try:
plot_dir=os.path.normpath(os.path.join(cwd,'output','anomaly_plot'))
if not os.path.isdir(plot_dir):
os.makedirs(plot_dir)
plotpath=str(plot_dir)+'/'+plot_name
except Exception as e:
self.log.info("<---- plot_dir path error. Error Msg: ---->"+str(e))
if os.path.exists(plotpath):
os.remove(plotpath)
plt.savefig(plotpath)
plt.clf()
plt.cla()
plt.close()
except Exception as e:
self.log.info("<---- dbscan error msg: ---->"+str(e))
self.log.info("<---- dbscan error msg (detailed): ---->"+str(traceback.format_exc()))
return best_model,outliers_final
## Inverse transform fn for categorical data
def inverse_transform(self,df,cat_cols,le_model):
df_new=pd.DataFrame()
df_new.index=df.index
df_reset_index=df.reset_index(drop=True)
for col in cat_cols:
df_reset_index[col] = le_model.inverse_transform(df_reset_index[col].astype(int))
df_reset_index.index=df_new.index
df=df_reset_index
return df
##If data comes without going via aion data profiler, we can use this below preprcessing fn ()
##Preprocess fn for categorical data , not used now.
def preprocessfn_categorical(self,df):
try:
cat_cols=self.cat_cols
preprocessed_df=None
le=preprocessing.LabelEncoder()
self.le_model=le
label_encoded_df = df.copy()
for col in cat_cols:
label_encoded_df[col]=le.fit_transform(label_encoded_df[col])
except Exception as e:
self.log.info("preprocessfn_categorical error traceback."+str(traceback.format_exc()))
return label_encoded_df,cat_cols
## Design pattern: Factory,Adapter. Detect antoencoder object or dbscan object based on input params. The interface can be used for anyother extention. Not created any abstract class.
##Main autoencoder based anomaly detection function, from here, sub modules will be called.
def mainAnomalyDetectionfn(self):
df=self.df
## reading post processed data from target->usecase->data directory
# df=self.read_inputdata()
## Below line overwrite incoming df with postprocesseddata
self.log.info("<----------- In autoencoder based anomaly detection algorithm main process module, the incoming dataframe information as below: \n")
buf_info=self.get_df_info(df)
self.log.info(buf_info)
model_location=None
time_series_data=None
# mv_unique_feature_ad=''
ae_hyperparameter=self.paramSpace
# self.log.info("mainAnomalyDetectionfn df: \n"+str(df))
self.log.info("paramSpace in mainano: \n"+str(self.paramSpace))
target=self.target
anomaly_algorithm=str(self.anomalyMethod)
mv_unique_feature_ad=self.mv_featurebased_ad_status
# df=self.dataload(datapath)
df.drop(df.filter(regex="Unname"),axis=1, inplace=True)
df.drop(df.filter(regex="truncated"),axis=1, inplace=True)
# cols2remove=df.columns[df.columns.str.startswith('Unname')]
# df.drop(cols2remove,axis=1,inplace=True)
# df.to_csv("C:\Users\jayaramakrishnans\AppData\Local\Programs\HCLTech\AION\data\target\actual_df.csv")
datetime_column=self.datetimeFeature
len_dtf=len(self.datetimeFeature)
##create datetime pattern source py file in target dir.
self.create_datetime_pyfile()
# cat_cols=None
if (self.datetimeFeature.lower() == 'na' or self.datetimeFeature==''):
len_dtf=0
if (len_dtf >= 1 ):
time_series_data="True"
else:
time_series_data="False"
self.datetimeFeature=''
try:
if (target != ''):
if (anomaly_algorithm.lower() == 'autoencoder'):
self.log.info("Supervised anomaly detection started.\n")
""" Below part for supervised time series anomaly detection.If timeseries anomaly (supervised) used in future, please uncomemnt and use below code snippet. """
# if (ae_hyperparameter['time_series_data'].lower() == 'true'):
# print("Given data is time series data and supervised learning category, because it target is labeled one.")
# datetime_column=str(self.datetimeFeature)
# time_steps=int(ae_hyperparameter['time_steps'])
# test_size_perc=int(ae_hyperparameter['test_size_perc'])
# df[datetime_column] = pd.to_datetime(df[datetime_column])
# df.set_index(datetime_column, inplace=True)
# self.aionAEAnomalyTS(df,test_size_perc,target,time_steps)
## Removing datetime feature for supervised autoencoder (currently timeseries anomaly not supported in supervised anomaly detection autoencoder)
test_size_perc=self.testSize*100
df=df.dropna()
model,anomaly_prediction_df,combined_df = self.ae_nontimeseriesmodelfn(df,target)
# print("*** End of Autoencoder based non time series Anomaly detection. *** \n")
self.log.info("*** End of Autoencoder based non time series Anomaly detection. ***")
features=df.columns
if (len(features)== 1):
# print("Problem type is Univariate time series anomaly detection.\n")
self.log.info("Problem type is Univariate time series anomaly detection.\n")
test_size_perc=self.testSize*100
df=df.dropna()
model,anomaly_prediction_df,combined_df = self.ae_nontimeseriesmodelfn(df,target)
elif (len(features) > 1):
df.drop(df.filter(regex="Unname"),axis=1, inplace=True)
test_size_perc=self.testSize*100
df=df.dropna()
model,anomaly_prediction_df,combined_df = self.ae_nontimeseriesmodelfn(df,target)
if (mv_unique_feature_ad.lower()=='true'):
self.log.info("\n\n *** Below is the anomaly values based on each feature of multivariate time series data. ***")
df.drop(df.filter(regex="Unname"),axis=1, inplace=True)
multivariate_cols= df.columns.values.tolist()
for col in multivariate_cols:
col=str(col)
## creating dataframe for one of feature in multivariate dataset.
multivariate_col_df = df[col].to_frame()
model,anomaly_prediction_df,combined_df = self.ae_nontimeseriesmodelfn(multivariate_col_df,target)
else:
if (anomaly_algorithm.lower() == 'dbscan'):
self.log.info("DBScan algorithm not supported for supervised training. \n")
else:
try:
self.log.info("Unsupervised anomaly detection started....\n")
if (anomaly_algorithm.lower() == 'autoencoder'):
test_size_perc=self.testSize*100
if (time_series_data.lower() == 'true'):
mv_unique_feature_ad=self.mv_featurebased_ad_status
dropout=float(ae_hyperparameter['dropout'])
datetime_column=str(self.datetimeFeature)
time_steps=int(ae_hyperparameter['time_steps'])
target=None
features=df.columns
if (len(features)== 1):
self.log.info("Problem type is Univariate time series anomaly detection.\n")
model,anomaly_prediction_df,combined_df = self.aionAEAnomalyTS(df,test_size_perc,target,time_steps,dropout,'False')
elif (len(features) > 1):
df.drop(df.filter(regex="Unname"),axis=1, inplace=True)
self.log.info("Problem type is Multivariate time series anomaly detection.\n")
self.log.info("*** Detecting anomaly in the given multi feature (Multivariate) data. Anomaly values based on all the features passed to the aion anomaly AE algorithm. ***")
model,anomaly_prediction_df,combined_df = self.aionAEAnomalyTS(df,test_size_perc,target,time_steps,dropout,'False')
if (mv_unique_feature_ad.lower()=='true'):
self.log.info("\n\n *** Below is the anomaly values based on each feature of multivariate time series data. ***")
df.drop(df.filter(regex="Unname"),axis=1, inplace=True)
multivariate_cols= df.columns.values.tolist()
# for i in range(1,len(features)):
for col in multivariate_cols:
col=str(col)
## creating dataframe for one of feature in multivariate dataset.
multivariate_col = df[col].to_frame()
feature_based_model,anomaly_prediction_df_mv,combined_df = self.aionAEAnomalyTS(multivariate_col,test_size_perc,target,time_steps,dropout,mv_unique_feature_ad)
""" Below code snippet is commented, the purpose is AION giving 3 option to user for time series, < Three types: univariate_tsd(single_timeseriesdetection), multivariate_tsd, multivariate_seperate_tsd>, based on that specific sub function called. """
# if (ae_hyperparameter['timeseries_ad_type'].lower() =='univariate_tsad'):
## univariate_tsd
# self.aionAEAnomalyTS(df,test_size_perc,target,time_steps,dropout)
# elif (ae_hyperparameter['timeseries_ad_type'].lower() =='multivariate_tsad'):
##multivariate_tsd
# if (len(features) <=1):
# # self.aionAEAnomalyTS(df,test_size_perc,target,time_steps,dropout)
# print("Given data looks like univariate data. Cannot apply multivariate. Check data and select appropriate timeseries anomaly detection option.")
# else:
# self.aionAEAnomalyTS(df,test_size_perc,target,time_steps,dropout)
# elif (ae_hyperparameter['timeseries_ad_type'].lower() =='multivariate_featurebased_tsad'):
## multivariate_seperate_tsd
# # features=df.columns
# df.drop(df.filter(regex="Unname"),axis=1, inplace=True)
# multivariate_cols= df.columns.values.tolist()
# print("multivariate_cols: \n",multivariate_cols)
# print("multivariate_cols type: \n",type(multivariate_cols))
# if (len(features) <=1):
# # self.aionAEAnomalyTS(df,test_size_perc,target,time_steps,dropout)
# print("Given data looks like univariate data. Cannot use multivariate.")
# else:
# # for i in range(1,len(features)):
# for col in multivariate_cols:
# print("processing multivariate feature name: ",col)
# col=str(col)
# multivariate_col = df[col].to_frame()
# print("multivariate_col type: \n",type(multivariate_col))
# self.aionAEAnomalyTS(multivariate_col,test_size_perc,target,time_steps,dropout)
# print("*** End of Autoencoder based time series Anomaly detection.*** \n")
self.log.info("*** End of Autoencoder based non time series Anomaly detection. ***")
else:
target=''
df=df.dropna()
model,anomaly_prediction_df,combined_df = self.ae_nontimeseriesmodelfn(df,target)
elif (anomaly_algorithm.lower() == 'dbscan'):
# df=df.dropna()
self.log.info("*** DBScan algorithm enabled. ***")
cols=df.columns
dbscan_hyperparameter=self.paramSpace
eps = list(dbscan_hyperparameter['eps'])
# eps=list(dbscan_hyperparameter['eps'])
# min_samples=list(dbscan_hyperparameter['min_samples'])
min_samples = list(dbscan_hyperparameter['min_samples'])
model,outliers = self.dbscan_ad(df,eps,min_samples,cols)
except Exception as e:
self.log.info("Unsupervised anomaly detection error msg: "+str(traceback.format_exc()))
##main ae AeDetector
except Exception as e:
self.log.info("<---- Main fn error msg of anomaly detection for debug purpose: ---->"+str(e))
self.log.info("<---- Main fn error msg of anomaly detection for debug purpose: (detailed): ---->"+str(traceback.format_exc()))
return model
## For testing standalone code
if __name__ == '__main__':
# print ("main function....")
target=None
df=None
hparams=None
AEAD=anomalyDetectionAE()
datapath,target,ae_hyperparameter = AEAD.mainAnomalyDetectionfn()
|
anomalyDetector.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import time
import os
import sys
import logging
from sklearn.metrics import accuracy_score, make_scorer
from sklearn.model_selection import train_test_split
from sklearn.svm import OneClassSVM
from sklearn.ensemble import IsolationForest
import pickle
from sklearn import metrics
import numpy as np
import pandas as pd
from learner.aion_matrix import aion_matrix
from learner.parameters import parametersDefine
from sklearn.metrics import f1_score
from sklearn import model_selection
from learner.anomalyDetectionAE import anomalyDetectionAE
class anomalyDetector(object):
def __init__(self):
self.log = logging.getLogger('eion')
def startanomalydetector(self,df,target,labelMaps,inlierLabels,learnerJson,model_type,saved_model,anomalyMethod,deployLocation,predicted_data_file,testPercentage,anomalyconfig,datetimeFeature,mv_featurebased_ad_status):
try:
self.log.info("startanomalydetector.... \n")
from io import StringIO
buf = StringIO()
df.info(buf=buf)
#self.log.info(buf.getvalue())
self.log.info("User data info : \n"+str(buf.getvalue()))
try:
df[datetimeFeature] = pd.to_datetime(df[datetimeFeature])
df.set_index(datetimeFeature, inplace=True)
#If still the datetime column exist in feature list, drop it. Because we already made datetime feature as index (datetimeindex)
df.drop(datetimeFeature,axis=1,inplace=True)
except Exception as e:
pass
ae_df=df
paramObj=parametersDefine()
anomalyMethod=anomalyMethod
inlierLabels=inlierLabels
anomalyDetectionType=""
inlierLabelList=inlierLabels.split(",")
self.log.info("<---- inlierLabels ---->"+inlierLabels)
self.log.info("<---- anomalyMethod ---->"+str(anomalyMethod))
if target != "":
self.log.info('Status:- |... AnomalyDetection: Supervised')
self.log.info("One class based anomaly Detection by relabeling data to fit one class models")
combinedString=""
dfStr=""
anomalyDetectionType="supervised"
if not anomalyMethod.lower() == "autoencoder": ##Added for auto encoder
self.log.info("startanomalydetector: df: \n"+str(df)) #task 12627
if labelMaps == {}:
for inlierVal in inlierLabelList:
inlier=inlierVal
dfStr = "x ==" + inlier + " or "
combinedString+= dfStr
func= combinedString.strip(" or ")
else:
for inlierVal in inlierLabelList:
try:
if inlierVal.isnumeric():
inlierVal = int(inlierVal)
# inlier=str(labelMaps[inlierVal]) ##Wrongly assigned inlier values to labelMaps dict key.
inlier = str(inlierVal)
dfStr = "x ==" + str(inlier) + " or "
combinedString+= dfStr
except Exception as e:
raise Exception(e)
func= combinedString.strip(" or ")
labelMaps={'InlierLabel':1,'NonInlierLabel':-1}
targetData=df[target]
df['anomaly'] = df[target].apply(lambda x: 1 if eval(func) else -1 )
anomtargetData=df['anomaly']
self.log.info("dataframe after relabeling the data")
self.log.info(df.head())
self.log.info("target column value counts with inliers and outliers")
self.log.info(df['anomaly'].value_counts())
df.drop([target, "anomaly"], axis=1, inplace=True)
outliers = anomtargetData[anomtargetData == -1]
self.log.info("outliers in data")
self.log.info(outliers.shape[0])
self.log.info("outlier fraction")
self.log.info(outliers.shape[0]/targetData.shape[0])
if int(testPercentage) != 0:
testSize= testPercentage/100
xtrain, xtest, ytrain, ytest = train_test_split(df, anomtargetData, test_size = testSize)
else:
xtrain =df
xtest =df
ytrain=anomtargetData
ytest =anomtargetData
if anomalyMethod.lower() == "isolationforest":
modelName="isolationforest"
paramSpace=anomalyconfig['modelParams']['IsolationForest']
paramDict =paramObj.paramDefine(paramSpace,'random')
ftwo_scorer = make_scorer(accuracy_score)
isolation_forest = model_selection.RandomizedSearchCV(IsolationForest(), paramDict, scoring=ftwo_scorer, n_iter=10)
mod = isolation_forest.fit(xtrain,ytrain)
model = mod.best_estimator_
elif anomalyMethod.lower() == "oneclasssvm":
modelName="oneClassSVM"
fthree_scorer = make_scorer(accuracy_score)
paramSpace=anomalyconfig['modelParams']['oneclassSVM']
paramDict =paramObj.paramDefine(paramSpace,'random')
one_class = model_selection.RandomizedSearchCV(OneClassSVM(), paramDict, scoring=fthree_scorer, n_iter=10)
mod = one_class.fit(xtrain,ytrain)
model = mod.best_estimator_
elif anomalyMethod.lower() == "autoencoder":
modelName='autoencoder'
testSize=testPercentage/100
self.log.info("Aion Autoencoder anomaly detection started..")
paramSpace=anomalyconfig['modelParams']['AutoEncoder']
adae_obj=anomalyDetectionAE(ae_df,paramSpace,deployLocation,target,anomalyMethod,testSize,datetimeFeature,mv_featurebased_ad_status)
model=adae_obj.mainAnomalyDetectionfn()
self.log.info("Aion Autoencoder anomaly detection completed..")
else:
self.log.info("IsolationForest, OneClassSVM and autoencoder are supported models")
modelName = ""
model = ""
else:
self.log.info('Status:- |... AnomalyDetection: Unsupervised')
self.log.info("unsupervised anomaly detection")
anomalyDetectionType="unsupervised"
model=None
xtrain =df
xtest = df
ytrain = pd.DataFrame()
if anomalyMethod.lower() == "isolationforest":
paramSpace=anomalyconfig['modelParams']['IsolationForest']
paramDict =paramObj.paramDefine(paramSpace,'random')
modelName="isolationforest"
def scorer_f(estimator, X):
return np.mean(estimator.score_samples(X))
isolation_forest = model_selection.RandomizedSearchCV(IsolationForest(), paramDict, scoring=scorer_f, n_iter=10)
mod = isolation_forest.fit(xtrain)
self.log.info('---------> Best Param: '+str(mod.best_params_))
model = mod.best_estimator_
elif anomalyMethod.lower() == "oneclasssvm":
paramSpace=anomalyconfig['modelParams']['oneclassSVM']
paramDict =paramObj.paramDefine(paramSpace,'random')
modelName="oneClassSVM"
def scorer_f1(estimator, X):
return np.mean(estimator.score_samples(X))
one_class = model_selection.RandomizedSearchCV(OneClassSVM(), paramDict, scoring=scorer_f1, n_iter=10)
model = one_class.fit(xtrain)
self.log.info('---------> Best Param: '+str(model.best_params_))
model = model.best_estimator_
elif anomalyMethod.lower() == "autoencoder":
ae_df.drop(ae_df.filter(regex="Unname"),axis=1, inplace=True)
modelName='autoencoder'
testSize= testPercentage/100
self.log.info("Aion Autoencoder anomaly detection started..")
paramSpace=anomalyconfig['modelParams']['AutoEncoder']
adae_obj=anomalyDetectionAE(ae_df,paramSpace,deployLocation,target,anomalyMethod,testSize,datetimeFeature,mv_featurebased_ad_status)
model=adae_obj.mainAnomalyDetectionfn()
self.log.info("Aion Autoencoder anomaly detection completed..")
elif anomalyMethod.lower() == "dbscan":
testSize=testPercentage/100
ae_df.drop(ae_df.filter(regex="Unname"),axis=1, inplace=True)
modelName='dbscan'
self.log.info("Aion DBScan anomaly detection started..")
paramSpace=anomalyconfig['modelParams']['DBScan']
adae_obj=anomalyDetectionAE(ae_df,paramSpace,deployLocation,target,anomalyMethod,testSize,datetimeFeature,mv_featurebased_ad_status)
model=adae_obj.mainAnomalyDetectionfn()
self.log.info("Aion DBScan anomaly detection completed..")
else:
self.log.info("IsolationForest,OneClassSVM,autoencoder and DBScan are supported models")
modelName = ""
model = ""
self.log.info('Status:- |... AnomalyDetection Algorithm applied: '+modelName)
if (anomalyMethod.lower() == "autoencoder" or anomalyMethod.lower() == "dbscan"):
if (anomalyMethod.lower() == "autoencoder"):
## Since autoencoder is implemented using tf.keras, saving the model in h5 format. If we save it in .sav format will give 'TensorSliceReader constructor' error.
saved_model=saved_model.replace('.sav','')
filename = os.path.join(deployLocation,'model',saved_model)
model.save(filename,save_format="tf")
elif (anomalyMethod.lower() == "dbscan"):
filename = os.path.join(deployLocation,'model',saved_model)
pickle.dump(model, open(filename, 'wb'))
matrix=''
trainmatrix=''
accuracy = 0
else:
filename = os.path.join(deployLocation,'model',saved_model)
pickle.dump(model, open(filename, 'wb'))
loaded_model=pickle.load(open(filename, 'rb'))
pred_testData=loaded_model.predict(xtest)
pred_trainData = loaded_model.predict(xtrain)
pred_trainScore = loaded_model.decision_function(xtrain)
self.log.info("<--- predicted values of testdata --->")
self.log.info(pred_testData)
if anomalyDetectionType == "supervised" :
df_predicted = pd.DataFrame()
df_predicted['actual'] = ytest
df_predicted['predict'] = pred_testData
df_predicted.to_csv(predicted_data_file)
preds = pred_testData
targs = ytest
unique_elements_ytest, counts_elements_ytest = np.unique(targs, return_counts=True)
unique_elements_pred, counts_elements_pred = np.unique(preds, return_counts=True)
'''
for i in range(0,len(unique_elements_ytest)):
self.log.info("unique value :" +str(unique_elements_ytest[i]) + " count in input testdata: " + str(counts_elements_ytest[i]) +" count in predicted testdata: " + str(counts_elements_pred[i]))
self.log.info("\n")
'''
self.log.info("\n======= Performance matrix on Test Data ======")
aion_matrixobj = aion_matrix()
self.log.info("-------> Test Matrix: ")
matrix = aion_matrixobj.getClassificationPerformaceMatrix(targs,preds,labelMaps)
self.log.info("-------> Train Matrix: ")
trainmatrix = aion_matrixobj.getClassificationPerformaceMatrix(ytrain,pred_trainData,labelMaps)
#self.log.info("-------> Confusion Matrix: ")
self.log.info(metrics.confusion_matrix(targs,preds))
self.log.info("-------> accuracy for inliers: ")
accuracy = metrics.accuracy_score(targs, preds)
self.log.info(metrics.accuracy_score(targs, preds))
self.log.info("-------> precision for inliers --->")
self.log.info(metrics.precision_score(targs, preds))
self.log.info("-------> recall for inliers ---> ")
self.log.info(metrics.recall_score(targs, preds))
self.log.info("-------> f1 for inliers--->")
self.log.info(metrics.f1_score(targs, preds))
self.log.info("-------> area under curve (auc) for inliers --->")
self.log.info(metrics.roc_auc_score(targs, preds))
self.log.info("-------> precision for outliers --->")
self.log.info(1-metrics.precision_score(targs, preds))
self.log.info("-------> recall for outliers ---> ")
self.log.info(1-metrics.recall_score(targs, preds))
self.log.info("-------> f1 for outliers--->")
self.log.info(1-metrics.f1_score(targs, preds))
self.log.info("======= Performance matrix on Test Data End ======\n")
else:
df_predicted = xtrain
df_predicted['predict'] = pred_trainData
df_predicted['score'] = pred_trainScore
df_predicted.to_csv(predicted_data_file, index=False)
matrix = ''
trainmatrix = ''
accuracy = 'NA'
return modelName,model,matrix,trainmatrix,accuracy,labelMaps
except Exception as inst:
self.log.info("Error: anomalyDetector failed "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno)) |
regressionModel.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
from learner.optimizetechnique import OptimizationTq
import warnings
from learner.parameters import parametersDefine
from learner.defaultAlgos import defaultParams
from hyperopt import fmin, tpe, hp, STATUS_OK, Trials
import time
import logging
import os
import sys
import json
from sklearn.svm import SVR
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Lasso
from sklearn.linear_model import Ridge
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import RandomForestRegressor
from xgboost import XGBRegressor
from lightgbm import LGBMRegressor
from catboost import CatBoostRegressor
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error,make_scorer
from sklearn.metrics import mean_squared_error
from learner.aion_matrix import aion_matrix
from uncertainties.aionUQ import aionUQ
import mlflow
class RegressionModel():
def __init__(self,modelList,params,scoreParam,cvSplit,numIter,geneticParam,trainX,trainY,testX,testY,method,deployLocation):
self.modelList =modelList
self.params =params
self.trainX =trainX
self.trainY =trainY
self.testX = testX
self.testY = testY
self.method =method
self.scoreParam=scoreParam
self.cvSplit=cvSplit
self.numIter=numIter
self.geneticParam=geneticParam
self.log = logging.getLogger('eion')
self.deployLocation = deployLocation
self.uq_x_train = trainX
self.uq_x_test = testX
self.uq_y_train = trainY
self.uq_y_test = testY
self.AlgorithmNames={'Linear Regression':'LinearRegression','Lasso':'Lasso','Ridge':'Ridge','Decision Tree':'DecisionTreeRegressor','Random Forest':'RandomForestRegressor','Extreme Gradient Boosting (XGBoost)':'XGBRegressor','Light Gradient Boosting (LightGBM)': 'LGBMRegressor',
'Categorical Boosting (CatBoost)': 'CatBoostRegressor','Bagging (Ensemble)':'BaggingRegressor','Stacking (Ensemble)':'StackingRegressor','Voting (Ensemble)':'VotingRegressor','Neural Architecture Search':'NAS'}
self.modelToAlgoNames = {value: key for key, value in self.AlgorithmNames.items()}
def logMlflow(self, runName, params, metrices, estimator, algoName=None):
with mlflow.start_run(run_name = runName):
for k,v in params.items():
mlflow.log_param(k, v)
for k,v in metrices.items():
mlflow.log_metric(k, v)
if algoName == 'CatBoostRegressor':
mlflow.catboost.log_model(estimator, "model")
else:
mlflow.sklearn.log_model(estimator, "model")
model_uri = mlflow.get_artifact_uri("model")
""" for some dataset evaluate takes more than 90 min, so commenting till some solution is not found
evaluate_data = self.testX.copy()
evaluate_data['target'] = self.testY.copy()
mlflow.evaluate(model_uri, data=evaluate_data, targets='target', model_type="regressor")
del evaluate_data
"""
def regressionModelling(self,modelOrFeatureBased, code_configure):
paramObj=parametersDefine()
bestModel=''
bestParams={}
import sys
bestScore=-sys.float_info.max #bugfix 11656
scoredetails = ''
self.log.info('\n---------- Regression Model has started ----------')
try:
self.log.info('Status:- |... Search Optimization Method applied: '+self.method)
for modelName in self.modelList:
objClf = aion_matrix()
if modelName in ['Bagging (Ensemble)','Voting (Ensemble)','Stacking (Ensemble)','Neural Architecture Search']:
if modelName == 'Bagging (Ensemble)':
from ensemble.ensemble_bagging import ensemble_bagging
ensemble_bagging_obj = ensemble_bagging(self.params[modelName],self.scoreParam,0,0)
estimator,modelParams,score,model = ensemble_bagging_obj.ensemble_bagging__regressor(self.trainX,self.trainY,self.testX,self.testY)
if modelName == 'Stacking (Ensemble)':
from ensemble.ensemble_stacking import ensemble_stacking
ensemble_stacking_obj = ensemble_stacking(self.params[modelName],self.scoreParam)
estimator,modelParams,score,model = ensemble_stacking_obj.ensemble_stacking__regressor(self.trainX,self.trainY,self.testX,self.testY,self.modelList)
if modelName == 'Voting (Ensemble)':
from ensemble.ensemble_voting import ensemble_voting
ensemble_voting_obj = ensemble_voting(self.params[modelName],self.scoreParam)
estimator,modelParams,score,model = ensemble_voting_obj.ensemble_voting__regressor(self.trainX,self.trainY,self.testX,self.testY,self.modelList)
'''
if modelName == 'Neural Architecture Search':
from nas.aionNAS import aionNAS
objNAS = aionNAS('Regression',self.params[modelName],self.trainX,self.testX,self.trainY,self.testY,self.deployLocation)
estimator,modelParams,score,model=objNAS.nasMain(self.scoreParam)
'''
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(score)+',"ModelUncertainty":"NA"}'
if self.scoreParam == "r2":
if score > bestScore:
bestScore =score
bestModel =model
bestParams=modelParams
bestEstimator=estimator
else:
if abs(score) < bestScore or bestScore == -sys.float_info.max:
bestScore =abs(score)
bestModel =model
bestParams=modelParams
bestEstimator=estimator
self.log.info('Status:- |... ML Algorithm applied: '+modelName)
self.log.info('Status:- |... Score: '+objClf.get_print_score(self.scoreParam)+'='+str(round(score,2))+'\n')
continue
if modelName not in self.params:
continue
paramSpace=self.params[modelName].copy()
algoName = self.AlgorithmNames[modelName]
paramDict =paramObj.paramDefine(paramSpace,self.method)
if self.method == 'bayesopt':
code_configure.add_model(algoName,paramSpace)
else:
paramDictCopy = paramDict
# numpy array is not json serializable
#numpy is already imported but still np.ndarray raise error
import numpy as np
for key,value in paramDictCopy.items():
if isinstance(value, np.ndarray):
paramDictCopy[key] = paramDictCopy[key].tolist()
code_configure.add_model(algoName,paramDictCopy)
if not self.method == 'bayesopt':
paramSize = paramObj.getParamSpaceSize(paramDict)
else:
paramSize = 0
if (self.method == 'bayesopt' and not paramDict) or (not self.method == 'bayesopt' and paramSize<=1):
try:
start = time.time()
#function call
defObj = defaultParams(algoName,paramDict,self.scoreParam,0,0,paramSize)
estimator, modelParams, model,score =defObj.startTrainingRegression(self.trainX,self.trainY,self.testX,self.testY)
executionTime = time.time() - start
self.log.info('---------> Total Execution: ' + str(executionTime) + '\n')
if (scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"' + self.modelToAlgoNames[model] + '","FeatureEngineering":"' + str(
modelOrFeatureBased) + '","Score":' + str(score) + ',"ModelUncertainty":"NA"}'
if self.scoreParam == "r2":
if score > bestScore:
bestScore = score
bestModel = model
bestParams = modelParams
bestEstimator = estimator
else:
if abs(score) < bestScore or bestScore == -sys.float_info.max:
bestScore = abs(score)
bestModel = model
bestParams = modelParams
bestEstimator = estimator
self.log.info('Status:- |... ML Algorithm applied: ' + modelName)
self.log.info('Status:- |... Score: ' + objClf.get_print_score(self.scoreParam) + '=' + str(
round(score, 2)) + '\n')
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------' + modelName + ' Model Execution failed!!!.' + str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
self.log.info('\n < ---------- Model Execution Failed End --------->')
continue
trainingStatus = 'Success'
if self.method =='grid':
try:
self.log.info("-------> Optimization Method :Grid Search")
self.log.info("-------> Model Name: "+str(modelName))
opTq =OptimizationTq(algoName,paramDict,self.cvSplit,self.scoreParam,self.numIter,self.trainX,self.trainY)
start = time.time()
model,modelParams,score,estimator=opTq.gridSearchOpt()
executionTime=time.time() - start
if not self.testX.empty:
predictedData = estimator.predict(self.testX)
if 'neg_mean_squared_error' in self.scoreParam:
meanssquatederror = mean_squared_error(self.testY,predictedData)
score = meanssquatederror
elif 'neg_root_mean_squared_error' in self.scoreParam:
rootmeanssquatederror=mean_squared_error(self.testY,predictedData,squared=False)
score = rootmeanssquatederror
elif 'mae' in self.scoreParam:
meanabsoluteerror=mean_absolute_error(self.testY,predictedData)
score = meanabsoluteerror
elif 'r2' in self.scoreParam:
r2score=r2_score(self.testY,predictedData)
score = r2score
problemName = estimator.__class__.__name__
runName = algoName + '_' + modelOrFeatureBased
metrices = {}
metrices["score"] = score
try:
self.logMlflow(runName, modelParams, metrices, estimator,algoName)
except Exception as e:
self.log.info('\n-----> ML flow error!!!.' + str(e))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
# raise
pass
uq_jsonobject = ''
try:
if model in ['XGBRegressor','LGBMRegressor','CatBoostRegressor']:
self.log.info('-----> Model Uncertainty Not Supported')
else:
problemName = estimator.__class__.__name__
uqObj=aionUQ(None,None,None,problemName,modelParams,estimator,None,None,self.deployLocation)
total_picp_percentage,total_Uncertainty_percentage,uq_medium,uq_best,scoringCriteria,uq_jsonobject=uqObj.uqMain_BBMRegression(self.uq_x_train,self.uq_x_test,self.uq_y_train,self.uq_y_test,"aionuq")
self.log.info("-------> model_confidence: "+str(total_picp_percentage)+str('%'))
self.log.info("-------> model_uncertainty: "+str(total_Uncertainty_percentage)+str('%'))
except:
pass
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(abs(score))+',"ModelUncertainty":'+str(json.dumps(uq_jsonobject))+'}'
self.log.info('---------> Total Execution: '+str(executionTime)+'\n')
if self.scoreParam == "r2":
if score > bestScore:
bestScore =score
bestModel =model
bestParams=modelParams
bestEstimator=estimator
else:
if abs(score) < bestScore or bestScore == -sys.float_info.max:
bestScore =abs(score)
bestModel =model
bestParams=modelParams
bestEstimator=estimator
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------'+ modelName+' Model Execution failed!!!.'+str(inst))
self.log.info('\n < ---------- Model Execution Failed End --------->')
trainingStatus = 'Error (Exception)'
elif self.method == 'random':
try:
self.log.info("-------> Optimization Method :Random Search")
self.log.info("-------> Model Name: "+str(modelName))
opTq =OptimizationTq(algoName,paramDict,self.cvSplit,self.scoreParam,self.numIter,self.trainX,self.trainY)
start = time.time()
model,modelParams,score,estimator=opTq.randomSearchOpt()
executionTime=time.time() - start
if not self.testX.empty:
predictedData = estimator.predict(self.testX)
if 'neg_mean_squared_error' in self.scoreParam:
meanssquatederror = mean_squared_error(self.testY,predictedData)
score = meanssquatederror
elif 'neg_root_mean_squared_error' in self.scoreParam:
rootmeanssquatederror=mean_squared_error(self.testY,predictedData,squared=False)
score = rootmeanssquatederror
elif 'mae' in self.scoreParam:
meanabsoluteerror=mean_absolute_error(self.testY,predictedData)
score = meanabsoluteerror
elif 'r2' in self.scoreParam:
r2score=r2_score(self.testY,predictedData)
score = r2score
if self.scoreParam == "r2":
if score>bestScore:
bestScore =score
bestModel =model
bestParams=modelParams
bestEstimator=estimator
else:
if abs(score) < bestScore or bestScore == -sys.float_info.max:
bestScore =abs(score)
bestModel =model
bestParams=modelParams
bestEstimator=estimator
problemName = estimator.__class__.__name__
runName = algoName + '_' + modelOrFeatureBased
metrices = {}
metrices["score"] = score
try:
self.logMlflow(runName, modelParams, metrices, estimator,algoName)
except Exception as e:
self.log.info('\n-----> ML flow error!!!.' + str(e))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
# raise
pass
uq_jsonobject = ''
try:
if model in ['XGBRegressor','LGBMRegressor','CatBoostRegressor']:
self.log.info('-----> Model Uncertainty Not Supported')
else:
problemName = estimator.__class__.__name__
uqObj=aionUQ(None,None,None,problemName,modelParams,estimator,None,None,self.deployLocation)
total_picp_percentage,total_Uncertainty_percentage,uq_medium,uq_best,scoringCriteria,uq_jsonobject=uqObj.uqMain_BBMRegression(self.uq_x_train,self.uq_x_test,self.uq_y_train,self.uq_y_test,"aionuq")
self.log.info("-------> model_confidence: "+str(total_picp_percentage)+str('%'))
self.log.info("-------> model_uncertainty: "+str(total_Uncertainty_percentage)+str('%'))
except Exception as e:
print(e)
pass
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(abs(score))+',"ModelUncertainty":'+str(json.dumps(uq_jsonobject))+'}'
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------'+ modelName+' Model Execution failed!!!.'+str(inst))
self.log.info('\n < ---------- Model Execution Failed End --------->')
trainingStatus = 'Error (Exception)'
elif self.method == 'bayesopt':
try:
self.log.info("-------> Optimization Method :BayesOpt Search")
self.log.info("-------> Model Name: "+str(modelName))
opTq =OptimizationTq(algoName,paramDict,self.cvSplit,self.scoreParam,self.numIter,self.trainX,self.trainY)
fun=opTq.f
trials = Trials()
start = time.time()
best = fmin(fun,paramDict,algo=tpe.suggest, max_evals=100, trials=trials)
executionTime=time.time() - start
results = sorted(trials.results, key = lambda x: x['loss'])
bestresult=results[0]
model=bestresult['model']
score=bestresult['score']
modelParams=bestresult['params']
res = ', '.join("{!s}={!r}".format(key,val) for (key,val) in modelParams.items())
modelObj=eval(model+'('+res+')')
estimator = modelObj.fit(self.trainX,self.trainY)
if not self.testX.empty:
predictedData = estimator.predict(self.testX)
if 'neg_mean_squared_error' in self.scoreParam:
meanssquatederror = mean_squared_error(self.testY,predictedData)
score = meanssquatederror
elif 'neg_root_mean_squared_error' in self.scoreParam:
rootmeanssquatederror=mean_squared_error(self.testY,predictedData,squared=False)
score = rootmeanssquatederror
elif 'mae' in self.scoreParam:
meanabsoluteerror=mean_absolute_error(self.testY,predictedData)
score = meanabsoluteerror
elif 'r2' in self.scoreParam:
r2score=r2_score(self.testY,predictedData)
score = r2score
problemName = estimator.__class__.__name__
runName = algoName + '_' + modelOrFeatureBased
metrices = {}
metrices["score"] = score
try:
self.logMlflow(runName, modelParams, metrices, estimator,algoName)
except Exception as e:
self.log.info('\n-----> ML flow error!!!.' + str(e))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
# raise
pass
if self.scoreParam == "r2":
if score>bestScore:
bestScore =score
bestModel =model
bestParams=modelParams
bestEstimator=estimator
else:
if abs(score) < bestScore or bestScore == -sys.float_info.max:
bestScore =abs(score)
bestModel =model
bestParams=modelParams
bestEstimator=estimator
uq_jsonobject = ''
try:
if model in ['XGBRegressor','LGBMRegressor','CatBoostRegressor']:
self.log.info('-----> Model Uncertainty Not Supported')
else:
problemName = estimator.__class__.__name__
uqObj=aionUQ(None,None,None,problemName,modelParams,estimator,None,None,self.deployLocation)
total_picp_percentage,total_Uncertainty_percentage,uq_medium,uq_best,scoringCriteria,uq_jsonobject=uqObj.uqMain_BBMRegression(self.uq_x_train,self.uq_x_test,self.uq_y_train,self.uq_y_test,"aionuq")
self.log.info("-------> model_confidence: "+str(total_picp_percentage)+str('%'))
self.log.info("-------> model_uncertainty: "+str(total_Uncertainty_percentage)+str('%'))
except:
pass
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(abs(score))+',"ModelUncertainty":'+str(json.dumps(uq_jsonobject))+'}'
self.log.info('---------> Total Execution: '+str(executionTime)+'\n')
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------'+ modelName+' Model Execution failed!!!.'+str(inst))
self.log.info('\n < ---------- Model Execution Failed End --------->')
trainingStatus = 'Error (Exception)'
else:
trainingStatus = 'Error (HyperTunning Algo Not Supported)'
pass
self.log.info('Status:- |... ML Algorithm applied: '+modelName)
if trainingStatus.lower() == 'success':
self.log.info('Status:- |... Score after hyperparameter tuning: '+objClf.get_print_score(self.scoreParam)+'='+str(round(score,2))+'\n')
else:
self.log.info('Status:- |... Training Error : '+trainingStatus+'\n')
if bestModel != 'None':
self.log.info('---------- Regression Model End ---------- \n')
self.log.info('\n------- Best Model and its parameters -------------')
self.log.info('Status:- |... Best Algorithm selected: '+str(self.modelToAlgoNames[bestModel])+' Score='+str(round(bestScore,2)))
self.log.info("-------> Best Name: "+str(bestModel))
self.log.info("-------> Best Score: "+str(bestScore))
else:
raise Exception("Sorry, no model is trained")
return self.modelToAlgoNames[bestModel],bestParams,bestScore,bestEstimator,scoredetails
except Exception as inst:
self.log.info( '\n-----> regressionModel failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno)) |
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
defaultAlgos.py | import numpy as np
# from learner.classificationModel import ClassifierModel
from learner.aion_matrix import aion_matrix
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error,make_scorer
from sklearn.metrics import mean_squared_error
class defaultParams():
def __init__(self, modelName, paramDict, scoreParam, MakeFP0, MakeFN0,paramSize):
self.modelName = modelName
self.paramDict = paramDict
self.scoreParam = scoreParam
self.MakeFP0 = MakeFP0
self.MakeFN0 = MakeFN0
self.dictsize = paramSize
def paramDictConvertion(self):
if self.dictsize != 0:
for keys in self.paramDict.keys():
self.paramDict[keys] = self.paramDict[keys][0]
def startTrainingClassification(self, trainX, trainY, testX, testY):
threshold = -1
precisionscore = -1
recallscore = -1
objClf = aion_matrix()
self.paramDictConvertion()
if self.modelName == 'LogisticRegression':
from sklearn import linear_model
estimator = linear_model.LogisticRegression()
if self.modelName == 'GaussianNB':
from sklearn.naive_bayes import GaussianNB
estimator = GaussianNB()
if self.modelName == 'SVC':
from sklearn import svm
estimator = svm.SVC()
if self.modelName == 'KNeighborsClassifier':
from sklearn.neighbors import KNeighborsClassifier
estimator = KNeighborsClassifier()
if self.modelName == 'DecisionTreeClassifier':
from sklearn.tree import DecisionTreeClassifier
estimator = DecisionTreeClassifier()
if self.modelName == 'RandomForestClassifier':
from sklearn.ensemble import RandomForestClassifier
estimator = RandomForestClassifier()
if self.modelName == 'GradientBoostingClassifier':
from sklearn.ensemble import GradientBoostingClassifier
estimator = GradientBoostingClassifier()
if self.modelName == 'XGBClassifier':
import xgboost as xgb
estimator = xgb.XGBClassifier()
if self.modelName == 'CatBoostClassifier':
from catboost import CatBoostClassifier
estimator = CatBoostClassifier()
if self.modelName == 'LGBMClassifier':
from lightgbm import LGBMClassifier
estimator = LGBMClassifier()
if self.dictsize != 0:
estimator.set_params(**self.paramDict)
estimator.fit(trainX, trainY)
if not testX.empty:
predictedData = estimator.predict(testX)
score = objClf.get_score(self.scoreParam, testY, predictedData)
if self.MakeFP0:
self.log.info('-------- Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange, endRange, stepsize)
threshold, precisionscore, recallscore = objClf.check_threshold(estimator, trainX, trainY,
threshold_range, 'FP', self.modelName)
self.log.info('-------- Calculate Threshold for FP End-------')
if self.MakeFN0:
self.log.info('-------- Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange, endRange, stepsize)
threshold, precisionscore, recallscore = objClf.check_threshold(estimator, trainX, trainY,
threshold_range, 'FN', self.modelName)
self.log.info('-------- Calculate Threshold for FN End-------')
else:
predictedData = estimator.predict(trainX)
score = objClf.get_score(self.scoreParam, trainY, predictedData)
if self.MakeFP0:
self.log.info('-------- Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange, endRange, stepsize)
threshold, precisionscore, recallscore = objClf.check_threshold(estimator, trainX, trainY,
threshold_range, 'FP', self.modelName)
self.log.info('-------- Calculate Threshold for FP End-------')
if self.MakeFN0:
self.log.info('-------- Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange, endRange, stepsize)
threshold, precisionscore, recallscore = objClf.check_threshold(estimator, trainX, trainY,
threshold_range, 'FN', self.modelName)
self.log.info('-------- Calculate Threshold for FN End-------')
# status, bscore, bthres, brscore, bpscore = objClf.getBestModel(self.MakeFP0,self.MakeFN0, threshold,
# bestthreshold, recallscore, bestrecallscore,
# precisionscore, bestprecisionscore, score,
# bestScore)
return estimator, estimator.get_params(), self.modelName, score, threshold, precisionscore, recallscore
def startTrainingRegression(self, trainX, trainY, testX, testY):
#objClf = aion_matrix()
try:
score = 0
self.paramDictConvertion()
if self.modelName=="LinearRegression":
from sklearn import linear_model
estimator = linear_model.LinearRegression()
if self.modelName=="Lasso":
from sklearn import linear_model
estimator = linear_model.Lasso()
if self.modelName=="Ridge":
from sklearn import linear_model
estimator = linear_model.Ridge()
if self.modelName=="DecisionTreeRegressor":
from sklearn.tree import DecisionTreeRegressor
estimator = DecisionTreeRegressor()
if self.modelName=="RandomForestRegressor":
from sklearn.ensemble import RandomForestRegressor
estimator = RandomForestRegressor()
if self.modelName== "XGBRegressor":
import xgboost as xgb
estimator = xgb.XGBRegressor()
if self.modelName == 'CatBoostRegressor':
from catboost import CatBoostRegressor
estimator = CatBoostRegressor()
if self.modelName == 'LGBMRegressor':
from lightgbm import LGBMRegressor
estimator = LGBMRegressor()
if self.dictsize != 0:
estimator.set_params(**self.paramDict)
estimator.fit(trainX, trainY)
except Exception as e:
print(e)
if not testX.empty:
predictedData = estimator.predict(testX)
if 'neg_mean_squared_error' in self.scoreParam:
meanssquatederror = mean_squared_error(testY, predictedData)
score = meanssquatederror
elif 'neg_root_mean_squared_error' in self.scoreParam:
rootmeanssquatederror = mean_squared_error(testY, predictedData, squared=False)
score = rootmeanssquatederror
elif 'mae' in self.scoreParam:
meanabsoluteerror = mean_absolute_error(testY, predictedData)
score = meanabsoluteerror
elif 'r2' in self.scoreParam:
r2score = r2_score(testY, predictedData)
score = r2score
else:
predictedData = estimator.predict(trainX)
if 'neg_mean_squared_error' in self.scoreParam:
meanssquatederror = mean_squared_error(trainY, predictedData)
score = meanssquatederror
elif 'neg_root_mean_squared_error' in self.scoreParam:
rootmeanssquatederror = mean_squared_error(trainY, predictedData, squared=False)
score = rootmeanssquatederror
elif 'mae' in self.scoreParam:
meanabsoluteerror = mean_absolute_error(trainY, predictedData)
score = meanabsoluteerror
elif 'r2' in self.scoreParam:
r2score = r2_score(trainY, predictedData)
score = r2score
return estimator, estimator.get_params(), self.modelName, score
|
objectDetector.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import os
import sys
import json
import shutil
import tarfile
import logging
import subprocess
from os.path import expanduser
import platform
from pathlib import Path, PurePosixPath
import tensorflow.compat.v2 as tf
os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2' # Suppress TensorFlow logging (1)
from google.protobuf import text_format
'''
from object_detection import model_lib_v2
from object_detection import model_main_tf2
from object_detection import exporter_lib_v2
from object_detection.utils import config_util
from object_detection.protos import pipeline_pb2
'''
from learner.cloudServer import awsGPUTraining
class objectDetector(object):
def __init__(self, dataLocation, pretrainedModels, modelDirName,remoteTrainingConfig):
self.log = logging.getLogger('eion')
self.dataLocation = dataLocation
self.pretrainedModels = Path(pretrainedModels)
self.modelDirName = Path(modelDirName['file'])
self.modelURLDict = modelDirName
self.gpu = remoteTrainingConfig['Enable']
self.serverConfig = remoteTrainingConfig
self.modelOutput = Path(dataLocation).parent/"export"
if remoteTrainingConfig['Enable']:
'''
if not Path(serverConfigFile).is_file():
raise ValueError("Gpu training is enabled but server config file is not present.")
with open(serverConfigFile) as fObj:
self.serverConfig = json.load(fObj)
'''
self.tfRecordLoc = PurePosixPath('aion/data/od')
self.pipelineLoc = PurePosixPath('aion/data/od')
self.labelMapLoc = PurePosixPath('aion/data/od')
self.gpuPretrainedModelPath = PurePosixPath('aion/pretrainedModels')/self.modelDirName
else:
self.tfRecordLoc = Path(dataLocation)
self.pipelineLoc = Path(dataLocation)
self.labelMapLoc = Path(dataLocation)
self.gpuPretrainedModelPath = None
def prepareConfig(self, detectionModel, num_classes, n_epoch, batch_size):
home = expanduser("~")
if platform.system() == 'Windows':
modelPath = os.path.join(home,'AppData','Local','HCLT','AION','PreTrainedModels','ObjectDetection')
else:
modelPath = os.path.join(home,'HCLT','AION','PreTrainedModels','ObjectDetection')
pipeline_config = str(modelPath/self.modelDirName/"pipeline.config")
checkPoint = "ckpt-0"
with open(str(modelPath/self.modelDirName/"checkpoint/checkpoint")) as f:
line = f.readline()
checkPoint = line.split(':')[1].strip()[1:-1] #(model_checkpoint_path: "ckpt-301") to ckpt-301
checkPoint = "checkpoint/"+checkPoint
from object_detection.utils import config_util
configs = config_util.get_configs_from_pipeline_file(pipeline_config)
model_config = configs['model']
if detectionModel.lower() == 'ssd':
model_config.ssd.num_classes = num_classes
configs['train_config'].fine_tune_checkpoint_type = "detection"
elif detectionModel.lower() == 'centernet':
model_config.center_net.num_classes = num_classes
configs['train_config'].fine_tune_checkpoint_type = "fine_tune"
elif detectionModel.lower() == 'fasterrcnn':
model_config.faster_rcnn.num_classes = num_classes
configs['train_config'].fine_tune_checkpoint_type = "detection"
else:
raise ValueError("{} Model is not supported for object detection.\n".format(detectionModel))
if self.gpu:
checkpointPath = str(self.gpuPretrainedModelPath / checkPoint)
else:
checkpointPath = str(modelPath/self.modelDirName/checkPoint)
configs['train_config'].fine_tune_checkpoint = checkpointPath
configs['train_config'].num_steps = n_epoch
configs['train_config'].batch_size = batch_size
configs['train_input_config'].tf_record_input_reader.input_path[:] = [str(self.tfRecordLoc/"train.tfrecord")]
configs['train_input_config'].label_map_path = str(self.labelMapLoc/"label_map.pbtxt")
configs['eval_input_config'].tf_record_input_reader.input_path[:] = [self.dataLocation + "/test.tfrecord"]
configs['eval_input_config'].label_map_path = self.dataLocation + "/label_map.pbtxt"
# Save new pipeline config
new_pipeline_proto = config_util.create_pipeline_proto_from_configs(configs)
config_util.save_pipeline_config(new_pipeline_proto, self.dataLocation)
def __exportModel(self):
self.log.info('-------> exporting trained Model')
from object_detection.protos import pipeline_pb2
from object_detection import exporter_lib_v2
pipeline_config = pipeline_pb2.TrainEvalPipelineConfig()
with tf.io.gfile.GFile(str(self.pipelineLoc/"pipeline.config"), 'r') as f:
text_format.Merge(f.read(), pipeline_config)
text_format.Merge('', pipeline_config)
exporter_lib_v2.export_inference_graph(
'image_tensor', pipeline_config, self.dataLocation,
str(self.modelOutput))
def startObjectDetector(self):
if self.gpu:
self.log.info('-------> Training on the cloud machine')
self.log.info('Status:- |...Remote Machine Training')
with open(self.dataLocation+'\model.config', 'w')as f:
json.dump( self.modelURLDict, f)
awsGpu = awsGPUTraining(self.serverConfig)
try:
awsGpu.start_instance()
awsGpu.copy_files_to_server(self.dataLocation)
awsGpu.start_executing_notebook()
self.log.info('-------> Downloading trained model file')
tarFile = awsGpu.copy_file_from_server(self.dataLocation)
with tarfile.open(tarFile) as tar:
tar.extractall(self.dataLocation)
awsGpu.stop_server_instance()
except:
awsGpu.stop_server_instance()
raise
extractedPath = Path(self.dataLocation)/Path(tarFile).name.split('.')[0]
filesList = extractedPath.glob('**/*')
for file in filesList:
if file.parent == extractedPath:
if file.name == "export":
shutil.copytree(file, self.modelOutput)
elif file.is_dir():
shutil.copytree(file, Path(self.dataLocation)/file.name)
else:
shutil.copy2(file, self.dataLocation)
shutil.rmtree(extractedPath)
Path(tarFile).unlink()
shutil.copy2(self.dataLocation + "/label_map.pbtxt", str(self.modelOutput))
else:
self.log.info('-------> Training on the local machine')
self.log.info('Status:- |...Local Machine Training')
tf.config.set_soft_device_placement(True)
strategy = tf.compat.v2.distribute.MirroredStrategy()
with strategy.scope():
try:
from object_detection import model_lib_v2
model_lib_v2.train_loop(
pipeline_config_path=str(self.pipelineLoc/"pipeline.config"),
model_dir=str(self.dataLocation))
except Exception:
raise
self.__exportModel()
shutil.copy2(str(self.labelMapLoc/"label_map.pbtxt"), str(self.modelOutput))
def evaluateObjectDetector(self, model_dir, pipeline_config_dir=None, checkpoint_dir=None):
if checkpoint_dir == None:
checkpoint_dir = model_dir
if pipeline_config_dir == None:
pipeline_config_dir = model_dir
self.log.info('-------> Evaluation started')
from object_detection import model_main_tf2
cmd = '"{}" "{}" --model_dir="{}" --pipeline_config_path="{}/pipeline.config" --checkpoint_dir="{}" --eval_timeout=6'.format(sys.executable, model_main_tf2.__file__, model_dir, model_dir, checkpoint_dir)
result = subprocess.run(cmd , capture_output=True, text=True,shell=True)
precisionParam = ['Average Precision', 'Average Recall']
text = result.stdout.split('\n')
stats = {}
keys = []
try:
for x in text:
for y in precisionParam:
indx = x.find(y)
if indx != -1:
keyValue = x[indx:].split(' = ')
stats[keyValue[0]] = keyValue[1]
keys.append(keyValue[0])
except Exception as e:
raise ValueError("Error in evaluation: " + str(e))
self.log.info('-------> Evaluation statistics:')
self.log.info(stats)
return stats, keys
|
optimizetechnique.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
from sklearn.model_selection import GridSearchCV
from sklearn.model_selection import RandomizedSearchCV
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.linear_model import SGDClassifier, PassiveAggressiveClassifier
from sklearn.linear_model import SGDRegressor, PassiveAggressiveRegressor
from sklearn.linear_model import SGDClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.model_selection import cross_val_score
from sklearn.svm import SVC
from hyperopt import fmin, tpe, hp, STATUS_OK, Trials
from sklearn.svm import SVR
import xgboost as xgb
from xgboost import XGBClassifier
from lightgbm import LGBMClassifier
from catboost import CatBoostClassifier
from xgboost import XGBRegressor
from lightgbm import LGBMRegressor
from catboost import CatBoostRegressor
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Lasso
from sklearn.linear_model import Ridge
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import RandomForestRegressor
import warnings
warnings.filterwarnings('ignore')
import time
import logging
import sys,os
class StreamToLogger(object):
def __init__(self, logger, log_level=logging.INFO):
self.logger = logger
self.log_level = log_level
self.linebuf = ''
def write(self, buf):
for line in buf.rstrip().splitlines():
self.logger.log(self.log_level, 'Model:- Iteration:: '+line.rstrip())
class OptimizationTq():
def __init__(self,modelName,tuneParams,cvSplit,scoreParam,nIter,trainX,trainY,geneticParam=None):
self.data = None
self.model=modelName
self.params =tuneParams
self.cvSplit=cvSplit
self.scoreParam=scoreParam
self.trainX =trainX
self.trainY = trainY
self.geneticParam=geneticParam if geneticParam else {}
self.nIter =nIter
self.count =0
self.best =0
self.log = logging.getLogger('eion')
def gridSearchOpt(self):
try:
sl = StreamToLogger(self.log, logging.INFO)
oldStdout = sys.stdout
sys.stdout = sl
self.log.info('Model:-Model Name:: '+str(self.model))
modelObj=eval(self.model+'()')
gridOp = GridSearchCV(modelObj, param_grid=self.params,scoring=self.scoreParam, cv=self.cvSplit,verbose=10)
gridFit=gridOp.fit(self.trainX,self.trainY)
self.log.info('Model:-Model Name:: '+str(self.model))
self.log.info('Model:-ScoringType:: '+str(gridFit.scorer_))
self.log.info('Model:-Best Param:: '+str(gridFit.best_params_))
self.log.info('Model:-Validation Score:: '+str(gridFit.best_score_))
self.log.info('Model:-CV Result:: '+str(gridFit.cv_results_))
self.log.info('Model:-Best Estimator:: '+str(gridFit.best_estimator_))
sys.stdout = oldStdout
return self.model,gridFit.best_params_,gridFit.best_score_,gridFit.best_estimator_
except Exception as inst:
self.log.info("gridSearchOpt failed ==>"+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def randomSearchOpt(self):
try:
sl = StreamToLogger(self.log, logging.INFO)
oldStdout = sys.stdout
sys.stdout = sl
self.log.info('Model:-Model Name:: '+str(self.model))
modelObj=eval(self.model+'()')
randomOp = RandomizedSearchCV(modelObj, param_distributions=self.params,scoring=self.scoreParam,n_iter=self.nIter,cv=self.cvSplit,verbose=10)
randomFit=randomOp.fit(self.trainX,self.trainY)
self.log.info('Model:-Model Name:: '+str(self.model))
self.log.info('Model:-ScoringType:: '+str(randomFit.scorer_))
self.log.info('Model:-Best Param:: '+str(randomFit.best_params_))
self.log.info('Model:-Validation Score:: '+str(randomFit.best_score_))
self.log.info('Model:-CV Result:: '+str(randomFit.cv_results_))
self.log.info('Model:-Best Estimator:: '+str(randomFit.best_estimator_))
sys.stdout = oldStdout
return self.model,randomFit.best_params_,randomFit.best_score_,randomFit.best_estimator_
except Exception as inst:
self.log.info("RandomsearchOptimization failed ==>"+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def bayesianOpt(self,params):
modelObj=eval(self.model+'(**'+str(params)+')')
score=cross_val_score(modelObj, self.trainX, self.trainY,scoring=self.scoreParam,cv=self.cvSplit)
return score.mean()
def f(self,params):
best=self.best
count=self.count
parameters=params
count += 1
classObj=OptimizationTq(self.model,self.params,self.cvSplit,self.scoreParam,self.nIter,self.trainX,self.trainY)
acc = classObj.bayesianOpt(parameters.copy())
return {'loss':-acc,'score': acc, 'status': STATUS_OK,'model' :self.model,'params': params}
|
cloudServer.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import boto3
import json
import time
import requests
import datetime
import uuid
import shutil
from websocket import create_connection
from botocore.exceptions import ClientError
import tarfile
from pathlib import Path, PurePosixPath
from stat import S_ISDIR
from fabric import Connection
import time
import logging
class awsGPUTraining():
def __init__(self, config):
local_config = {"location":{"data":"aion/data/od", "code":"", "pretrainedModel":"aion/pretrainedModels"},
"jupyter":{"header":{"Authorization":"Token f3af05d5348301997fb014f245569e872d27bb9018fd70d2"}, "portNo":"8888",
"notebook_path":"aion/code/AWS_GPU_OD_Training.ipynb"}}
self.serverConfig = config["server"]
self.sshConfig = config["ssh"]
self.log = logging.getLogger('eion')
self.codeLocation = local_config["location"]["code"]
self.dataLocation = local_config["location"]["data"]
self.pretrainedModelLocation = local_config["location"]["pretrainedModel"]
self.jupyterConfig = local_config["jupyter"]
self.serverIP = ""
if self.serverConfig["awsAccessKeyId"] == "" or self.serverConfig["awsSecretAccessKey"] == "":
raise ValueError("Cloud server configuration is not available.")
if len(self.serverConfig["InstanceIds"]) == 0 and self.serverConfig["amiId"] == "":
raise ValueError("Please provide either InstanceIds or amiId in server config")
self.instanceId = []
self.separate_instance = False
if self.serverConfig["amiId"] != "":
self.separate_instance = True
else:
if len(self.serverConfig["InstanceIds"]):
if isinstance(self.serverConfig["InstanceIds"], list):
self.instanceId = self.serverConfig["InstanceIds"]
elif isinstance(self.serverConfig["InstanceIds"], str):
self.instanceId = [self.serverConfig["InstanceIds"]]
self.ec2_client = boto3.client(self.serverConfig["serverName"], region_name=self.serverConfig["regionName"], aws_access_key_id=self.serverConfig["awsAccessKeyId"], aws_secret_access_key=self.serverConfig["awsSecretAccessKey"])
def __sftp_exists(self, sftp, path):
try:
sftp.stat(path)
return True
except:# IOError, e:
#if e.errno == errno.ENOENT:
return False
def __rmtree(self, sftp, remotepath, level=0):
for f in sftp.listdir_attr(remotepath):
rpath = str(PurePosixPath(remotepath)/f.filename)
if S_ISDIR(f.st_mode):
self.__rmtree(sftp, rpath, level=(level + 1))
sftp.rmdir(rpath)
else:
rpath = str(PurePosixPath(remotepath)/f.filename)
sftp.remove(rpath)
def copy_files_to_server(self, location):
try:
client = Connection(
host=self.serverIP,
user=self.sshConfig["userName"],
connect_kwargs={
"key_filename": self.sshConfig["keyFilePath"],
},
)
client.sudo('rm -rf {}/*'.format(self.dataLocation))
tarFile = str((PurePosixPath(self.dataLocation).parent/PurePosixPath(self.dataLocation).name).with_suffix(".tar.gz"))
client.put(location+'/test.tfrecord', self.dataLocation+'/test.tfrecord')
client.put(location+'/train.tfrecord', self.dataLocation+'/train.tfrecord')
client.put(location+'/pipeline.config', self.dataLocation+'/pipeline.config')
client.put(location+'/label_map.pbtxt', self.dataLocation+'/label_map.pbtxt')
client.put(location+'/model.config', self.dataLocation+'/model.config')
if self.jupyterConfig != "":
client.run("touch {}".format(self.dataLocation+'/log.txt'))
except Exception as e:
raise ValueError("Error in copying data to cloud server. " + str(e))
def __myexec(self, ssh, cmd, timeout, want_exitcode=False):
# one channel per command
stdin, stdout, stderr = ssh.exec_command(cmd)
# get the shared channel for stdout/stderr/stdin
channel = stdout.channel
# we do not need stdin.
stdin.close()
# indicate that we're not going to write to that channel anymore
channel.shutdown_write()
# read stdout/stderr in order to prevent read block hangs
stdout_chunks = []
stdout_chunks.append(stdout.channel.recv(len(stdout.channel.in_buffer)))
# chunked read to prevent stalls
while not channel.closed or channel.recv_ready() or channel.recv_stderr_ready():
# stop if channel was closed prematurely, and there is no data in the buffers.
got_chunk = False
readq, _, _ = select.select([stdout.channel], [], [], timeout)
for c in readq:
if c.recv_ready():
stdout_chunks.append(stdout.channel.recv(len(c.in_buffer)))
got_chunk = True
if c.recv_stderr_ready():
# make sure to read stderr to prevent stall
stderr.channel.recv_stderr(len(c.in_stderr_buffer))
got_chunk = True
'''
1) make sure that there are at least 2 cycles with no data in the input buffers in order to not exit too early (i.e. cat on a >200k file).
2) if no data arrived in the last loop, check if we already received the exit code
3) check if input buffers are empty
4) exit the loop
'''
if not got_chunk \
and stdout.channel.exit_status_ready() \
and not stderr.channel.recv_stderr_ready() \
and not stdout.channel.recv_ready():
# indicate that we're not going to read from this channel anymore
stdout.channel.shutdown_read()
# close the channel
stdout.channel.close()
break # exit as remote side is finished and our bufferes are empty
# close all the pseudofiles
stdout.close()
stderr.close()
if want_exitcode:
# exit code is always ready at this point
return (''.join(stdout_chunks), stdout.channel.recv_exit_status())
return ''.join(stdout_chunks)
def __myexec1(self, ssh, cmd, timeout, want_exitcode=False):
# one channel per command
stdin, stdout, stderr = ssh.exec_command(cmd, get_pty=True)
for line in iter(stderr.readline, ""):
print(line, end="")
stdin.close()
stdout.close()
stderr.close()
def executeCode(self):
try:
client = Connection(
host=self.serverIP,
user=self.sshConfig["userName"],
connect_kwargs={
"key_filename": self.sshConfig["keyFilePath"],
},
)
cmd = 'python3.8 {} {} {}'.format(self.codeLocation, self.dataLocation, self.pretrainedModelLocation)
client.run( cmd)
except Exception as e:
raise ValueError("Error in running code on cloud server. " + str(e))
def start_executing_notebook(self):
try:
publicIp_Port = self.serverIP + ":" + self.jupyterConfig["portNo"]
conURL = "ws://" + publicIp_Port
base = 'http://' + publicIp_Port + ''
headers = self.jupyterConfig["header"]
url = base + '/api/kernels'
flag = True
while flag: # deadlock need to add timeout
response = requests.post(url, headers=headers)
flag = False
kernel = json.loads(response.text)
# Load the notebook and get the code of each cell
url = base + '/api/contents/' + self.jupyterConfig["notebook_path"]
response = requests.get(url, headers=headers)
file = json.loads(response.text)
code = [c['source'] for c in file['content']['cells'] if len(c['source']) > 0 and c['cell_type']=='code' ]
ws = create_connection(conURL + "/api/kernels/" + kernel["id"] + "/channels",
header=headers)
def send_execute_request(code):
msg_type = 'execute_request';
content = {'code': code, 'silent': False}
hdr = {'msg_id': uuid.uuid1().hex,
'username': 'test',
'session': uuid.uuid1().hex,
'data': datetime.datetime.now().isoformat(),
'msg_type': msg_type,
'version': '5.0'}
msg = {'header': hdr, 'parent_header': hdr,
'metadata': {},
'content': content}
return msg
for c in code:
ws.send(json.dumps(send_execute_request(c)))
# We ignore all the other messages, we just get the code execution output
# (this needs to be improved for production to take into account errors, large cell output, images, etc.)
error_msg = ''
traceback_msg = ''
for i in range(0, len(code)):
msg_type = '';
while msg_type != "stream":
rsp = json.loads(ws.recv())
msg_type = rsp["msg_type"]
if msg_type == 'error':
raise ValueError("Error on Cloud machine: "+rsp['content']['evalue'])
ws.close()
self.log.info('Status:- |...Execution Started`')
except ClientError as e:
raise ValueError(e)
def __wait_for_completion(self, sftp, remoteLogFile, localLogFile):
waiting = True
error_msg = ""
while waiting:
time.sleep(5 * 60)
try:
sftp.get(str(remoteLogFile), str(localLogFile))
with open(localLogFile, "r") as f:
content = f.readlines()
for x in content:
if "Error" in x:
waiting = False
error_msg = x
if "success" in x:
waiting = False
except:
raise (str(e))
return error_msg
def copy_file_from_server(self, localPath):
try:
client = Connection(
host=self.serverIP,
user=self.sshConfig["userName"],
connect_kwargs={
"key_filename": self.sshConfig["keyFilePath"],
},
)
remoteLogFile = PurePosixPath(self.dataLocation)/'log.txt'
localLogFile = Path(localPath)/'remote_log.txt'
client.get(str(remoteLogFile), str(localLogFile))
tarFile = (PurePosixPath(self.dataLocation).parent/PurePosixPath(self.dataLocation).name).with_suffix(".tar.gz")
client.get(str(tarFile), str(Path(localPath)/tarFile.name))
except:
raise
return str(Path(localPath)/tarFile.name)
def create_instance(self):
instances = self.ec2_client.run_instances(
ImageId=self.serverConfig["amiId"],
MinCount=1,
MaxCount=1,
InstanceType="t2.xlarge",
KeyName="AION_GPU",
SecurityGroupIds = ["sg-02c3a6c8dd67edb74"]
)
self.instanceId = [instances['Instances'][0]['InstanceId']]
def start_instance(self):
if self.separate_instance:
self.create_instance()
try:
response = self.ec2_client.start_instances(InstanceIds=self.instanceId, DryRun=True)
except Exception as e:
if 'DryRunOperation' not in str(e):
raise ValueError("Error in starting the EC2 instance, check server configuration. " + str(e))
try:
running_state_code = 16
response = self.ec2_client.start_instances(InstanceIds=self.instanceId, DryRun=False)
instance_status_code = 0
while instance_status_code != running_state_code:
response = self.ec2_client.describe_instances(InstanceIds=self.instanceId)
instance_status_code = response['Reservations'][0]['Instances'][0]['State']['Code']
if instance_status_code == running_state_code:
self.serverIP = response['Reservations'][0]['Instances'][0]['PublicIpAddress']
break
except ClientError as e:
raise ValueError("Error in starting the EC2 instance. " + str(e))
def terminate_instance(self):
ec2 = boto3.resource(self.serverConfig["serverName"], region_name=self.serverConfig["regionName"], aws_access_key_id=self.serverConfig["awsAccessKeyId"], aws_secret_access_key=self.serverConfig["awsSecretAccessKey"])
ec2.instances.filter(InstanceIds=self.instanceId).terminate() # for terminating an ec2 instance
def stop_server_instance(self):
try:
self.ec2_client.stop_instances(InstanceIds=self.instanceId, DryRun=True)
except Exception as e:
if 'DryRunOperation' not in str(e):
raise
stopped_state_code = 80
# Dry run succeeded, call stop_instances without dryrun
try:
response = self.ec2_client.stop_instances(InstanceIds=self.instanceId, DryRun=False)
response = self.ec2_client.describe_instances(InstanceIds=self.instanceId)
instance_status_code = 0
while instance_status_code != stopped_state_code:
response = self.ec2_client.describe_instances(InstanceIds=self.instanceId)
instance_status_code = response['Reservations'][0]['Instances'][0]['State']['Code']
if instance_status_code == stopped_state_code:
break
except:
raise ValueError("Error in stopping the EC2 instance {}.Please stop it manually ".format(self.instanceId[0]))
if self.separate_instance:
try:
self.terminate_instance()
except:
raise ValueError("Error in terminating the EC2 instance {}.Please terminate it manually ".format(self.instanceId[0]))
|
classificationModel.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import time
import os
import sys
import numpy as np
from numpy import arange
from numpy import argmax
import json
from sklearn.metrics import accuracy_score
from sklearn.metrics import confusion_matrix
from sklearn.metrics import recall_score
from sklearn.metrics import precision_score
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.linear_model import SGDClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import roc_curve
from sklearn.metrics import roc_auc_score
from sklearn.metrics import f1_score
from sklearn.svm import SVC
from xgboost import XGBClassifier
from lightgbm import LGBMClassifier
from catboost import CatBoostClassifier
from sklearn.preprocessing import binarize
from learner.optimizetechnique import OptimizationTq
from learner.defaultAlgos import defaultParams
from learner.parameters import parametersDefine
from hyperopt import fmin, tpe, hp, STATUS_OK, Trials
import logging
from learner.aion_matrix import aion_matrix
import mlflow
from pathlib import Path
from uncertainties.aionUQ import aionUQ
# apply threshold to positive probabilities to create labels
def to_labels(pos_probs, threshold):
return (pos_probs >= threshold).astype('int')
class ClassifierModel():
def __init__(self,modelList,params,scoreParam,cvSplit,numIter,geneticParam,trainX,trainY,testX,testY,method,modelType,MakeFP0,MakeFN0,deployLocation):
self.modelList =modelList
self.params =params
self.trainX =trainX
self.X =trainX
self.trainY =trainY
self.testX = testX
self.testY = testY
self.method =method
self.scoreParam=scoreParam
self.cvSplit=cvSplit
self.numIter=numIter
self.geneticParam=geneticParam
self.MakeFP0= MakeFP0
self.MakeFN0=MakeFN0
self.log = logging.getLogger('eion')
self.modelType = modelType
self.uq_x_train = trainX
self.uq_x_test = testX
self.uq_y_train = trainY
self.uq_y_test = testY
self.deployLocation = deployLocation
self.AlgorithmNames={'Logistic Regression':'LogisticRegression','Stochastic Gradient Descent':'SGDClassifier','Naive Bayes':'GaussianNB','Support Vector Machine':'SVC','K Nearest Neighbors':'KNeighborsClassifier','Decision Tree':'DecisionTreeClassifier','Random Forest':'RandomForestClassifier','Gradient Boosting':'GradientBoostingClassifier','Extreme Gradient Boosting (XGBoost)':'XGBClassifier','Categorical Boosting (CatBoost)': 'CatBoostClassifier','Light Gradient Boosting (LightGBM)': 'LGBMClassifier','Bagging (Ensemble)':'BaggingClassifier','Stacking (Ensemble)':'StackingClassifier','Voting (Ensemble)':'VotingClassifier','Deep Q Network':'DQN','Dueling Deep Q Network':'DDQN','Neural Architecture Search':'NAS'}
self.modelToAlgoNames = {value: key for key, value in self.AlgorithmNames.items()}
def check_threshold(self,estimator,testX,testY,threshold_range,checkParameter,modelName):
thresholdx = -1
for threshold in threshold_range:
predictedData = estimator.predict_proba(testX)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold)
p_score = precision_score(testY, predictedData)
#self.log.info('-------------> Precision:'+str(p_score))
r_score = recall_score(testY, predictedData)
#self.log.info('-------------> Rscore:'+str(r_score))
#self.log.info(confusion_matrix(testY, predictedData))
tn, fp, fn, tp = confusion_matrix(testY, predictedData).ravel()
if(checkParameter.lower() == 'fp'):
if fp == 0:
if(p_score == 1):
thresholdx = threshold
self.log.info('---------------> Best Threshold:'+str(threshold))
self.log.info('---------------> Best Precision:'+str(p_score))
self.log.info('---------------> Best Recall:'+str(r_score))
self.log.info('---------------> TN:'+str(tn))
self.log.info('---------------> FP:'+str(fp))
self.log.info('---------------> FN:'+str(fn))
self.log.info('---------------> TP:'+str(tp))
break
if(checkParameter.lower() == 'fn'):
if fn == 0:
if(r_score == 1):
thresholdx = threshold
self.log.info('---------------> Best Threshold:'+str(threshold))
self.log.info('---------------> Best Precision:'+str(p_score))
self.log.info('---------------> Best Recall:'+str(r_score))
self.log.info('---------------> TN:'+str(tn))
self.log.info('---------------> FP:'+str(fp))
self.log.info('---------------> FN:'+str(fn))
self.log.info('---------------> TP:'+str(tp))
break
return(thresholdx,p_score,r_score)
def getBestModel(self,fp0,fn0,threshold,bestthreshold,rscore,brscore,pscore,bpscore,tscore,btscore):
cmodel = False
if(threshold != -1):
if(bestthreshold == -1):
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif fp0:
if rscore > brscore:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif rscore == brscore:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif fn0:
if pscore > bpscore:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif pscore == bpscore:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
else:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
btscore = tscore
else:
if(bestthreshold == -1):
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
btscore = tscore
return cmodel,btscore,bestthreshold,brscore,bpscore
def logMlflow(self, runName, params, metrices, estimator, algoName=None):
with mlflow.start_run(run_name = runName):
for k,v in params.items():
mlflow.log_param(k, v)
for k,v in metrices.items():
mlflow.log_metric(k, v)
if algoName == 'CatBoostClassifier':
mlflow.catboost.log_model(estimator, "model")
else:
mlflow.sklearn.log_model(estimator, "model")
model_uri = mlflow.get_artifact_uri("model")
""" for some dataset evaluate takes more than 90 min, so commenting till some solution is not found
evaluate_data = self.testX.copy()
evaluate_data['label'] = self.testY.copy()
mlflow.evaluate(model_uri, data=evaluate_data, targets='label', model_type="classifier")
del evaluate_data
"""
def classModelling(self, modelOrFeatureBased,code_configure):
paramObj=parametersDefine()
bestModel='None'
bestParams={}
bestScore=-0xFFFF
bestEstimator = 'None'
bestpipelineModel='None'
scoredetails = ''
threshold = -1
bestthreshold = -1
precisionscore =-1
bestprecisionscore=-1
recallscore = -1
bestrecallscore=-1
self.log.info('\n---------- ClassifierModel has started ----------')
objClf = aion_matrix()
try:
self.log.info('Status:- |... Search Optimization Method applied: '+self.method)
for modelName in self.modelList:
if modelName in ['Bagging (Ensemble)','Voting (Ensemble)','Stacking (Ensemble)','Dueling Deep Q Network','Deep Q Network','Neural Architecture Search']:
if modelName == 'Bagging (Ensemble)':
from ensemble.ensemble_bagging import ensemble_bagging
ensemble_bagging_obj = ensemble_bagging(self.params[modelName],self.scoreParam,self.MakeFP0,self.MakeFN0)
estimator,modelParams,score,model,threshold,precisionscore,recallscore = ensemble_bagging_obj.ensemble_bagging_classifier(self.trainX,self.trainY,self.testX,self.testY)
if modelName == 'Stacking (Ensemble)':
from ensemble.ensemble_stacking import ensemble_stacking
ensemble_stacking_obj = ensemble_stacking(self.params[modelName],self.scoreParam)
estimator,modelParams,score,model,threshold,precisionscore,recallscore = ensemble_stacking_obj.ensemble_stacking_classifier(self.trainX,self.trainY,self.testX,self.testY,self.MakeFP0,self.MakeFN0,self.modelList)
if modelName == 'Voting (Ensemble)':
from ensemble.ensemble_voting import ensemble_voting
ensemble_voting_obj = ensemble_voting("",self.scoreParam)
#bug 12437
status,estimator,modelParams,score,model,threshold,precisionscore,recallscore = ensemble_voting_obj.ensemble_voting_classifier(self.trainX,self.trainY,self.testX,self.testY,self.MakeFP0,self.MakeFN0,self.modelList)
if status != "SUCCESS": #bug 12437
continue
if modelName == 'Deep Q Network':
from reinforcement.DRL_train import ReinformentLearning
rlObj = ReinformentLearning(self.params[modelName],self.scoreParam,'Classification')
estimator,modelParams,score,model,threshold,precisionscore,recallscore = rlObj.TrainRL(self.trainX,self.trainY,self.testX,self.testY,'DQN',self.deployLocation)
if modelName == 'Dueling Deep Q Network':
from reinforcement.DRL_train import ReinformentLearning
rlObj = ReinformentLearning(self.params[modelName],self.scoreParam,'Classification')
estimator,modelParams,score,model,threshold,precisionscore,recallscore = rlObj.TrainRL(self.trainX,self.trainY,self.testX,self.testY,'DDQN',self.deployLocation)
'''
if modelName == 'Neural Architecture Search':
from nas.aionNAS import aionNAS
objNAS = aionNAS('Classification',self.params[modelName],self.trainX,self.testX,self.trainY,self.testY,self.deployLocation)
estimator,modelParams,score,model,threshold,precisionscore,recallscore=objNAS.nasMain(self.scoreParam)
'''
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(score)+',"ModelUncertainty":"NA"}'
status,bscore,bthres,brscore,bpscore = self.getBestModel(self.MakeFP0,self.MakeFN0,threshold,bestthreshold,recallscore,bestrecallscore,precisionscore,bestprecisionscore,score,bestScore)
if status:
bestScore =bscore
bestModel =model
bestParams=modelParams
bestEstimator=estimator
bestthreshold = bthres
bestrecallscore = brscore
bestprecisionscore = bpscore
self.log.info('Status:- |... ML Algorithm applied: '+modelName)
self.log.info('Status:- |... Score: '+objClf.get_print_score(self.scoreParam)+'='+str(round(score,2))+'\n')
continue
paramSpace=self.params[modelName].copy()
algoName = self.AlgorithmNames[modelName]
paramDict =paramObj.paramDefine(paramSpace,self.method)
if not self.method == 'bayesopt':
paramSize = paramObj.getParamSpaceSize(paramDict)
else:
paramSize = 0
if (self.method == 'bayesopt' and not paramDict) or (not self.method == 'bayesopt' and paramSize<=0):
try:
start = time.time()
#function call
defObj = defaultParams(algoName,paramDict,self.scoreParam,self.MakeFP0, self.MakeFN0,paramSize)
estimator, modelParams, model,score, threshold, precisionscore, recallscore =defObj.startTrainingClassification(self.trainX,self.trainY,self.testX,self.testY)
executionTime = time.time() - start
if (scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"' + self.modelToAlgoNames[model] + '","FeatureEngineering":"' + str(
modelOrFeatureBased) + '","Score":' + str(score) + ',"ModelUncertainty":"NA"}'
status, bscore, bthres, brscore, bpscore = self.getBestModel(self.MakeFP0, self.MakeFN0,threshold, bestthreshold,recallscore, bestrecallscore,precisionscore, bestprecisionscore,score, bestScore)
self.log.info('---------> Total Execution: ' + str(executionTime) + '\n')
if status:
bestScore = bscore
bestModel = model
bestParams = modelParams
bestEstimator = estimator
bestthreshold = bthres
bestrecallscore = brscore
bestprecisionscore = bpscore
self.log.info('Status:- |... ML Algorithm applied: ' + modelName)
self.log.info('Status:- |... Score: ' + objClf.get_print_score(self.scoreParam) + '=' + str(
round(score, 2)) + '\n')
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------' + modelName + ' Model Execution failed!!!.' + str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno))
self.log.info('\n < ---------- Model Execution Failed End --------->')
continue
# call algorithms with default valuepass
if self.method == 'bayesopt':
code_configure.add_model(algoName,paramSpace)
else:
paramDictCopy = paramDict
# numpy array is not json serializable
#numpy is already imported but still np.ndarray raise error
import numpy as np
for key,value in paramDictCopy.items():
if isinstance(value, np.ndarray):
paramDictCopy[key] = paramDictCopy[key].tolist()
code_configure.add_model(algoName,paramDictCopy)
trainingStatus = 'Success'
if self.method =='grid':
try:
self.log.info("-------> Optimization Method :Grid Search")
self.log.info("-------> Model Name: "+str(modelName))
opTq =OptimizationTq(algoName,paramDict,self.cvSplit,self.scoreParam,self.numIter,self.trainX,self.trainY)
start = time.time()
model,modelParams,score,estimator=opTq.gridSearchOpt()
executionTime=time.time() - start
if not self.testX.empty:
predictedData = estimator.predict(self.testX)
score = objClf.get_score(self.scoreParam,self.testY,predictedData)
else:
score = score*100
problemName = estimator.__class__.__name__
runName = algoName + '_' + modelOrFeatureBased
metrices = {}
metrices["score"] = score
try:
self.logMlflow(runName, modelParams, metrices, estimator, algoName)
except Exception as e:
self.log.info('----------> ML Flow error!!!. ' + str(e)) # usnish
pass
output_jsonobject = ""
problemName = estimator.__class__.__name__
self.log.info('----------> Testing Score: '+str(score))
try:
if ((estimator.__class__.__name__ == "ABCMeta") or (model in ['SGDClassifier','XGBClassifier','CatBoostClassifier','LGBMClassifier']) ):
self.log.info('-----> Model Uncertainty Not Supported')
else:
uqObj=aionUQ(None,None,None,problemName,modelParams,estimator,None,None,self.deployLocation)
accuracy,uq_ece,output_jsonobject,model_confidence_per,model_uncertainty_per=uqObj.uqMain_BBMClassification(self.uq_x_train,self.uq_x_test,self.uq_y_train,self.uq_y_test,"aionuq")
self.log.info("-------> model_confidence: "+str(model_confidence_per)+str('%'))
self.log.info("-------> model_uncertainty: "+str(model_uncertainty_per)+str('%'))
except:
pass
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(score)+',"ModelUncertainty":'+str(json.dumps(output_jsonobject))+'}'
self.log.info('----------> Testing Score: '+str(score))
import numpy as np
if self.MakeFP0:
self.log.info('-------- Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = self.check_threshold(estimator,self.trainX,self.trainY,threshold_range,'FP',algoName)
self.log.info('-------- Calculate Threshold for FP End-------')
if self.MakeFN0:
self.log.info('-------- Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = self.check_threshold(estimator,self.trainX,self.trainY,threshold_range,'FN',algoName)
self.log.info('-------- Calculate Threshold for FN End-------')
self.log.info('----------> Total Execution: '+str(executionTime)+'\n')
status,bscore,bthres,brscore,bpscore = self.getBestModel(self.MakeFP0,self.MakeFN0,threshold,bestthreshold,recallscore,bestrecallscore,precisionscore,bestprecisionscore,score,bestScore)
if status:
bestScore =bscore
bestModel =model
bestParams=modelParams
bestEstimator=estimator
bestthreshold = bthres
bestrecallscore = brscore
bestprecisionscore = bpscore
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------'+ modelName+' Model Execution failed!!!.'+str(inst))
self.log.info('\n < ---------- Model Execution Failed End --------->')
trainingStatus = 'Error (Exception)'
elif self.method == 'random':
try:
self.log.info("-------> Optimization Method :Random Search")
self.log.info("-------> Model Name: "+str(modelName))
start = time.time()
opTq =OptimizationTq(algoName,paramDict,self.cvSplit,self.scoreParam,self.numIter,self.trainX,self.trainY)
model,modelParams,score,estimator=opTq.randomSearchOpt()
executionTime=time.time() - start
if not self.testX.empty:
predictedData = estimator.predict(self.testX)
score = objClf.get_score(self.scoreParam,self.testY,predictedData)
else:
score = score*100
problemName = estimator.__class__.__name__
runName = algoName + '_' + modelOrFeatureBased
metrices = {}
metrices["score"] = score
try:
self.logMlflow(runName, modelParams, metrices, estimator, algoName)
except Exception as e:
self.log.info('----------> ML Flow error!!!. ' + str(e)) # usnish
pass
import numpy as np
if self.MakeFP0:
self.log.info('-------- Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = self.check_threshold(estimator,self.trainX,self.trainY,threshold_range,'FP',algoName)
self.log.info('-------- Calculate Threshold for FP End-------')
if self.MakeFN0:
self.log.info('-------- Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = self.check_threshold(estimator,self.trainX,self.trainY,threshold_range,'FN',algoName)
self.log.info('-------- Calculate Threshold for FN End-------')
if threshold != -1:
if not self.testX.empty:
predictedData = estimator.predict_proba(self.testX)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold)
score = objClf.get_score(self.scoreParam,self.testY,predictedData)
else:
predictedData = estimator.predict_proba(self.trainX)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold)
score = objClf.get_score(self.scoreParam,self.trainY,predictedData)
self.log.info('---------> Total Execution: '+str(executionTime)+'\n')
output_jsonobject = ""
problemName = estimator.__class__.__name__
self.log.info('----------> Testing Score: '+str(score))
try:
if ((estimator.__class__.__name__ == "ABCMeta") or (model in ['SGDClassifier','XGBClassifier','CatBoostClassifier','LGBMClassifier']) ):
self.log.info('-----> Model Uncertainty Not Supported')
else:
uqObj=aionUQ(None,None,None,problemName,modelParams,estimator,None,None,self.deployLocation)
accuracy,uq_ece,output_jsonobject,model_confidence_per,model_uncertainty_per=uqObj.uqMain_BBMClassification(self.uq_x_train,self.uq_x_test,self.uq_y_train,self.uq_y_test,"aionuq")
self.log.info("-------> model_confidence: "+str(model_confidence_per)+str('%'))
self.log.info("-------> model_uncertainty: "+str(model_uncertainty_per)+str('%'))
except Exception as e:
pass
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(score)+',"ModelUncertainty":'+str(json.dumps(output_jsonobject))+'}'
status,bscore,bthres,brscore,bpscore = self.getBestModel(self.MakeFP0,self.MakeFN0,threshold,bestthreshold,recallscore,bestrecallscore,precisionscore,bestprecisionscore,score,bestScore)
if status:
bestScore =bscore
bestModel =model
bestParams=modelParams
bestEstimator=estimator
bestthreshold = threshold
bestrecallscore = recallscore
bestprecisionscore = precisionscore
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------'+ modelName+' Model Execution failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
self.log.info('\n < ---------- Model Execution Failed End --------->')
trainingStatus = 'Error (Exception)'
elif self.method == 'bayesopt':
try:
self.log.info("-------> Optimization Method :BayesOpt")
self.log.info("-------> Model Name: "+str(modelName))
opTq =OptimizationTq(algoName,paramDict,self.cvSplit,self.scoreParam,self.numIter,self.trainX,self.trainY)
fun=opTq.f
trials = Trials()
start = time.time()
best = fmin(fun,paramDict,algo=tpe.suggest, max_evals=self.numIter, trials=trials)
executionTime=time.time() - start
results = sorted(trials.results, key = lambda x: x['loss'])
bestresult=results[0]
model=bestresult['model']
score=bestresult['score']
modelParams=bestresult['params']
executionTime=time.time() - start
res = ', '.join("{!s}={!r}".format(key,val) for (key,val) in modelParams.items())
modelObj=eval(model+'('+res+')')
estimator = modelObj.fit(self.trainX,self.trainY)
if not self.testX.empty:
predictedData = estimator.predict(self.testX)
score = objClf.get_score(self.scoreParam,self.testY,predictedData)
problemName = estimator.__class__.__name__
runName = algoName + '_' + modelOrFeatureBased
metrices = {}
metrices["score"] = score
try:
self.logMlflow(runName, modelParams, metrices, estimator, algoName)
except Exception as e:
self.log.info('----------> ML Flow error!!!. ' + str(e)) # usnish
pass
output_jsonobject = ""
problemName = estimator.__class__.__name__
self.log.info('----------> Testing Score: '+str(score))
try:
if ((estimator.__class__.__name__ == "ABCMeta") or (model in ['SGDClassifier','XGBClassifier','CatBoostClassifier','LGBMClassifier']) ):
self.log.info('-----> Model Uncertainty Not Supported')
else:
uqObj=aionUQ(None,None,None,problemName,modelParams,estimator,None,None,self.deployLocation)
accuracy,uq_ece,output_jsonobject,model_confidence_per,model_uncertainty_per=uqObj.uqMain_BBMClassification(self.uq_x_train,self.uq_x_test,self.uq_y_train,self.uq_y_test,"aionuq")
self.log.info("-------> model_confidence: "+str(model_confidence_per)+str('%'))
self.log.info("-------> model_uncertainty: "+str(model_uncertainty_per)+str('%'))
except:
pass
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","FeatureEngineering":"'+str(modelOrFeatureBased)+'","Score":'+str(score)+',"ModelUncertainty":'+str(json.dumps(output_jsonobject))+'}'
'''
test_accuracy = accuracy_score(self.testY,predictedData)
test_precision = precision_score(self.testY,predictedData,average='macro')
self.log.info('---------> Test Accuracy: '+str(test_accuracy))
self.log.info('---------> Test Precision: '+str(test_precision))
'''
import numpy as np
if self.MakeFP0:
self.log.info('-------- Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = self.check_threshold(estimator,self.testX,self.testY,threshold_range,'FP',algoName)
self.log.info('-------- Calculate Threshold for FP End-------')
if self.MakeFN0:
self.log.info('-------- Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = self.check_threshold(estimator,self.testX,self.testY,threshold_range,'FN',algoName)
self.log.info('-------- Calculate Threshold for FN End-------')
self.log.info('---------> Total Execution: '+str(executionTime)+'\n')
status,bscore,bthres,brscore,bpscore = self.getBestModel(self.MakeFP0,self.MakeFN0,threshold,bestthreshold,recallscore,bestrecallscore,precisionscore,bestprecisionscore,score,bestScore)
if status:
bestScore =score
bestModel =model
bestParams=modelParams
res = ', '.join("{!s}={!r}".format(key,val) for (key,val) in bestParams.items())
modelObj=eval(bestModel+'('+res+')')
bestEstimator=estimator
bestthreshold = threshold
bestrecallscore = recallscore
bestprecisionscore = precisionscore
except Exception as inst:
self.log.info('\n < ---------- Model Execution Failed Start--------->')
self.log.info('\n<-------'+ modelName+' Model Execution failed!!!.'+str(inst))
self.log.info('\n < ---------- Model Execution Failed End --------->')
trainingStatus = 'Error (Exception)'
else:
trainingStatus = 'Error (HyperTunning Algo Not Supported)'
pass
self.log.info('Status:- |... ML Algorithm applied: '+modelName)
if trainingStatus.lower() == 'success':
self.log.info('Status:- |... Score after hyperparameter tuning: '+objClf.get_print_score(self.scoreParam)+'='+str(round(score,2))+'\n')
else:
self.log.info('Status:- |... Training Error : '+trainingStatus+'\n')
self.log.info('---------- ClassifierModel End ---------- \n')
if bestModel != 'None':
self.log.info('\n------- Best Model and its parameters -------------')
self.log.info('Status:- |... Best Algorithm selected: '+str(self.modelToAlgoNames[bestModel])+' Score='+str(round(bestScore,2)))
self.log.info("-------> Best Name: "+str(bestModel))
self.log.info("-------> Best Score: "+str(bestScore))
return self.modelToAlgoNames[bestModel],bestParams,bestScore,bestEstimator,scoredetails,bestthreshold,bestprecisionscore,bestrecallscore
else:
raise Exception("Sorry, no model is trained")
except Exception as inst:
self.log.info( '\n-----> ClassifierModel failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno)) |
aion_matrix.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import warnings
warnings.filterwarnings('ignore')
import pandas as pd
from sklearn.metrics import confusion_matrix
from sklearn.metrics import classification_report
from sklearn.metrics import roc_curve, auc
from sklearn.metrics import roc_auc_score
from sklearn.metrics import accuracy_score
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error,make_scorer
from sklearn.metrics import mean_squared_error
from sklearn.metrics import recall_score
from sklearn.metrics import precision_score
from sklearn.metrics import f1_score
import logging
import numpy as np
from sklearn.preprocessing import binarize
from sklearn.preprocessing import LabelBinarizer
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error,make_scorer
from sklearn.metrics import mean_squared_error
class aion_matrix:
def __init__(self):
self.log = logging.getLogger('eion')
def get_print_score(self,matrix):
if 'accuracy' in str(matrix).lower():
return 'Accuracy'
elif 'recall' in str(matrix).lower():
return 'Recall'
elif 'precision' in str(matrix).lower():
return 'Precision'
elif 'f1_score' in str(matrix).lower():
return 'F1_Score'
elif 'roc_auc' in str(matrix).lower():
return 'ROC_AUC'
elif 'mse' in str(matrix).lower() or 'neg_mean_squared_error' in str(matrix).lower():
return 'Mean Squared Error(MSE)'
elif 'rmse' in str(matrix).lower() or 'neg_root_mean_squared_error' in str(matrix).lower():
return 'Root Mean Suared Error(RMSE)'
elif 'mae' in str(matrix).lower() or 'neg_mean_absolute_error' in str(matrix).lower():
return 'Mean Absolute Error (MAE)'
elif 'r2' in str(matrix).lower():
return 'R-Squared(R2)'
else:
return 'Unknown'
def get_score(self,matrix,actual,predict):
if 'accuracy' in str(matrix).lower():
ensemble_score = accuracy_score(actual,predict)
ensemble_score = ensemble_score*100
elif 'recall' in str(matrix).lower():
ensemble_score = recall_score(actual,predict,average='macro')
ensemble_score = ensemble_score*100
elif 'precision' in str(matrix).lower():
ensemble_score = precision_score(actual,predict,average='macro')
ensemble_score = ensemble_score*100
elif 'f1_score' in str(matrix).lower():
ensemble_score = f1_score(actual,predict, average='macro')
ensemble_score = ensemble_score*100
elif 'roc_auc' in str(matrix).lower():
try:
ensemble_score = roc_auc_score(actual,predict,average="macro")
except:
try:
actual = pd.get_dummies(actual)
predict = pd.get_dummies(predict)
ensemble_score = roc_auc_score(actual,predict, average='weighted', multi_class='ovr')
except:
ensemble_score = 0
ensemble_score = ensemble_score*100
elif ('mse' in str(matrix).lower()) or ('neg_mean_squared_error' in str(matrix).lower()):
ensemble_score = mean_squared_error(actual,predict)
elif ('rmse' in str(matrix).lower()) or ('neg_root_mean_squared_error' in str(matrix).lower()):
ensemble_score=mean_squared_error(actual,predict,squared=False)
elif ('mae' in str(matrix).lower()) or ('neg_mean_absolute_error' in str(matrix).lower()):
ensemble_score=mean_absolute_error(actual,predict)
elif 'r2' in str(matrix).lower():
ensemble_score=r2_score(actual,predict)
return round(ensemble_score,2)
def getClassificationPerformaceMatrix(self,le_trainY,predictedData,labelMaps):
setOfyTrue = set(le_trainY)
unqClassLst = list(setOfyTrue)
if(str(labelMaps) != '{}'):
inv_mapping_dict = {v: k for k, v in labelMaps.items()}
unqClassLst2 = (pd.Series(unqClassLst)).map(inv_mapping_dict)
unqClassLst2 = list(unqClassLst2)
else:
unqClassLst2 = unqClassLst
indexName = []
columnName = []
targetnames=[]
for item in unqClassLst2:
indexName.append("act:"+str(item))
columnName.append("pre:"+str(item))
targetnames.append(str(item))
matrixconfusion = pd.DataFrame(confusion_matrix(le_trainY,predictedData, labels = unqClassLst),index = indexName, columns = columnName)
#pd.set_option('expand_frame_repr', False)
pd.set_option('display.max_columns',len(targetnames)+2)
self.log.info('-------> Confusion Matrix: ')
self.log.info(matrixconfusion)
pd.reset_option('display.max_columns')
#pd.reset_option('expand_frame_repr')
#self.log.info('-------> Confusion Matrix With Labels: ')
#self.log.info(confusion_matrix(le_trainY,predictedData, labels = unqClassLst))
#print(unqClassLst2)
classificationreport = pd.DataFrame(classification_report(le_trainY, predictedData, labels = unqClassLst,target_names=targetnames,output_dict=True)).transpose()
self.log.info('-------> Classification Report: ')
self.log.info(classificationreport)
lb = LabelBinarizer()
lb.fit(le_trainY)
transformTarget= lb.transform(le_trainY)
transformPredict = lb.transform(predictedData)
rocaucscore = roc_auc_score(transformTarget,transformPredict,average="macro")
self.log.info('-------> ROC AUC SCORE :'+str(rocaucscore))
matrixconfusion = matrixconfusion.to_json(orient='index')
classificationreport = classificationreport.to_json(orient='index')
matrix = '"ConfusionMatrix":'+matrixconfusion+',"ClassificationReport":'+classificationreport+',"ROC_AUC_SCORE":'+str(rocaucscore)
return(matrix)
def get_regression_matrix(self,targetData,predictedData):
r2score=r2_score(targetData, predictedData)
self.log.info('-------> R2_score :'+str(r2score))
meanabsoluteerror=(mean_absolute_error(targetData, predictedData))
self.log.info('-------> MAE :'+str(meanabsoluteerror))
meanssquatederror=mean_squared_error(targetData, predictedData)
self.log.info('-------> MSE :'+str(meanssquatederror))
rootmeanssquatederror=mean_squared_error(targetData, predictedData,squared=False)
self.log.info('-------> RMSE :'+str(rootmeanssquatederror))
targetArray, predictedArray = np.array(targetData), np.array(predictedData)
try:
EPSILON = 1e-10
meanpercentageerror=np.mean(np.abs((targetArray - predictedArray) / (targetArray+EPSILON)))*100
except ZeroDivisionError:
meanpercentageerror = 0
self.log.info('-------> MAPE :'+str(meanpercentageerror))
try:
normalised_rmse_percentage = round(((rootmeanssquatederror/ ( np.max(targetData) - np.min(targetData) )) * 100), 4)
except Exception as e:
normalised_rmse_percentage = -1
self.log.info('-------> Normalised RMSE percentage :'+str(normalised_rmse_percentage))
matrix = '"MAE":'+str(meanabsoluteerror)+',"R2Score":'+str(r2score)+',"MSE":'+str(meanssquatederror)+',"MAPE":'+str(meanpercentageerror)+',"RMSE":'+str(rootmeanssquatederror)+',"Normalised RMSE(%)":'+str(normalised_rmse_percentage)
return matrix
def getbestfeatureModel(self,modelType,scoreParam,score1,score2,model1,model2,threshold1,pscore1,rscore1,threshold2,pscore2,rscore2):
best_feature_model = 'Model1'
self.log.info('\n ---------- Summary Start ------------')
if modelType.lower() == "classification":
if(threshold1 == -1 and threshold2 == -1):
if score1> score2:
self.log.info('-------> Best Features: Model1')
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model1'
else:
self.log.info('-------> Best Features: Model2')
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model2'
elif(threshold1 == -1):
self.log.info('-------> Best Features: Model2')
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model2'
elif(threshold1 == -2):
self.log.info('-------> Best Features: Model1')
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model1'
else:
if pscore1 == pscore2:
if rscore1 > rscore2:
self.log.info('-------> Best Features: Model1')
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model1'
else:
self.log.info('-------> Best Features: Model2')
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model2'
elif rscore1 == rscore2:
if pscore1 > pscore2:
self.log.info('-------> Best Features: Model1')
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model1'
else:
self.log.info('-------> Best Features: Model2')
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model2'
elif modelType.lower() == "regression":
if scoreParam == "r2" or scoreParam == "explained_variance":
if score1> score2 :
self.log.info('-------> Best Features: Model1')
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model1'
else:
self.log.info('-------> Best Features: Model2')
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model2'
else:
if score1< score2 :
self.log.info('-------> Best Features: Model1')
self.log.info('-------> Best Model: '+str(model1))
self.log.info('-------> Best Score: '+str(score1))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model1'
else:
self.log.info('-------> Best Features: Model2')
self.log.info('-------> Best Model: '+str(model2))
self.log.info('-------> Best Score: '+str(score2))
self.log.info('-------> Scoring Param: '+str(scoreParam))
best_feature_model = 'Model2'
self.log.info('---------- Summary End ------------\n')
return(best_feature_model)
def check_threshold(self,estimator,testX,testY,threshold_range,checkParameter,modelName):
thresholdx = -1
for threshold in threshold_range:
predictedData = estimator.predict_proba(testX)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold)#bug 12437
p_score = precision_score(testY, predictedData)
#self.log.info('-------------> Precision:'+str(p_score))
r_score = recall_score(testY, predictedData)
#self.log.info('-------------> Rscore:'+str(r_score))
#self.log.info(confusion_matrix(testY, predictedData))
tn, fp, fn, tp = confusion_matrix(testY, predictedData).ravel()
if(checkParameter.lower() == 'fp'):
if fp == 0:
if(p_score == 1):
thresholdx = threshold
self.log.info('---------------> Best Threshold:'+str(threshold))
self.log.info('---------------> Best Precision:'+str(p_score))
self.log.info('---------------> Best Recall:'+str(r_score))
self.log.info('---------------> TN:'+str(tn))
self.log.info('---------------> FP:'+str(fp))
self.log.info('---------------> FN:'+str(fn))
self.log.info('---------------> TP:'+str(tp))
break
if(checkParameter.lower() == 'fn'):
if fn == 0:
if(r_score == 1):
thresholdx = threshold
self.log.info('---------------> Best Threshold:'+str(threshold))
self.log.info('---------------> Best Precision:'+str(p_score))
self.log.info('---------------> Best Recall:'+str(r_score))
self.log.info('---------------> TN:'+str(tn))
self.log.info('---------------> FP:'+str(fp))
self.log.info('---------------> FN:'+str(fn))
self.log.info('---------------> TP:'+str(tp))
break
return(thresholdx,p_score,r_score)
def getBestModel(self,fp0,fn0,threshold,bestthreshold,rscore,brscore,pscore,bpscore,tscore,btscore):
cmodel = False
if(threshold != -1):
if(bestthreshold == -1):
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif fp0:
if rscore > brscore:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif rscore == brscore:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif fn0:
if pscore > bpscore:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif pscore == bpscore:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
else:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
btscore = tscore
else:
if(bestthreshold == -1):
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
btscore = tscore
return cmodel,btscore,bestthreshold,brscore,bpscore |
ImageLearning.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import os
import tensorflow
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense, Dropout, Flatten
from tensorflow.keras.layers import Conv2D, MaxPooling2D
from tensorflow.keras.utils import to_categorical
from tensorflow.keras.preprocessing import image
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from tensorflow.keras.layers import Input
from tensorflow.keras.models import Model
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.applications import VGG16
from tensorflow.keras.callbacks import EarlyStopping
import logging
from sklearn.preprocessing import LabelEncoder
from statistics import mean
import sys
from learner.machinelearning import machinelearning
from learner.aion_matrix import aion_matrix
from profiler.imageAug import ImageAugmentation
from pathlib import Path
class ImageLearning:
def __init__(self,dataFrame,input_directory,outputdir,modelname,hyperParam, AugEnabled,keepAugImages,operations,augConf):
self.image_list = dataFrame
self.input_directory = input_directory
self.outputdir = outputdir
self.modelname = modelname
self.hyperParam = hyperParam
self.labelMapping={}
self.log = logging.getLogger('eion')
self.AIONNumericDtypes=['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
self.AugEnabled = AugEnabled
self.keepAugImages = keepAugImages
self.operations = operations
self.augConf = augConf
def TrainCAST(self,predicted_data_file):
datatype = self.image_list['Label'].dtypes
if datatype not in self.AIONNumericDtypes:
labelEncode= LabelEncoder()
self.image_list['Label'] = self.image_list['Label'].apply(str)
self.image_list['Label'] = labelEncode.fit_transform(self.image_list['Label'])
self.labelMapping = dict(zip(labelEncode.classes_, labelEncode.transform(labelEncode.classes_)))
self.log.info('\n-------> First Ten Rows of Input Data After Encoding: ')
self.log.info(self.image_list.head(10))
self.log.info('Status:- |... Target Feature Encoding Done')
if not os.path.exists(self.outputdir):
os.makedirs(self.outputdir)
train_df, test_df = train_test_split(self.image_list, random_state=42, test_size=self.hyperParam['test_split_ratio'])
if self.AugEnabled:
csv_file = "tempTrainDf.csv"
train_df.to_csv(csv_file, index=False)
ia = ImageAugmentation(self.input_directory, csv_file)
csv_file = ia.augment("imageclassification", self.operations,None,self.augConf)
train_df = pd.read_csv(csv_file)
Path(csv_file).unlink()
train_image = []
train_df.reset_index(drop=True, inplace=True)
for i in range(train_df.shape[0]):
#print(os.path.join(self.input_directory,str(self.image_list['File'][i])))
img = image.load_img(os.path.join(self.input_directory,str(train_df['File'][i])), target_size=(self.hyperParam['img_width'],self.hyperParam['img_height'],self.hyperParam['img_channel']), grayscale=False)
img = image.img_to_array(img)
img = img/255
train_image.append(img)
test_image = []
test_df.reset_index(drop=True, inplace=True)
for i in range(test_df.shape[0]):
#print(os.path.join(self.input_directory,str(self.image_list['File'][i])))
img = image.load_img(os.path.join(self.input_directory,str(test_df['File'][i])), target_size=(self.hyperParam['img_width'],self.hyperParam['img_height'],self.hyperParam['img_channel']), grayscale=False)
img = image.img_to_array(img)
img = img/255
test_image.append(img)
self.log.info('Status:- |... Image Loading Done')
X_train = np.array(train_image)
y_train = train_df['Label']
X_test = np.array(test_image)
y_test = test_df['Label']
ytrain = y_train.values
ytrain = to_categorical(ytrain)
ytest = y_test.values
ytest = to_categorical(ytest)
#print(y)
self.log.info("Loading Imagenet Weights...")
if self.modelname == "densenet":
self.log.info('Loading Densenet model')
baseModel = tensorflow.keras.applications.DenseNet121(weights="imagenet", include_top=False, input_tensor=Input(shape=(self.hyperParam['img_width'], self.hyperParam['img_height'], self.hyperParam['img_channel']))) #98
elif self.modelname == "inception":
self.log.info('Loading Inception model')
baseModel = tensorflow.keras.applications.InceptionV3(weights="imagenet", include_top=False, input_tensor=Input(shape=(self.hyperParam['img_width'], self.hyperParam['img_height'], self.hyperParam['img_channel']))) #97
headModel = baseModel.output
headModel = Flatten(name="flatten")(headModel)
headModel = Dense(1024, activation='relu')(headModel)
headModel = Dropout(0.5)(headModel)
headModel = Dense(2, activation='sigmoid')(headModel)
model = Model(inputs=baseModel.input, outputs=headModel)
self.log.info("[INFO] compiling model...")
opt = Adam(lr=self.hyperParam['lr'])
model.compile(loss="binary_crossentropy", optimizer=opt, metrics=["accuracy"])
#early_stop = EarlyStopping(monitor='val_loss',patience=2)
#history = model.fit(X_train, y_train, epochs=hyperparam_config['epochs'], validation_data=(X_test, y_test), callbacks=[early_stop])
history = model.fit(X_train, ytrain, epochs=self.hyperParam['epochs'], validation_data=(X_test, ytest))
self.log.info('Status:- |... Image Classification Algorithm applied:'+str(self.modelname))
#Saving trained model weights
model.save_weights(os.path.join(self.outputdir, self.modelname))
saved_model = self.modelname
modelname = self.modelname
prediction = model.predict(X_train)
predictedData = np.argmax(prediction,axis=1)
mlobj = machinelearning()
self.log.info('\n--------- Performance Matrix with Train Data ---------')
trainingperformancematrix = mlobj.getClassificationPerformaceMatrix(y_train, predictedData,self.labelMapping)
prediction = model.predict(X_test)
predictedData = np.argmax(prediction,axis=1)
self.log.info('\n--------- Performance Matrix with Test Data ---------')
performancematrix = mlobj.getClassificationPerformaceMatrix(y_test, predictedData,self.labelMapping)
df_test = pd.DataFrame()
df_test['actual'] = y_test
df_test['predict'] = predictedData
df_test.to_csv(predicted_data_file)
objClf = aion_matrix()
scoring_param = 'Accuracy'
score = objClf.get_score(scoring_param,y_test,predictedData)
#score = mean(history.history['accuracy'])
if self.AugEnabled and not self.keepAugImages:
ia.removeAugmentedImages(train_df)
scoredetails = '{"Model":"'+modelname+'","Score":'+str(round(score,2))+'}'
self.log.info('Status:- |... Score Accuracy: '+str(round(score,2)))
return saved_model,modelname,'ImageClassification',scoring_param,score,scoredetails,self.labelMapping,trainingperformancematrix,performancematrix
|
image_eda.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import warnings
warnings.simplefilter("ignore")
import os
import numpy as np
from numpy import asarray
import cv2
import sys
import random
import glob as glob
import math as m
# for gamma function, called
from scipy.special import gamma as tgamma
import matplotlib.image as mpimg
import skimage
from libsvm import svmutil,svm
#import svmutil
from svmutil import *
from svm import *
from PIL import Image
from collections import Counter
from imutils import paths
import matplotlib.pyplot as plt
import json
###################################################################################
#Input - # AGGD fit model, takes input as the MSCN Image / Pair-wise Product
#Output - best values of image parameters
#Defination - used as internal method to measure_ImageQualityScore
###################################################################################
def AGGDfit(structdis):
# variables to count positive pixels / negative pixels and their squared sum
poscount = 0
negcount = 0
possqsum = 0
negsqsum = 0
abssum = 0
poscount = len(structdis[structdis > 0]) # number of positive pixels
negcount = len(structdis[structdis < 0]) # number of negative pixels
# calculate squared sum of positive pixels and negative pixels
possqsum = np.sum(np.power(structdis[structdis > 0], 2))
negsqsum = np.sum(np.power(structdis[structdis < 0], 2))
# absolute squared sum
abssum = np.sum(structdis[structdis > 0]) + np.sum(-1 * structdis[structdis < 0])
# calculate left sigma variance and right sigma variance
lsigma_best = np.sqrt((negsqsum/negcount))
rsigma_best = np.sqrt((possqsum/poscount))
gammahat = lsigma_best/rsigma_best
# total number of pixels - totalcount
totalcount = structdis.shape[1] * structdis.shape[0]
rhat = m.pow(abssum/totalcount, 2)/((negsqsum + possqsum)/totalcount)
rhatnorm = rhat * (m.pow(gammahat, 3) + 1) * (gammahat + 1)/(m.pow(m.pow(gammahat, 2) + 1, 2))
prevgamma = 0
prevdiff = 1e10
sampling = 0.001
gam = 0.2
# vectorized function call for best fitting parameters
vectfunc = np.vectorize(func, otypes = [np.float], cache = False)
# calculate best fit params
gamma_best = vectfunc(gam, prevgamma, prevdiff, sampling, rhatnorm)
return [lsigma_best, rsigma_best, gamma_best]
def func(gam, prevgamma, prevdiff, sampling, rhatnorm):
while(gam < 10):
r_gam = tgamma(2/gam) * tgamma(2/gam) / (tgamma(1/gam) * tgamma(3/gam))
diff = abs(r_gam - rhatnorm)
if(diff > prevdiff): break
prevdiff = diff
prevgamma = gam
gam += sampling
gamma_best = prevgamma
return gamma_best
def compute_features(img):
scalenum = 2
feat = []
# make a copy of the image
im_original = img.copy()
# scale the images twice
for itr_scale in range(scalenum):
im = im_original.copy()
# normalize the image
im = im / 255.0
# calculating MSCN coefficients
mu = cv2.GaussianBlur(im, (7, 7), 1.166)
mu_sq = mu * mu
sigma = cv2.GaussianBlur(im*im, (7, 7), 1.166)
sigma = (sigma - mu_sq)**0.5
# structdis is the MSCN image
structdis = im - mu
structdis /= (sigma + 1.0/255)
# calculate best fitted parameters from MSCN image
best_fit_params = AGGDfit(structdis)
# unwrap the best fit parameters
lsigma_best = best_fit_params[0]
rsigma_best = best_fit_params[1]
gamma_best = best_fit_params[2]
# append the best fit parameters for MSCN image
feat.append(gamma_best)
feat.append((lsigma_best*lsigma_best + rsigma_best*rsigma_best)/2)
# shifting indices for creating pair-wise products
shifts = [[0,1], [1,0], [1,1], [-1,1]] # H V D1 D2
for itr_shift in range(1, len(shifts) + 1):
OrigArr = structdis
reqshift = shifts[itr_shift-1] # shifting index
# create transformation matrix for warpAffine function
M = np.float32([[1, 0, reqshift[1]], [0, 1, reqshift[0]]])
ShiftArr = cv2.warpAffine(OrigArr, M, (structdis.shape[1], structdis.shape[0]))
Shifted_new_structdis = ShiftArr
Shifted_new_structdis = Shifted_new_structdis * structdis
# shifted_new_structdis is the pairwise product
# best fit the pairwise product
best_fit_params = AGGDfit(Shifted_new_structdis)
lsigma_best = best_fit_params[0]
rsigma_best = best_fit_params[1]
gamma_best = best_fit_params[2]
constant = m.pow(tgamma(1/gamma_best), 0.5)/m.pow(tgamma(3/gamma_best), 0.5)
meanparam = (rsigma_best - lsigma_best) * (tgamma(2/gamma_best)/tgamma(1/gamma_best)) * constant
# append the best fit calculated parameters
feat.append(gamma_best) # gamma best
feat.append(meanparam) # mean shape
feat.append(m.pow(lsigma_best, 2)) # left variance square
feat.append(m.pow(rsigma_best, 2)) # right variance square
# resize the image on next iteration
im_original = cv2.resize(im_original, (0,0), fx=0.5, fy=0.5, interpolation=cv2.INTER_CUBIC)
return feat
def img_MeasureImageQuality(dataset_directory):
"""
####################################################################################
#Input - img_path
#Output - Quality index of input image
#Defination - function to calculate BRISQUE quality score in range of 0 and 100 [0:good;100:bad]
####################################################################################
"""
imgfile_dict = {}
for file in os.listdir(dataset_directory):
if (file.endswith(".jfif") or file.endswith(".png") or file.endswith(".jpg") or file.endswith(".jpeg")):
filename = os.path.join(dataset_directory , file)
if os.path.isfile(filename)==False:
sys.exit()
file_extension = os.path.splitext(filename)[1]
if file_extension==".jfif":
extension=".jfif"
if file_extension==".png":
extension=".png"
if file_extension==".jpg":
extension=".jpg"
if file_extension==".jpeg":
extension=".jpeg"
if (extension not in [".jpg",".jpeg",".jfif",".png"]):
sys.exit()
try:
# read image from given path
dis = cv2.imread(filename, 1)
if(dis is None):
sys.exit(0)
# convert to gray scale
dis = cv2.cvtColor(dis, cv2.COLOR_BGR2GRAY)
# compute feature vectors of the image
features = compute_features(dis)
# rescale the brisqueFeatures vector from -1 to 1
x = [0]
# pre loaded lists from C++ Module to rescale brisquefeatures vector to [-1, 1]
min_= [0.336999 ,0.019667 ,0.230000 ,-0.125959 ,0.000167 ,0.000616 ,0.231000 ,-0.125873 ,0.000165 ,0.000600 ,0.241000 ,-0.128814 ,0.000179 ,0.000386 ,0.243000 ,- 0.133080 ,0.000182 ,0.000421 ,0.436998 ,0.016929 ,0.247000 ,-0.200231 ,0.000104 ,0.000834 ,0.257000 ,-0.200017 ,0.000112 ,0.000876 ,0.257000 ,-0.155072 , 0.000112 ,0.000356 ,0.258000 ,-0.154374 ,0.000117 ,0.000351]
max_= [9.999411, 0.807472, 1.644021, 0.202917, 0.712384, 0.468672, 1.644021, 0.169548, 0.713132, 0.467896, 1.553016, 0.101368, 0.687324, 0.533087, 1.554016, 0.101000 , 0.689177, 0.533133, 3.639918, 0.800955, 1.096995, 0.175286, 0.755547, 0.399270, 1.095995, 0.155928, 0.751488, 0.402398, 1.041992, 0.093209, 0.623516, 0.532925, 1.042992, 0.093714, 0.621958, 0.534484]
# append the rescaled vector to x
for i in range(0, 36):
min = min_[i]
max = max_[i]
x.append(-1 + (2.0/(max - min) * (features[i] - min)))
modelPath = os.path.join(os.path.dirname(os.path.abspath(__file__)),'allmodel.txt')
# load model
model = svmutil.svm_load_model(modelPath)
# create svm node array from python list
x, idx = gen_svm_nodearray(x[1:], isKernel=(model.param.kernel_type == PRECOMPUTED))
x[36].index = -1 # set last index to -1 to indicate the end.
# get important parameters from model
svm_type = model.get_svm_type()
is_prob_model = model.is_probability_model()
nr_class = model.get_nr_class()
if svm_type in (ONE_CLASS, EPSILON_SVR, NU_SVC):
# here svm_type is EPSILON_SVR as it's regression problem
nr_classifier = 1
dec_values = (c_double * nr_classifier)()
# calculate the quality score of the image using the model and svm_node_array
qualityscore = svmutil.libsvm.svm_predict_probability(model, x, dec_values)
imgfile_dict[file] = round(qualityscore,2)
#print ("Quality Score of the given image is: ", qualityscore, "[0:Good;100:Bad]")
except:
pass
finally:
warnings.simplefilter("ignore")
#print(imgfile_dict)
return imgfile_dict
# calculate moode
def mode(arr):
if arr==[]:
return None
else:
return max(set(arr), key=arr.count)
def img_EDA(dataset_directory):
"""
####################################################################################
#Input - dataset_directory with all type of Images
#Output - mean,median and mode image size, channels type, extensions, recommendation of images etc
#Defination - img_EDA takes the all images and print the EDA results
####################################################################################
"""
imgeda_dict = {}
# check input directory
if os.path.isdir(dataset_directory)==False:
print("folder does not exist")
sys.exit()
width_list=[]
height_list=[]
k=[]
c=[]
cnum=[]
v=[]
ext=[]
cnt=0
for item in os.listdir(dataset_directory):
if (item.endswith(".jfif") or item.endswith(".png") or item.endswith(".jpg") or item.endswith(".jpeg")):
if os.path.isfile(os.path.join(dataset_directory , item)):
im = Image.open(os.path.join(dataset_directory , item))
c.append(im.mode)
cnum.append(len(im.mode))
width_list.append(im.width)
height_list.append(im.height)
k.append(im.size)
v.append(im.width*im.height)
f, e = os.path.splitext(os.path.join(dataset_directory , item))
ext.append(e)
cnt=cnt+1
# calculate biggest and smallest image
img_dict={}
for key, val in zip(k, v):
img_dict[key] = val
max_key = max(img_dict, key=img_dict.get)
#max_key
min_key = min(img_dict, key=img_dict.get)
#min_key
imgeda_dict['Channels'] = set(c)
imgeda_dict['Extensions'] = set(ext)
imgeda_dict['Total_Images'] = cnt
imgeda_dict['Smallest_Image'] = min_key
imgeda_dict['Largest_Image'] = max_key
imgeda_dict['Mean_Width'] = int(np.mean(width_list))
imgeda_dict['Mean_Height'] = int(np.mean(height_list))
imgeda_dict['Median_Width'] = int(np.median(width_list))
imgeda_dict['Median_Height'] = int(np.median(height_list))
imgeda_dict['Mode_Width'] = int(mode(width_list))
imgeda_dict['Mode_Height'] = int(mode(height_list))
imgeda_dict['Recomended_Mean_Width_Height'] = (int(np.mean(width_list)),int(np.mean(height_list)))
imgeda_dict['Recomended_Median_Width_Height'] = (int(np.median(width_list)),int(np.median(height_list)))
imgeda_dict['Recomended_Mode_Width_Height'] = (int(mode(width_list)),int(mode(height_list)))
imgeda_dict['Size_Distribution'] = dict(Counter(k).items())
imgeda_dict['Channel_Mean'] = np.mean(cnum)
imgeda_dict['Channel_Standard_Deviation'] = np.std(cnum)
'''
print('*-----------------------<<< RESULTS >>>-------------------------*')
print()
print('%-30s | ' % 'Channels', set(c))
print('%-30s | ' % 'Extensions', set(ext))
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'Total Images', cnt)
print('%-30s | ' % 'Smallest Image', min_key)
print('%-30s | ' % 'Largest Image', max_key)
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'Mean Width', int(np.mean(width_list)))
print('%-30s | ' % 'Mean Height', int(np.mean(height_list)))
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'Median Width', int(np.median(width_list)))
print('%-30s | ' % 'Median Height', int(np.median(height_list)))
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'Mode Width', int(mode(width_list)))
print('%-30s | ' % 'Mode Height', int(mode(height_list)))
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'recommended size by mean(w,h)',(int(np.mean(width_list)),int(np.mean(height_list))))
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'recommended size by median(w,h)',(int(np.median(width_list)),int(np.median(height_list))))
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'recommended size by mode(w,h)',(int(mode(width_list)),int(mode(height_list))))
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'distribution of sizes',dict(Counter(k).items()) )
print('*---------------------------------------------------------------*')
print('%-30s | ' % 'channel mean',np.mean(cnum))
print('%-30s | ' % 'channel standard deviation',np.std(cnum))
'''
#print(imgeda_dict)
return imgeda_dict
def dhash(image, hashSize=8):
# convert the image to grayscale and resize the grayscale image,
# adding a single column (width) so we can compute the horizontal
# gradient
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
resized = cv2.resize(gray, (hashSize + 1, hashSize))
# compute the (relative) horizontal gradient between adjacent
# column pixels
diff = resized[:, 1:] > resized[:, :-1]
# convert the difference image to a hash and return it
return sum([2 ** i for (i, v) in enumerate(diff.flatten()) if v])
def img_duplicatefinder(dataset_directory):
# grab the paths to all images in our input dataset directory and
# then initialize our hashes dictionary
print("[INFO] computing image hashes...")
imagePaths = list(paths.list_images(dataset_directory))
hashes = {}
duplimg_list = []
remove_file = 0
# loop over our image paths
for imagePath in imagePaths:
# load the input image and compute the hash
image = cv2.imread(imagePath)
h = dhash(image)
# grab all image paths with that hash, add the current image
# path to it, and store the list back in the hashes dictionary
p = hashes.get(h, [])
p.append(imagePath)
hashes[h] = p
# loop over the image hashes
for (h, hashedPaths) in hashes.items():
# check to see if there is more than one image with the same hash
if len(hashedPaths) > 1:
#print(hashedPaths)
duplimg_list.append(hashedPaths)
return duplimg_list
def img_plot_colour_hist(dataset_directory):
import io, base64, urllib
red_values = []; green_values = []; blue_values = []; all_channels = []
imagePaths = list(paths.list_images(dataset_directory))
for imagePath in imagePaths:
img = np.array(Image.open(imagePath))
red_values.append(np.mean(img[:, :, 0]))
green_values.append(np.mean(img[:, :, 1]))
blue_values.append(np.mean(img[:, :, 2]))
all_channels.append(np.mean(img))
_, axes = plt.subplots(ncols=4, nrows=1, constrained_layout=True, figsize=(16, 3), sharey=True)
for ax, column, vals, c in zip(
axes,
['red', 'green', 'blue', 'all colours'],
[red_values, green_values, blue_values, all_channels],
'rgbk'
):
ax.hist(vals, bins=100, color=c)
ax.set_title(f'{column} hist')
plt.suptitle("Image Dataset Colour Distribution")
buf = io.BytesIO()
plt.savefig(buf, format='png')
buf.seek(0)
string = base64.b64encode(buf.read())
uri = 'data:image/png;base64,' + urllib.parse.quote(string)
return uri |
predict.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import os
import keras
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras.utils import to_categorical
from keras.preprocessing import image
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from keras.utils import to_categorical
from Fkeras.layers import Input
from keras.models import Model
from keras.optimizers import Adam
from keras.applications import VGG16
from tensorflow.keras.callbacks import EarlyStopping
from sklearn.metrics import classification_report,confusion_matrix,precision_recall_curve
import seaborn as sns
def PredictCAST(test_csv, test_dataset_directory, load_model_dir, model_name, hparams_config_file):
hyperparam_config = hparams_config_file['img_classifier']
print("[Info] loading imagenet weights...")
#baseModel = keras.applications.ResNet101(weights="imagenet", include_top=False, input_tensor=Input(shape=(128, 128, 3)))
if model_name == "densenet":
print('Loading Densenet model')
baseModel = keras.applications.DenseNet121(weights="imagenet", include_top=False, input_tensor=Input(shape=(hyperparam_config['img_width'],hyperparam_config['img_height'],hyperparam_config['img_channel']))) #98
elif model_name == "inception":
print('Loading Inception model')
baseModel = keras.applications.InceptionV3(weights="imagenet", include_top=False, input_tensor=Input(shape=(hyperparam_config['img_width'],hyperparam_config['img_height'],hyperparam_config['img_channel']))) #97
headModel = baseModel.output
headModel = Flatten(name="flatten")(headModel)
headModel = Dense(1024, activation='relu')(headModel)
headModel = Dropout(0.5)(headModel)
headModel = Dense(2, activation='sigmoid')(headModel)
model = Model(inputs=baseModel.input, outputs=headModel)
print("[INFO] compiling model...")
opt = Adam(lr=hyperparam_config['lr'])
model.compile(loss="binary_crossentropy", optimizer=opt, metrics=["accuracy"])
model.load_weights(os.path.join(load_model_dir, model_name))
#model.load_weights(load_model_dir)
test_image = []
for i in range(test_csv.shape[0]):
img = image.load_img(test_dataset_directory + '/' + str(test_csv['file_name'][i]), target_size=(hyperparam_config['img_width'],hyperparam_config['img_height'],hyperparam_config['img_channel']), grayscale=False)
img = image.img_to_array(img)
img = img/255
test_image.append(img)
test_images = np.array(test_image)
test_labels = test_csv['class'].values
test_labels = to_categorical(test_labels)
# making predictions
prediction = model.predict(test_images)
prediction = np.argmax(prediction,axis=1)
print('Classification Report : ')
print(classification_report(test_csv['class'],prediction))
sns.heatmap(confusion_matrix(test_csv['class'],prediction),annot=True)
plt.show()
print('Confusion matrix : ')
print(confusion_matrix(test_csv['class'],prediction))
print("[INFO] Evaluating model accuracy and loss...Take some moment...")
test_loss, test_acc = model.evaluate(test_images, test_labels, verbose=2)
print('\nTest accuracy:', test_acc)
print('\nTest loss:', test_loss)
print("Prediction Completed...")
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
predict_single.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import os
import keras
from keras.models import Sequential
from keras.layers import Dense, Dropout, Flatten
from keras.layers import Conv2D, MaxPooling2D
from keras.utils import to_categorical
from keras.preprocessing import image
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from keras.utils import to_categorical
from keras.layers import Input
from keras.models import Model
from keras.optimizers import Adam
from keras.applications import VGG16
from tensorflow.keras.callbacks import EarlyStopping
from sklearn.metrics import classification_report,confusion_matrix,precision_recall_curve
import seaborn as sns
import cv2
def PredictCAST(test_image, load_model_dir, model_name, hparams_config_file):
hyperparam_config = hparams_config_file['img_classifier']
print("[Info] loading imagenet weights...")
#baseModel = keras.applications.ResNet101(weights="imagenet", include_top=False, input_tensor=Input(shape=(128, 128, 3)))
if model_name == "densenet":
print('Loading Densenet model')
baseModel = keras.applications.DenseNet121(weights="imagenet", include_top=False, input_tensor=Input(shape=(hyperparam_config['img_width'],hyperparam_config['img_height'],hyperparam_config['img_channel']))) #98
elif model_name == "inception":
print('Loading Inception model')
baseModel = keras.applications.InceptionV3(weights="imagenet", include_top=False, input_tensor=Input(shape=(hyperparam_config['img_width'],hyperparam_config['img_height'],hyperparam_config['img_channel']))) #97
headModel = baseModel.output
headModel = Flatten(name="flatten")(headModel)
headModel = Dense(1024, activation='relu')(headModel)
headModel = Dropout(0.5)(headModel)
headModel = Dense(2, activation='sigmoid')(headModel)
model = Model(inputs=baseModel.input, outputs=headModel)
print("[INFO] compiling model...")
opt = Adam(lr=hyperparam_config['lr'])
model.compile(loss="binary_crossentropy", optimizer=opt, metrics=["accuracy"])
model.load_weights(os.path.join(load_model_dir, model_name))
img = cv2.imread(test_image)
img = cv2.resize(img, (hyperparam_config['img_width'],hyperparam_config['img_height']))
orig = img.copy()
img = image.img_to_array(img)
img = np.expand_dims(img, axis=0)
img = img/255
print("[Info] predicting output")
#prediction = model.predict_classes(img)
prediction = model.predict(img)
prediction = np.argmax(prediction,axis=1)
print(prediction)
if (prediction<0.5):
print("def_front")
cv2.putText(orig, "def_front", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
else:
print("ok_front")
cv2.putText(orig, "ok_front", (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (0, 255, 0), 2)
plt.imshow(orig)
plt.axis('off')
plt.show()
print("Prediction Completed...")
|
incMachineLearning.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import warnings
warnings.filterwarnings('ignore')
import logging
import sklearn
from random import sample
from numpy.random import uniform
import numpy as np
import math
import pickle
import os
import json
from math import isnan
from sklearn.preprocessing import binarize
from sklearn.preprocessing import LabelEncoder
import pandas as pd
from sklearn.preprocessing import LabelBinarizer
from sklearn.model_selection import train_test_split
from incremental.incClassificationModel import incClassifierModel
from incremental.incRegressionModel import incRegressionModel
class incMachineLearning(object):
def __init__(self,mlobj):
self.features=[]
self.mlobj=mlobj
self.log = logging.getLogger('eion')
def startLearning(self,mlconfig,modelType,modelParams,modelList,scoreParam,features,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,targetType,deployLocation,iterName,iterVersion,trained_data_file,predicted_data_file,labelMaps):
model = 'None'
params = 'None'
score = 0xFFFF
estimator = None
model_tried = ''
threshold = -1
pscore = -1
rscore = -1
topics = {}
if(targetColumn != ''):
targetData = dataFrame[targetColumn]
datacolumns=list(dataFrame.columns)
if targetColumn in datacolumns:
datacolumns.remove(targetColumn)
scoreParam = self.mlobj.setScoreParams(scoreParam,modelType,categoryCountList)
self.log.info('\n-------------- Training ML: Start --------------')
model_type,model,params, score, estimator,model_tried,xtrain,ytrain,xtest,ytest,threshold,pscore,rscore,method,incObj=self.startLearnerModule(mlconfig,modelType,modelParams,modelList,scoreParam,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,targetType,deployLocation,iterName,iterVersion,trained_data_file,labelMaps)
self.log.info('-------------- Training ML: End --------------\n')
filename = os.path.join(deployLocation,'production','model',model+'.pkl')
saved_model = model+'.pkl'
pickle.dump(estimator, open(filename, 'wb'))
df_test = xtest.copy()
df_test.reset_index(inplace = True,drop=True)
trainPredictedData = incObj.bestTrainPredictedData
predictedData = incObj.bestPredictedData
try:
if(model_type == 'Classification'):
self.log.info('\n--------- Performance Matrix with Train Data ---------')
train_matrix = self.mlobj.getClassificationPerformaceMatrix(ytrain,trainPredictedData,labelMaps)
self.log.info('--------- Performance Matrix with Train Data End ---------\n')
self.log.info('\n--------- Performance Matrix with Test Data ---------')
performancematrix = self.mlobj.getClassificationPerformaceMatrix(ytest,predictedData,labelMaps)
ytest.reset_index(inplace=True,drop=True)
df_test['actual'] = ytest
df_test['predict'] = predictedData
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
matrix = performancematrix
elif(model_type == 'Regression'):
self.log.info('\n--------- Performance Matrix with Train Data ---------')
train_matrix = self.mlobj.get_regression_matrix(ytrain, trainPredictedData)
self.log.info('--------- Performance Matrix with Train Data End ---------\n')
self.log.info('\n--------- Performance Matrix with Test Data ---------')
matrix = self.mlobj.get_regression_matrix(ytest, predictedData)
ytest.reset_index(inplace=True, drop=True)
df_test['actual'] = ytest
df_test['predict'] = predictedData
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
except Exception as Inst:
self.log.info('--------- Error Performance Matrix ---------\n')
self.log.info(str(Inst))
df_test['predict'] = predictedData
matrix = ""
train_matrix = ""
self.log.info('--------- Performance Matrix with Test Data End ---------\n')
df_test.to_csv(predicted_data_file)
return 'Success',model_type,model,saved_model,matrix,train_matrix,xtrain.shape,model_tried,score,filename,self.features,threshold,pscore,rscore,method,estimator,xtrain,ytrain,xtest,ytest,topics,params
def startLearnerModule(self,mlconfig,modelType,modelParams,modelList,scoreParam,targetColumn,dataFrame,xtrain,ytrain,xtest,ytest,categoryCountList,targetType,deployLocation,iterName,iterVersion,trained_data_file,labelMaps):
matrix = ''
threshold = -1
pscore = -1
rscore = -1
datacolumns=list(xtrain.columns)
if targetColumn in datacolumns:
datacolumns.remove(targetColumn)
self.features =datacolumns
self.log.info('-------> Features Used For Training the Model: '+(str(self.features))[:500])
xtrain = xtrain[self.features]
xtest = xtest[self.features]
method = mlconfig['optimizationMethod']
method = method.lower()
geneticParam = ''
optimizationHyperParameter = mlconfig['optimizationHyperParameter']
cvSplit = optimizationHyperParameter['trainTestCVSplit']
nIter = int(optimizationHyperParameter['iterations'])
if(method.lower() == 'genetic'):
geneticParam = optimizationHyperParameter['geneticparams']
scoreParam = scoreParam
if 'thresholdTunning' in mlconfig:
thresholdTunning = mlconfig['thresholdTunning']
else:
thresholdTunning = 'NA'
if cvSplit == "":
cvSplit =None
else:
cvSplit =int(cvSplit)
if modelType == 'classification':
model_type = "Classification"
MakeFP0 = False
MakeFN0 = False
if(len(categoryCountList) == 2):
if(thresholdTunning.lower() == 'fp0'):
MakeFP0 = True
elif(thresholdTunning.lower() == 'fn0'):
MakeFN0 = True
noOfClasses= len(labelMaps)
incObjClf = incClassifierModel(noOfClasses,modelList, modelParams, scoreParam, cvSplit, nIter,geneticParam, xtrain,ytrain,xtest,ytest,method,modelType,MakeFP0,MakeFN0,deployLocation)
model, params, score, estimator,model_tried,threshold,pscore,rscore = incObjClf.firstFit()
incObj = incObjClf
elif modelType == 'regression':
model_type = "Regression"
incObjReg = incRegressionModel(modelList, modelParams, scoreParam, cvSplit, nIter,geneticParam, xtrain,ytrain,xtest,ytest,method,deployLocation)
model,params,score,estimator,model_tried = incObjReg.firstFit()
incObj = incObjReg
return model_type,model,params, score, estimator,model_tried,xtrain,ytrain,xtest,ytest,threshold,pscore,rscore,method, incObj |
incRegressionModel.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
from learner.optimizetechnique import OptimizationTq
from learner.parameters import parametersDefine
import time
import logging
import os
import sys
from sklearn.metrics import r2_score
from sklearn.metrics import mean_absolute_error,make_scorer
from sklearn.metrics import mean_squared_error
from learner.aion_matrix import aion_matrix
class incRegressionModel():
def __init__(self,modelList,params,scoreParam,cvSplit,numIter,geneticParam,trainX,trainY,testX,testY,method,deployLocation):
self.modelList =modelList
self.params =params
self.trainX =trainX
self.trainY =trainY
self.testX = testX
self.testY = testY
self.method =method
self.scoreParam=scoreParam
self.cvSplit=cvSplit
self.numIter=numIter
self.geneticParam=geneticParam
self.log = logging.getLogger('eion')
self.deployLocation = deployLocation
self.bestTrainPredictedData = None
self.bestPredictedData = None
self.AlgorithmNames={'Online Linear Regression':'Online Linear Regression', 'Online Decision Tree Regressor':'Online Decision Tree Regressor', 'Online KNN Regressor':'Online KNN Regressor'}
self.modelToAlgoNames = {value: key for key, value in self.AlgorithmNames.items()}
def firstFit(self):
bestModel=''
bestParams={}
import sys
bestScore=-sys.float_info.max #bugfix 11656
scoredetails = ''
self.log.info('\n---------- Regression Model has started ----------')
try:
for modelName in self.modelList:
if modelName not in self.params:
continue
paramSpace=self.params[modelName]
algoName = self.AlgorithmNames[modelName]
from incremental.riverML import riverML
riverMLObj = riverML()
self.log.info("-------> Model Name: "+str(modelName))
start = time.time()
model, modelParams, estimator, trainPredictedData = riverMLObj.startLearn('regression',algoName,paramSpace,self.trainX, self.trainY)
modelParams = str(modelParams)
executionTime=time.time() - start
self.log.info('---------> Total Execution: '+str(executionTime))
predictedData = riverMLObj.getPrediction(estimator,self.testX)
if 'neg_mean_squared_error' in self.scoreParam:
meanssquatederror = mean_squared_error(self.testY,predictedData)
score = meanssquatederror
elif 'neg_root_mean_squared_error' in self.scoreParam:
rootmeanssquatederror=mean_squared_error(self.testY,predictedData,squared=False)
score = rootmeanssquatederror
elif 'neg_mean_absolute_error' in self.scoreParam:
meanabsoluteerror=mean_absolute_error(self.testY,predictedData)
score = meanabsoluteerror
elif 'r2' in self.scoreParam:
r2score=r2_score(self.testY,predictedData)
score = round(r2score*100, 2)
if self.scoreParam == "r2":
if score>bestScore:
bestScore =score
bestModel =model
bestParams=modelParams
bestEstimator=estimator
self.bestTrainPredictedData = trainPredictedData
self.bestPredictedData = predictedData
else:
if abs(score) < bestScore or bestScore == -sys.float_info.max:
bestScore =abs(score)
bestModel =model
bestParams=modelParams
bestEstimator=estimator
self.bestTrainPredictedData = trainPredictedData
self.bestPredictedData = predictedData
metrices = {}
metrices["score"] = score
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","Score":'+str(abs(score))+'}'
self.log.info('Status:- |... ML Algorithm applied: '+modelName)
self.log.info("Status:- |... Testing Score: "+str(score))
self.log.info('---------- Regression Model End ---------- \n')
self.log.info('\n------- Best Model and its parameters -------------')
self.log.info('Status:- |... Best Algorithm selected: '+str(self.modelToAlgoNames[bestModel])+' Score='+str(round(bestScore,2)))
self.log.info("-------> Best Name: "+str(bestModel))
self.log.info("-------> Best Score: "+str(bestScore))
return self.modelToAlgoNames[bestModel],bestParams,bestScore,bestEstimator,scoredetails
except Exception as inst:
self.log.info( '\n-----> regressionModel failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno)) |
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
incClassificationModel.py |
'''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import time
import os
import sys
import numpy as np
from sklearn.metrics import confusion_matrix
from sklearn.metrics import recall_score
from sklearn.metrics import precision_score
from sklearn.preprocessing import binarize
from learner.optimizetechnique import OptimizationTq
from learner.parameters import parametersDefine
import logging
from learner.aion_matrix import aion_matrix
# apply threshold to positive probabilities to create labels
def to_labels(pos_probs, threshold):
return (pos_probs >= threshold).astype('int')
class incClassifierModel():
def __init__(self,noOfClasses,modelList,params,scoreParam,cvSplit,numIter,geneticParam,trainX,trainY,testX,testY,method,modelType,MakeFP0,MakeFN0,deployLocation):
self.noOfClasses = noOfClasses
self.modelList =modelList
self.params =params
self.trainX =trainX
self.X =trainX
self.trainY =trainY
self.testX = testX
self.testY = testY
self.method =method
self.scoreParam=scoreParam
self.cvSplit=cvSplit
self.numIter=numIter
self.geneticParam=geneticParam
self.MakeFP0= MakeFP0
self.MakeFN0=MakeFN0
self.log = logging.getLogger('eion')
self.modelType = modelType
self.deployLocation = deployLocation
self.isRiverModel = False
self.AlgorithmNames={'Online Logistic Regression':'Online Logistic Regression', 'Online Softmax Regression':'Online Softmax Regression', 'Online Decision Tree Classifier':'Online Decision Tree Classifier', 'Online KNN Classifier':'Online KNN Classifier'}
self.modelToAlgoNames = {value: key for key, value in self.AlgorithmNames.items()}
def check_threshold(self,estimator,testX,testY,threshold_range,checkParameter,modelName):
thresholdx = -1
for threshold in threshold_range:
predictedData = estimator.predict_proba(testX)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold)#bug 12437
p_score = precision_score(testY, predictedData)
r_score = recall_score(testY, predictedData)
tn, fp, fn, tp = confusion_matrix(testY, predictedData).ravel()
if(checkParameter.lower() == 'fp'):
if fp == 0:
if(p_score == 1):
thresholdx = threshold
self.log.info('---------------> Best Threshold:'+str(threshold))
self.log.info('---------------> Best Precision:'+str(p_score))
self.log.info('---------------> Best Recall:'+str(r_score))
self.log.info('---------------> TN:'+str(tn))
self.log.info('---------------> FP:'+str(fp))
self.log.info('---------------> FN:'+str(fn))
self.log.info('---------------> TP:'+str(tp))
break
if(checkParameter.lower() == 'fn'):
if fn == 0:
if(r_score == 1):
thresholdx = threshold
self.log.info('---------------> Best Threshold:'+str(threshold))
self.log.info('---------------> Best Precision:'+str(p_score))
self.log.info('---------------> Best Recall:'+str(r_score))
self.log.info('---------------> TN:'+str(tn))
self.log.info('---------------> FP:'+str(fp))
self.log.info('---------------> FN:'+str(fn))
self.log.info('---------------> TP:'+str(tp))
break
return(thresholdx,p_score,r_score)
def getBestModel(self,fp0,fn0,threshold,bestthreshold,rscore,brscore,pscore,bpscore,tscore,btscore):
cmodel = False
if(threshold != -1):
if(bestthreshold == -1):
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif fp0:
if rscore > brscore:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif rscore == brscore:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif fn0:
if pscore > bpscore:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
elif pscore == bpscore:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
bestthreshold = threshold
brscore = rscore
bpscore = pscore
btscore = tscore
else:
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
btscore = tscore
else:
if(bestthreshold == -1):
if tscore > btscore or btscore == -0xFFFF:
cmodel = True
btscore = tscore
return cmodel,btscore,bestthreshold,brscore,bpscore
def firstFit(self):
bestModel='None'
bestParams={}
bestScore=-0xFFFF
bestEstimator = 'None'
scoredetails = ''
threshold = -1
bestthreshold = -1
precisionscore =-1
bestprecisionscore=-1
recallscore = -1
bestrecallscore=-1
self.bestTrainPredictedData = None
self.bestPredictedData = None
self.log.info('\n---------- ClassifierModel has started ----------')
objClf = aion_matrix()
try:
for modelName in self.modelList:
paramSpace=self.params[modelName]
algoName = self.AlgorithmNames[modelName]
from incremental.riverML import riverML
riverMLObj = riverML()
self.log.info("-------> Model Name: "+str(modelName))
start = time.time()
model, modelParams, estimator, trainPredictedData = riverMLObj.startLearn('classification',algoName,paramSpace,self.trainX, self.trainY, self.noOfClasses)
modelParams = str(modelParams)
predictedData = riverMLObj.getPrediction(estimator,self.testX)
executionTime=time.time() - start
self.testY.reset_index(inplace=True, drop=True)
score = objClf.get_score(self.scoreParam,self.testY.values.flatten(),predictedData.values.flatten())
self.log.info(str(score))
metrices = {}
metrices["score"] = score
threshold = -1
precisionscore = precision_score(self.testY, predictedData, average='macro')
recallscore = recall_score(self.testY, predictedData, average='macro')
self.log.info('---------> Total Execution: '+str(executionTime))
if(scoredetails != ''):
scoredetails += ','
scoredetails += '{"Model":"'+self.modelToAlgoNames[model]+'","Score":'+str(score)+'}'
status,bscore,bthres,brscore,bpscore = self.getBestModel(self.MakeFP0,self.MakeFN0,threshold,bestthreshold,recallscore,bestrecallscore,precisionscore,bestprecisionscore,score,bestScore)
if status:
bestScore =bscore
bestModel =model
bestParams=modelParams
bestEstimator=estimator
bestthreshold = threshold
bestrecallscore = recallscore
bestprecisionscore = precisionscore
self.bestTrainPredictedData = trainPredictedData
self.bestPredictedData = predictedData
self.log.info('Status:- |... ML Algorithm applied: '+modelName)
self.log.info("Status:- |... Testing Score: "+str(score))
self.log.info('---------- ClassifierModel End ---------- \n')
self.log.info('\n------- Best Model and its parameters -------------')
self.log.info('Status:- |... Best Algorithm selected: '+str(self.modelToAlgoNames[bestModel])+' Score='+str(round(bestScore,2)))
self.log.info("-------> Best Name: "+str(bestModel))
self.log.info("-------> Best Score: "+str(bestScore))
return self.modelToAlgoNames[bestModel],bestParams,bestScore,bestEstimator,scoredetails,bestthreshold,bestprecisionscore,bestrecallscore
except Exception as inst:
self.log.info( '\n-----> ClassifierModel failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
|
incProfiler.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
#System imports
import logging
import os
import sys
import pickle
#Sci-Tools imports
import numpy as np
import pandas as pd
from sklearn.preprocessing import LabelEncoder
from scipy import stats
from word2number import w2n
#river imports
from river.preprocessing import StatImputer
from river import stats, compose, anomaly
class incProfiler():
def __init__(self):
self.DtypesDic={}
self.pandasNumericDtypes=['int16', 'int32', 'int64', 'float16', 'float32', 'float64']
self.allNumberTypeCols = [] #all number type columns
self.allNumCols = [] #only numerical columns which includes num features and target if it is numerical
self.allCatCols = []
self.numFtrs = []
self.catFtrs = []
self.textFtrs = []
self.textVectorFtrs = []
self.numDiscreteCols = []
self.numContinuousCols = []
self.wordToNumericFeatures=[]
self.emptyCols=[]
self.missingCols = []
self.targetColumn = ""
self.le_dict = {}
self.configDict = {}
self.incFill = None
self.incLabelMapping = None
self.incCatEncoder = None
self.incScaler = None
self.incOutlierRem = None
self.log = logging.getLogger('eion')
def pickleDump(self, model, path):
if model is not None:
with open(path, 'wb') as f:
pickle.dump(model, f)
def saveProfilerModels(self, deployLocation):
if isinstance(self.incFill['num_fill'], StatImputer) or isinstance(self.incFill['cat_fill'], StatImputer):
self.pickleDump(self.incFill, os.path.join(deployLocation,'production','profiler','incFill.pkl'))
self.pickleDump(self.incLabelMapping, os.path.join(deployLocation,'production','profiler','incLabelMapping.pkl'))
self.pickleDump(self.incCatEncoder, os.path.join(deployLocation,'production','profiler','incCatEncoder.pkl'))
self.pickleDump(self.incScaler, os.path.join(deployLocation,'production','profiler','incScaler.pkl'))
self.pickleDump(self.incOutlierRem, os.path.join(deployLocation,'production','profiler','incOutlierRem.pkl'))
def featureAnalysis(self, df, conf_json, targetFeature):
try:
self.log.info('-------> Remove Duplicate Rows')
noofdplicaterows = df.duplicated(keep='first').sum()
df = df.drop_duplicates(keep="first")
df = df.reset_index(drop=True)
self.log.info('Status:- |... Duplicate row treatment done: '+str(noofdplicaterows))
self.log.info(df.head(5))
self.log.info( '\n----------- Inspecting Features -----------')
ctn_count = 0
df = df.replace('-', np.nan)
df = df.replace('?', np.nan)
dataFDtypes=self.dataFramecolType(df)
numerical_ratio = float(conf_json['numericFeatureRatio'])
categoricalMaxLabel = int(conf_json['categoryMaxLabel'])
indexFeatures = []
numOfRows = df.shape[0]
dataCols = df.columns
for item in dataFDtypes:
if(item[1] == 'object'):
filteredDf,checkFlag = self.smartFilter(item[0],df,numerical_ratio)
if(checkFlag):
self.wordToNumericFeatures.append(item[0])
self.log.info('----------> Data Type Converting to numeric :Yes')
try:
df[item[0]]=filteredDf[item[0]].astype(float)
except:
pass
ctn_count = ctn_count+1
else:
count = (df[item[0]] - df[item[0]].shift() == 1).sum()
if((numOfRows - count) == 1):
self.log.info( '-------> Feature :'+str(item[0]))
self.log.info('----------> Sequence Feature')
indexFeatures.append(item[0])
self.configDict['wordToNumCols'] = self.wordToNumericFeatures
self.configDict['emptyFtrs'] = indexFeatures
self.log.info('Status:- |... Feature inspection done for numeric data: '+str(ctn_count)+' feature(s) converted to numeric')
self.log.info('Status:- |... Feature word to numeric treatment done: '+str(self.wordToNumericFeatures))
self.log.info( '----------- Inspecting Features End -----------\n')
except Exception as inst:
self.log.info("Error in Feature inspection: "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
try:
self.log.info('\n---------- Dropping Index features ----------')
self.log.info('Index Features to remove '+str(indexFeatures))
if len(indexFeatures) > 0:
dataCols = list(set(dataCols) - set(indexFeatures))
for empCol in indexFeatures:
self.log.info('-------> Drop Feature: '+empCol)
df = df.drop(columns=[empCol])
self.log.info('---------- Dropping Index features End----------\n')
dataFDtypes=self.dataFramecolType(df)
categoricalMaxLabel = int(conf_json['categoryMaxLabel'])
for item in dataFDtypes:
self.DtypesDic[item[0]] = item[1]
nUnique=len(df[item[0]].unique().tolist())
if item[1] in self.pandasNumericDtypes:
self.allNumberTypeCols.append(item[0])
if nUnique >= categoricalMaxLabel:
self.allNumCols.append(item[0]) #pure numerical
if item[1] in ['int16', 'int32', 'int64']:
self.numDiscreteCols.append(item[0])
elif item[1] in ['float16', 'float32', 'float64']:
self.numContinuousCols.append(item[0])
else:
self.allCatCols.append(item[0])
elif item[1] != 'bool':
if (nUnique >= categoricalMaxLabel) and targetFeature != item[0]:
self.textFtrs.append(item[0])
else:
col = item[0]
if (max(df[col].astype(str).str.split().str.len()) > 10) and targetFeature != item[0]:
self.textFtrs.append(item[0])
else:
self.allCatCols.append(item[0])
else:
self.allCatCols.append(item[0])
misval_ratio = float(conf_json['misValueRatio'])
self.configDict['misval_ratio'] = misval_ratio
missingCols, emptyCols = self.getMissingVals(df, dataCols, misval_ratio)
if targetFeature in emptyCols:
raise Exception('Target column '+str(targetFeature)+' cannot be empty')
dataCols = list(set(dataCols) - set(emptyCols))
self.log.info('\n---------- Dropping empty features ----------')
for empCol in emptyCols:
self.log.info('-------> Drop Feature: '+empCol)
df = df.drop(columns=[empCol])
self.log.info('---------- Dropping empty features End----------\n')
self.log.info('Status:- |... Empty feature treatment done: '+str(len(emptyCols))+' empty feature(s) found')
self.log.info('-------> Data Frame Shape After Dropping (Rows,Columns): '+str(df.shape))
self.allNumCols = list(set(self.allNumCols) - set(emptyCols))
self.allCatCols = list(set(self.allCatCols) - set(emptyCols))
self.textFtrs = list(set(self.textFtrs) - set(emptyCols))
missingValFtrs = list(set(missingCols) - set(emptyCols))
self.log.info(str(len(missingValFtrs))+' feature(s) found with missing value(s)')
self.log.info('\n-------> Numerical continuous columns :'+(str(self.numContinuousCols))[:500])
self.log.info('-------> Numerical discrete columns :'+(str(self.numDiscreteCols))[:500])
self.log.info('-------> Non numerical columns :'+(str(self.allCatCols))[:500])
self.log.info('-------> Text columns :'+(str(self.textFtrs))[:500])
except Exception as inst:
self.log.info("Error in segregating numerical and categorical columns: "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return df, missingValFtrs, emptyCols, dataCols, self.allNumCols, self.allCatCols, self.textFtrs
def createIncProfiler(self, df, conf_json, allNumCols, numFtrs, allCatCols, textFtrs, missingValFtrs):
self.incLabelMapping = None
catFtrs = allCatCols.copy()
#LabelEncoding
if self.targetColumn in allCatCols:
catFtrs.remove(self.targetColumn)
self.incLabelMapping = LabelEncoder()
df[self.targetColumn] = df[self.targetColumn].apply(str)
self.incLabelMapping.fit(df[self.targetColumn])
self.le_dict = dict(zip(self.incLabelMapping.classes_, self.incLabelMapping.transform(self.incLabelMapping.classes_)))
self.log.info('----------> Encoded Values of Target Labels: '+(str(self.le_dict))[:500])
#self.incFill --> {num_fill:SI/0.0/'drop', cat_fill:SI/0.0/'drop'}
#fill
self.incFill = {}
self.incCatEncoder = None
self.incScaler = None
self.incOutlierRem = None
num_fill_method = 'Mean'
for x in list(conf_json['numericalFillMethod'].keys()):
if conf_json['numericalFillMethod'][x] == 'True':
num_fill_method = x
break
if num_fill_method.lower() =='mean':
num_fill = [(col, stats.Mean()) for col in allNumCols]
self.incFill['num_fill'] = StatImputer(*num_fill)
elif num_fill_method.lower() =='min':
num_fill = [(col, stats.Min()) for col in allNumCols]
self.incFill['num_fill'] = StatImputer(*num_fill)
elif num_fill_method.lower() == 'max':
num_fill = [(col, stats.Max()) for col in allNumCols]
self.incFill['num_fill'] = StatImputer(*num_fill)
elif num_fill_method.lower() =='zero':
self.incFill['num_fill'] = 'zero'
elif num_fill_method.lower() =='drop':
self.incFill['num_fill'] = 'drop'
else:
num_fill = [(col, stats.Mean()) for col in allNumCols]
self.incFill['num_fill'] = StatImputer(*num_fill)
cat_fill_method = 'Mode'
for x in list(conf_json['categoricalFillMethod'].keys()):
if conf_json['categoricalFillMethod'][x] == 'True':
cat_fill_method = x
break
if cat_fill_method.lower() =='zero':
self.incFill['cat_fill'] = 'zero'
elif cat_fill_method.lower() == 'mode':
cat_fill = [(col, stats.Mode()) for col in allCatCols]
self.incFill['cat_fill'] = StatImputer(*cat_fill)
elif cat_fill_method.lower() =='drop':
self.incFill['cat_fill'] = 'drop'
#CatEncoding
for x in list(conf_json['categoryEncoding'].keys()):
if conf_json['categoryEncoding'][x] == 'True':
catEncoder = x
break
catEncHow = 'Mean'
for x in list(conf_json['targetEncodingParams']['how'].keys()):
if conf_json['targetEncodingParams']['how'][x] == 'True':
catEncHow = x
break
if self.targetColumn in catFtrs:
catFtrs.remove(self.targetColumn)
if len(catFtrs) > 0:
from river.feature_extraction import TargetAgg
if catEncHow.lower() == 'mean':
agg_stat = stats.Mean()
if catEncHow.lower() == 'bayesianmean' or catEncHow.lower() == 'bayesian mean':
agg_stat = stats.BayesianMean(prior=0.5, prior_weight=50)
self.incCatEncoder = TargetAgg(
by=catFtrs[0], how=agg_stat)
for col in catFtrs[1:]:
self.incCatEncoder += TargetAgg(
by=col, how=agg_stat)
self.incCatEncoder|= compose.Discard(*catFtrs)
#Scaling
normalization_status = 'False'
normalization_method = ""
if 'normalization' in conf_json:
nor_supported_methods = conf_json['normalization']
for k in nor_supported_methods.keys():
if conf_json['normalization'][k].lower() == 'true':
normalization_status='True'
normalization_method =k
break
if normalization_status.lower() == "true" and len(numFtrs) > 0:
from sklearn.preprocessing import MinMaxScaler, StandardScaler, MaxAbsScaler
if self.targetColumn in numFtrs:
numFtrs.remove(self.targetColumn)
if normalization_method.lower() =='standardscaler':
self.incScaler = StandardScaler()
elif normalization_method.lower() =='minmaxscaler' or normalization_method.lower() =='minmax':
self.incScaler = MinMaxScaler()
elif normalization_method.lower() =='maxabsscaler' or normalization_method.lower() =='maxabs':
self.incScaler = MaxAbsScaler()
else:
self.incScaler = None
#OutlierRemoval
outlier_status = 'False'
outlier_method = 'None'
for x in list(conf_json['outlierDetection'].keys()):
if conf_json['outlierDetection'][x] == 'True':
outlier_method = x
outlier_status = 'True'
break
if outlier_status and numFtrs:
outlierMethodNames = list(conf_json['outlierDetectionParams'].keys())
if outlier_method.lower() == 'oneclasssvm' or outlier_method.lower() == 'one class svm':
for x in outlierMethodNames:
if x[0].lower() == 'o':
key = x
break
params = conf_json['outlierDetectionParams'][key]
self.log.info('<--- one class SVM with quantile filter --->')
self.incOutlierRem = anomaly.QuantileFilter(anomaly.OneClassSVM(nu=float(params['nu'])),q=float(params['q']))
elif outlier_method.lower() =='halfspacetrees' or outlier_method.lower() =='half space trees':
for x in outlierMethodNames:
if x[0].lower() == 'h':
key = x
break
params = conf_json['outlierDetectionParams'][key]
self.log.info('<--- Half space trees with quantile filter --->')
self.incOutlierRem = anomaly.QuantileFilter(anomaly.HalfSpaceTrees(n_trees=int(params['n_trees']),height=int(params['height']), window_size=int(params['window_size'])) ,q=float(params['q']))
else:
self.log.info("No method is provided for outlier analysis")
def getMissingVals(self,dataframe,columns,misval_ratio):
try:
self.log.info( '\n----------- Detecting for Missing Values -----------')
nonNAArray=[]
numOfRows = dataframe.shape[0]
for i in columns:
numNa=dataframe.loc[(pd.isna(dataframe[i])),i ].shape[0]
nonNAArray.append(tuple([i,numNa]))
for item in nonNAArray:
numofMissingVals = item[1]
if(numofMissingVals !=0):
self.log.info('-------> Feature '+str(item[0]))
self.log.info('----------> Number of Empty Rows '+str(numofMissingVals))
self.missingCols.append(item[0])
if(numofMissingVals >= numOfRows * misval_ratio):
self.log.info('----------> Empty: Yes')
self.log.info('----------> Permitted Rows: '+str(int(numOfRows * misval_ratio)))
self.emptyCols.append(item[0])
if(len(self.missingCols) !=0):
self.log.info( '----------- Detecting for Missing Values End -----------\n')
return self.missingCols, self.emptyCols
else:
self.log.info( '-------> Missing Value Features :Not Any')
self.log.info( '----------- Detecting for Missing Values End -----------\n')
return self.missingCols, self.emptyCols
except Exception as e:
self.log.info("getMissingVals failed ==>" +str(e))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1]
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
return None, None
def startIncProfiler(self,df,conf_json,targetFeature,deployLocation,problemType):
try:
self.targetColumn = targetFeature
df, missingValFtrs, emptyFtrs, dataFtrs, allNumCols, allCatCols, textFtrs = self.featureAnalysis(df, conf_json, self.targetColumn)
if len(textFtrs)>0:
self.log.info('Text Features are not supported. Dropping '+str(textFtrs)[:500])
df = df.drop(columns=textFtrs)
catFtrs = allCatCols.copy()
numFtrs = allNumCols.copy()
if self.targetColumn in catFtrs:
catFtrs.remove(self.targetColumn)
if targetFeature in allNumCols:
numFtrs.remove(targetFeature)
self.configDict['targetCol'] = self.targetColumn
self.configDict['numFtrs'] = numFtrs
self.configDict['catFtrs'] = catFtrs
self.configDict['allNumCols'] = allNumCols
self.configDict['allCatCols'] = allCatCols
self.configDict['allFtrs'] = numFtrs+catFtrs
try:
self.log.info('\n---------- Creating Incremental profiler models ----------')
self.createIncProfiler(df, conf_json, allNumCols, numFtrs, allCatCols, textFtrs, missingValFtrs)
self.log.info('\n--------- Incremental profiler models have been created ---------')
except Exception as inst:
self.log.info("Error in creating Incremental profiler models"+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
raise
try:
#mvt
# if missingValFtrs:
if self.incFill['num_fill'] == 'drop':
df = df.dropna(axis = 0, subset=allNumCols)
self.configDict['num_fill'] = 'drop'
elif self.incFill['num_fill'] == 'zero':
df[allNumCols] = df[allNumCols].fillna(value = 0.0)
self.configDict['num_fill'] = 'zero'
else:
df = df.astype(object).where(df.notna(), None)
df[allNumCols]= df[allNumCols].apply(lambda row: self.apply_river_model(row.to_dict(), self.incFill
['num_fill']), axis='columns')
self.configDict['num_fill'] = {col:self.incFill['num_fill'].stats[col].get() for col in allNumCols}
if self.incFill['cat_fill'] == 'drop':
df = df.dropna(axis = 0, subset=allCatCols)
self.configDict['cat_fill'] = 'drop'
elif self.incFill['cat_fill'] == 'zero':
df[allCatCols] = df[allCatCols].fillna(value = 0.0)
self.configDict['cat_fill'] = 'zero'
else:
df = df.astype(object).where(df.notna(), None)
df[allCatCols]= df[allCatCols].apply(lambda row: self.apply_river_model(row.to_dict(), self.incFill['cat_fill']), axis='columns')
self.configDict['cat_fill'] = {col:self.incFill['cat_fill'].stats[col].get() for col in allCatCols}
self.log.info('\nStatus:- |... Missing value treatment done')
except Exception as inst:
self.log.info("Error in Missing value treatment "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
raise
try:
#labelenc
if self.incLabelMapping:
df[targetFeature] = self.incLabelMapping.transform(df[targetFeature])
# self.configDict['labelMapping'] = self.le_dict
except Exception as inst:
self.log.info("Error in Label mapping "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
raise
try:
#catenc
if self.incCatEncoder:
self.log.info('\n--------- Converting Non Numerical Categorical Features to Numerical Features ---------')
self.encTarget = targetFeature
if problemType.lower() == 'regression':
from sklearn.preprocessing import StandardScaler
sc = StandardScaler()
self.encTarget = 'scaledTarget'
df['scaledTarget'] = sc.fit_transform(df[targetFeature].to_numpy().reshape(-1,1))
encCols = catFtrs.copy()
encCols.append(self.encTarget)
self.configDict['encCols'] = encCols
self.configDict['encTarget'] = self.encTarget
transformed_data = df[encCols].apply(lambda row: self.apply_enc(row.to_dict()), axis='columns')
if targetFeature in transformed_data.columns:
transformed_data.drop(targetFeature, inplace=True, axis = 1)
if problemType.lower() == 'regression':
df.drop('scaledTarget', inplace=True, axis = 1)
df[catFtrs] = transformed_data
# self.log.info('Status:- |... Target Encoding state is as follows: ')
self.configDict['catEnc'] = []
if len(catFtrs) == 1:
col = catFtrs[0]
self.configDict['catEnc'].append({col:self.incCatEncoder['TargetAgg'].state.to_dict()})
else:
for i, col in enumerate(catFtrs):
if i==0:
no = ''
else:
no = str(i)
self.configDict['catEnc'].append({col:self.incCatEncoder['TransformerUnion']['TargetAgg'+no].state.to_dict()})
# print(self.incCatEncoder['TransformerUnion']['TargetAgg'].state)
# self.log.info(self.incCatEncoder)
self.log.info('Status:- |... Categorical to numeric feature conversion done: '+str(len(catFtrs))+' features converted')
except Exception as inst:
self.log.info("Error in categorical encoding "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
raise
try:
#scaler
if self.incScaler:
self.log.info("\n---------- Data Normalization has started ----------")
self.incScaler = self.incScaler.partial_fit(df[numFtrs])
df[numFtrs] = self.incScaler.transform(df[numFtrs])
self.log.info( "---------- Normalization Done on Following features ----------")
self.log.info(numFtrs)
self.log.info('Status:- |... Normalization treatment done')
except Exception as inst:
self.log.info("Error in normalization "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
raise
try:
#outlierrem
if self.incOutlierRem:
self.log.info('\n---------- Performing outlier analysis ----------')
df = df[df[numFtrs].apply(lambda x: False if self.apply_od_pipe(x.to_dict()) else True, axis=1)]
self.log.info('\n <--- dataframe after outlier analysis --->')
df.reset_index(drop=True, inplace=True)
self.log.info(df.head(5))
self.log.info('Status:- |... Outlier treatment done')
self.log.info('\n <--- shape of dataframe after outlier analysis --->')
self.log.info(df.shape)
except Exception as inst:
self.log.info("Error in outlier treatment "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
raise
#saveprofiler
self.log.info('\n---------- Saving profiler models ----------')
self.saveProfilerModels(deployLocation)
self.log.info('<--- Profiler models saved at '+deployLocation+' --->')
return df,targetFeature,missingValFtrs,numFtrs,catFtrs,self.le_dict,self.configDict,textFtrs,emptyFtrs,self.wordToNumericFeatures
except Exception as inst:
self.log.info("Error: dataProfiler failed "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def transformData(self, df, targetFeature, missingValFtrs,numFtrs, catFtrs, textFtrs):
try:
df = df.drop_duplicates(keep="first")
df = df.reset_index(drop=True)
df = df.replace('-', np.nan)
df = df.replace('?', np.nan)
text_mv_cols = list(set(missingValFtrs).intersection(set(textFtrs)))
if len(text_mv_cols)>0:
df[text_mv_cols] = df[text_mv_cols].fillna(value = 'NA')
if 'num_fill' in self.configDict:
if self.configDict['num_fill'] == 'drop':
df = df.dropna(axis = 0, subset=self.allNumCols)
elif self.configDict['num_fill'] == 'zero':
df[self.allNumCols] = df[self.allNumCols].fillna(value = 0.0)
else:
for x in self.allNumCols:
df[x] = df[x].fillna(value = self.configDict['num_fill'][x])
if 'cat_fill' in self.configDict:
if self.configDict['cat_fill'] == 'drop':
df = df.dropna(axis = 0, subset=self.allCatCols)
elif self.configDict['cat_fill'] == 'zero':
df[self.allCatCols] = df[self.allCatCols].fillna(value = 0.0)
else:
for x in self.allCatCols:
df[x] = df[x].fillna(value = self.configDict['cat_fill'][x])
if self.incLabelMapping:
df[targetFeature] = self.incLabelMapping.transform(df[targetFeature])
if self.incCatEncoder:
transformed_data = df[catFtrs].apply(lambda row: self.apply_enc(row.to_dict(), isTrain=False), axis='columns')
df[catFtrs] = transformed_data
if self.incScaler:
df[numFtrs] = self.incScaler.transform(df[numFtrs])
return df
except Exception as inst:
self.log.info("Error: DataProfiling transformation failed "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def checknumStr(self,dataframe,feature):
try:
dataframe[feature]=dataframe[feature].apply(lambda x: self.testStr(x))
return dataframe
except:
self.log.info("checknumStr failed")
return dataframe
#test whether the value is numeric /string
def testStr(self,value):
try:
x=eval(value)
return np.nan
except:
return value
"""
Missing values analysis
Detects number of missing values in each column of dataframe
"""
def checksRows(self,dataframe,target_column,dataColumns):
self.log.info( '\n----------- Checking Target Feature Empty Rows -----------')
if self.targetColumn != '':
numNa=dataframe.loc[(pd.isna(dataframe[self.targetColumn])),self.targetColumn].shape[0]
self.log.info('------->No of Empty Rows in Target Fields: '+str(numNa))
if numNa >0:
self.log.info('-------> Remove Empty Target Field Rows')
dataframe = dataframe.dropna(axis=0, subset=[self.targetColumn])
self.log.info('-------> Remove Duplicate Rows')
dataframe = dataframe.dropna(axis=0,how='all',subset=dataColumns)
noofdplicaterows = dataframe.duplicated(keep='first').sum()
dataframe = dataframe.drop_duplicates(keep="first")
dataframe = dataframe.reset_index(drop=True)
return dataframe,noofdplicaterows
def apply_river_model(self, x, profModel):
profModel.learn_one(x)
return pd.Series(profModel.transform_one(x))
def apply_enc(self, x, isTrain=True):
if isTrain:
y = x[self.encTarget]
self.incCatEncoder.learn_one(x, y)
return pd.Series(self.incCatEncoder.transform_one(x))
def apply_od_pipe(self, x):
score = self.incOutlierRem.score_one(x)
is_anomaly = self.incOutlierRem.classify(score)
self.incOutlierRem.learn_one(x)
return is_anomaly
#Convert Words To Number
def s2n(self,value):
try:
x=eval(value)
return x
except:
try:
return w2n.word_to_num(value)
except:
return np.nan
def convertWordToNumeric(self,dataframe,feature):
try:
dataframe[feature]=dataframe[feature].apply(lambda x: self.s2n(x))
return dataframe
except Exception as inst:
self.log.info("convertWordToNumeric Failed ===>"+str(inst))
return dataframe
#test whether the value is numeric /string
def testNum(self,value):
try:
x=eval(value)
return x
except:
return np.nan
##check for numeric values in string column
def checkNumeric(self,dataframe,feature):
try:
dataframe[feature]=dataframe[feature].apply(lambda x: self.testNum(x))
return dataframe
except Exception as inst:
self.log.info("checkNumeric Failed ===>"+str(inst))
return dataframe
def smartFilter(self,feature,df,numericRatio):
try:
distinctCount = len(df[feature].unique())
numOfRows = df.shape[0]
tempDataFrame=df.copy(deep=True)
if(distinctCount != 1):
self.log.info('-------> Feature :'+str(feature))
testDf = self.checkNumeric(tempDataFrame,feature)
tempDf = testDf[feature]
tempDf = tempDf.dropna()
numberOfNonNullVals = tempDf.count()
if(numberOfNonNullVals > int(numOfRows * numericRatio)):
tempDataFrame=df.copy(deep=True)
testDf = self.convertWordToNumeric(tempDataFrame,feature)
tempDf = testDf[feature]
tempDf = tempDf.dropna()
self.log.info('----------> Numeric Status :Yes')
return testDf,True
else:
#Wasnt't a numerical feature
self.log.info('----------> Numeric Status :No')
#numDf = self.checknumStr(df,feature)
return df,False
else:
self.log.info( '\n---> No Numerics found in :' +str(feature))
return df,False
except:
self.log.info( '\n---> No Numerics found in :'+str(feature))
return df,False
def dataFramecolType(self,dataFrame):
dataFDtypes=[]
try:
dataColumns=list(dataFrame.columns)
for i in dataColumns:
dataType=dataFrame[i].dtypes
dataFDtypes.append(tuple([i,str(dataType)]))
return dataFDtypes
except:
self.log.info("error in dataFramecolyType")
return dataFDtypes
|
riverML.py |
import logging
import pickle
import os
import sys
import pandas as pd
from river import stream
from river.linear_model import LogisticRegression, SoftmaxRegression, LinearRegression
from river.tree import ExtremelyFastDecisionTreeClassifier, HoeffdingAdaptiveTreeRegressor
# from river.ensemble import AdaptiveRandomForestRegressor, AdaptiveRandomForestClassifier
from river.neighbors import KNNClassifier, KNNRegressor
from river.multiclass import OneVsRestClassifier
from river.optim import SGD, Adam, AdaDelta, NesterovMomentum, RMSProp
# from river.optim.losses import CrossEntropy, Log, MultiClassLoss, Poisson, RegressionLoss, BinaryLoss, Huber
# from river.optim.initializers import Normal
class riverML(object):
def __init__(self):
self.algoDict={'Online Logistic Regression':LogisticRegression, 'Online Softmax Regression':SoftmaxRegression, 'Online Decision Tree Classifier':ExtremelyFastDecisionTreeClassifier, 'Online KNN Classifier':KNNClassifier,'Online Linear Regression':LinearRegression, 'Online Decision Tree Regressor':HoeffdingAdaptiveTreeRegressor, 'Online KNN Regressor':KNNRegressor}
self.optDict={'sgd': SGD, 'adam':Adam, 'adadelta':AdaDelta, 'nesterovmomentum':NesterovMomentum, 'rmsprop':RMSProp}
self.log = logging.getLogger('eion')
def getPrediction(self, model,X):
testStream = stream.iter_pandas(X)
preds = []
for (xi,yi) in testStream:
pred = model.predict_one(xi)
preds.append(pred)
return pd.DataFrame(preds)
def startLearn(self,problemType,algoName,params,xtrain,ytrain,noOfClasses=None):
try:
model = self.algoDict[algoName]
params = self.parseParams(params, algoName)
if problemType == 'classification':
if noOfClasses>2:
model = OneVsRestClassifier(classifier=model(**params))
else:
model = model(**params)
else:
model = model(**params)
trainStream = stream.iter_pandas(xtrain, ytrain)
#head start
for i, (xi, yi) in enumerate(trainStream):
if i>100:
break
if yi!=None:
model.learn_one(xi, yi)
trainPredictedData = []
trainStream = stream.iter_pandas(xtrain, ytrain)
for i, (xi, yi) in enumerate(trainStream):
if yi!=None:
trainPredictedData.append(model.predict_one(xi))
model.learn_one(xi, yi)
trainPredictedData = pd.DataFrame(trainPredictedData)
return algoName, params, model, trainPredictedData
except Exception as inst:
self.log.info( '\n-----> '+algoName+' failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def parseParams(self, params, algoName):
try:
from learner.parameters import parametersDefine
paramsObj = parametersDefine()
paramDict =paramsObj.paramDefine(params,method=None)
paramDict = {k:v[0] for k,v in paramDict.items()}
if algoName=='Online Logistic Regression' or algoName=='Online Softmax Regression' or algoName=='Online Linear Regression':
opt = self.optDict[paramDict.pop('optimizer').lower()]
lr = float(paramDict.pop('optimizer_lr'))
paramDict['optimizer'] = opt(lr)
return paramDict
except Exception as inst:
self.log.info( '\n-----> Parameter parsing failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
ensemble_bagging.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import time
import os
import sys
import numpy as np
import pandas as pd
from sklearn import model_selection
from sklearn.model_selection import train_test_split, KFold, cross_val_score
from sklearn.model_selection import KFold
#Classification metrics lib
import logging
import warnings
warnings.filterwarnings('always') # "error", "ignore", "always", "default", "module" or "once"
from learner.aion_matrix import aion_matrix
from sklearn.preprocessing import binarize
class ensemble_bagging():
def __init__(self,ensemble_params,scoreParam,MakeFP0,MakeFN0):
self.ensemble_params = ensemble_params
self.scoreParam=scoreParam
self.MakeFP0 = MakeFP0
self.MakeFN0 = MakeFN0
self.log = logging.getLogger('eion')
def add_alg2dict(self,k,v):
b_dict={}
b_dict[k]=v
return b_dict
def getSelected_algs_params(self,problemType,ensembleType,ensembleConfig):
from learner.parameters import parametersDefine
paramObj=parametersDefine()
ensClass_algs_params={}
algs_status={}
for key,val in ensembleConfig.items():
for s,p in val.items():
if (s == "enable" and p == "True"):
params = val['param']
params_eval = paramObj.paramDefine(params,None)
params_eval = {param_key: param_value[0] for param_key, param_value in params_eval.items()}
ensClass_algs_params[key]=params_eval
else:
pass
return ensClass_algs_params
def listEnsembleClassBaggingAlgs(self,ensClass_algs_params):
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.naive_bayes import GaussianNB
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import BaggingClassifier, ExtraTreesClassifier, RandomForestClassifier
ensembleBaggingClassList=list()
for key,val in ensClass_algs_params.items():
if (key == 'Logistic Regression'):
lr=LogisticRegression()
lr=lr.set_params(**val)
ensembleBaggingClassList.append(lr)
elif (key == 'Support Vector Machine'):
svm=SVC()
svm=svm.set_params(**val)
ensembleBaggingClassList.append(svm)
elif (key == 'Naive Bayes'):
nb=GaussianNB()
nb=nb.set_params(**val)
ensembleBaggingClassList.append(nb)
elif (key == 'K Nearest Neighbors'):
knn=KNeighborsClassifier()
knn=knn.set_params(**val)
ensembleBaggingClassList.append(knn)
elif (key == 'Decision Tree'):
dt=DecisionTreeClassifier()
dt=dt.set_params(**val)
ensembleBaggingClassList.append(dt)
elif (key == 'Random Forest'):
rf=RandomForestClassifier()
rf=rf.set_params(**val)
ensembleBaggingClassList.append(rf)
else:
pass
return ensembleBaggingClassList
def listEnsembleRegBaggingAlgs(self,ensReg_algs_params):
from sklearn.linear_model import Ridge
from sklearn.linear_model import LinearRegression
from sklearn.ensemble import RandomForestRegressor
from sklearn.tree import DecisionTreeRegressor
ensembleBaggingRegList=list()
for key,val in ensReg_algs_params.items():
if (key == 'Linear Regression'):
lir=LinearRegression()
lir=lir.set_params(**val)
ensembleBaggingRegList.append(lir)
elif (key == 'Decision Tree'):
dtr=DecisionTreeRegressor()
dtr=dtr.set_params(**val)
ensembleBaggingRegList.append(dtr)
elif (key == 'Ridge'):
ridge=Ridge()
ridge=ridge.set_params(**val)
ensembleBaggingRegList.append(ridge)
else:
ensembleBaggingRegList=[]
return ensembleBaggingRegList
def ensemble_bagging_classifier(self,X_train,y_train, X_test, y_test):
## New changes
from sklearn.ensemble import BaggingClassifier, ExtraTreesClassifier, RandomForestClassifier
ensemble_method = "Bagging_classifier"
problemType='classification'
ensembleType='bagging'
model_dict=self.ensemble_params
ensClass_algs_params = self.getSelected_algs_params(problemType,ensembleType,model_dict)
ensembleBaggingList = self.listEnsembleClassBaggingAlgs(ensClass_algs_params)
# clf_array = model_list
clf_array=ensembleBaggingList
# no. of base classifier
num_trees = len(clf_array)
# max_samples=float(max_samples)
n_estimators = num_trees
# random_state=seed
bagging_mean={}
bagging_std={}
accuracy_basealgs_train={}
accuracy_basealgs_test={}
blable=""
accuracy_score_test=0
kfold = model_selection.KFold(n_splits=10, random_state=None)
bestScore=-0xFFFF
scoredetails = ''
threshold = -1
bestthreshold = -1
precisionscore =-1
bestprecisionscore=-1
recallscore = -1
bestrecallscore=-1
objClf = aion_matrix()
if (ensemble_method == "Bagging_classifier"):
#bagging ensemble of base classifier .e.g. KNeighborsClassifier base estimators, each built on random subsets of 40% of the samples and 50% of the features.
for clf in clf_array:
self.log.info('-----------> Ensemble Algorithm '+str(clf.__class__.__name__))
clf.fit(X_train, y_train)
bagging_clf = BaggingClassifier(clf,n_estimators = num_trees, random_state=10)
bagging_clf.fit(X_train, y_train)
bagging_scores = cross_val_score(bagging_clf, X_train, y_train, cv=kfold,n_jobs=-1)
#bagging_ensemble_t=bagging_clf.fit(X_train, y_train)
if not X_test.empty:
bag_predict=bagging_clf.predict(X_test)
accuracy_score_test = objClf.get_score(self.scoreParam,y_test,bag_predict)
else:
accuracy_score_test = bagging_scores
MakeFP0 = False
MakeFN0 = False
if self.MakeFP0:
self.log.info('-------- Ensemble: Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = objClf.check_threshold(bagging_clf,X_train,y_train,threshold_range,'FP','')
MakeFP0 = True
self.log.info('-------- Calculate Threshold for FP End-------')
if self.MakeFN0:
self.log.info('-------- Ensemble: Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = objClf.check_threshold(bagging_clf,X_train,y_train,threshold_range,'FN','')
MakeFN0 = True
self.log.info('-------- Calculate Threshold for FN End-------')
if threshold != -1:
if not X_test.empty:
predictedData = bagging_clf.predict_proba(X_test)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold) #bug 12437
accuracy_score_test = objClf.get_score(self.scoreParam,y_test,predictedData)
status,bscore,bthres,brscore,bpscore = objClf.getBestModel(MakeFP0,MakeFN0,threshold,bestthreshold,recallscore,bestrecallscore,precisionscore,bestprecisionscore,accuracy_score_test,bestScore)
if status:
bestScore =bscore
bestModel =bagging_clf.__class__.__name__
bestEstimator=bagging_clf
bestthreshold = bthres
bestBaseModel = clf.__class__.__name__
bestrecallscore = brscore
bestprecisionscore = bpscore
else:
pass
best_alg_name=bestEstimator.__class__.__name__
self.log.info('-----------> Best Bagging Classifier Model '+str(bestBaseModel))
self.log.info('-----------> Best Score '+str(bestScore))
# self.log.info('-----------> Threshold '+str(bestthreshold)) #bug 12438
if bestthreshold != -1:
if not X_test.empty:
predictedData_test = bestEstimator.predict_proba(X_test)
predictedData_test = binarize(predictedData_test[:,1].reshape(-1, 1),threshold=bestthreshold) #bug 12437
predictedData_train = bestEstimator.predict_proba(X_train)
predictedData_train = binarize(predictedData_train[:,1].reshape(-1, 1),threshold=bestthreshold) #bug 12437
else:
if not X_test.empty:
predictedData_test = bestEstimator.predict(X_test)
predictedData_train = bestEstimator.predict(X_train)
return bestEstimator,bestEstimator.get_params(),bestScore,best_alg_name,bestthreshold,bestprecisionscore,bestrecallscore
def ensemble_bagging__regressor(self,X_train,y_train, X_test, y_test):
from sklearn.ensemble import BaggingRegressor
ensemble_method='Bagging_regressor'
problemType='regression'
ensembleType='bagging'
model_dict=self.ensemble_params
ensReg_algs_params = self.getSelected_algs_params(problemType,ensembleType,model_dict)
ensembleBaggingList = self.listEnsembleRegBaggingAlgs(ensReg_algs_params)
scoredetails = ''
aion_matrixobj = aion_matrix()
reg_array = ensembleBaggingList
num_trees = len(reg_array)
#self.log.info(num_trees)
# max_samples=float(max_samples)
n_estimators = num_trees
r_state=10
bestModel=''
bestParams={}
bestScore=-sys.float_info.max #extension of bugfix 11656
objClf = aion_matrix()
for reg in reg_array:
self.log.info('-----------> Ensemble Algorithm '+str(reg.__class__.__name__))
nmodel=reg.fit(X_train, y_train)
model = reg.__class__.__name__
estimator = BaggingRegressor(base_estimator=reg, random_state=r_state)
bagging_ensemble_t=estimator.fit(X_train, y_train)
predictedData = estimator.predict(X_test)
score = objClf.get_score(self.scoreParam,y_test,predictedData)
if self.scoreParam == "r2":
if score > bestScore:
bestScore =score
bestModel =model
bestEstimator=estimator
else:
if abs(score) < bestScore or bestScore == -sys.float_info.max: #extension of bugfix 11656
bestScore =abs(score)
bestModel =model
bestEstimator=estimator
best_alg_name=bestEstimator.__class__.__name__
self.log.info('-----------> Best Ensemble Algorithm '+str(bestModel))
return bestEstimator,bestEstimator.get_params(),bestScore,best_alg_name
|
ensemble_voting.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import numpy as np
import logging
import warnings
from sklearn.ensemble import VotingClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.linear_model import Ridge
from sklearn.preprocessing import binarize
from sklearn.ensemble import VotingRegressor
from sklearn.svm import SVC
from sklearn.linear_model import LinearRegression
from sklearn.ensemble import RandomForestRegressor
from sklearn.tree import DecisionTreeRegressor
from learner.aion_matrix import aion_matrix
warnings.filterwarnings('always')
class ensemble_voting():
def __init__(self,ensemble_params,scoreParam):
self.ensemble_params = ensemble_params
self.scoreParam=scoreParam
self.final_estimator_r=''
self.final_estimator_c=''
self.log = logging.getLogger('eion')
''' Read the aion config "Ensemble-Voting", parse the algorithm and associated params based on enable or True status.Not used now '''
def getSelected_algs_params(self,problemType,ensembleType,ensembleConfig):
from learner.parameters import parametersDefine
paramObj=parametersDefine()
ensClass_algs_params={}
# algs_status={}
for key,val in ensembleConfig.items():
for s,p in val.items():
if (s == "enable" and p == "True"):
params = val['param']
params_eval = paramObj.paramDefine(params,None)
params_eval = {param_key: param_value[0] for param_key, param_value in params_eval.items()}
ensClass_algs_params[key]=params_eval
else:
pass
return ensClass_algs_params
''' To make array of voting algorithm based on user config list. Not used now, in future if needed similar line with bagging ensemble, please use this. '''
def listEnsembleClassVotingAlgs(self,ensClass_algs_params):
ensembleVotingClassList=list()
for key,val in ensClass_algs_params.items():
if (key == 'Logistic Regression'):
lr=LogisticRegression()
lr=lr.set_params(**val)
ensembleVotingClassList.append(lr)
elif (key == 'Support Vector Machine'):
svm=SVC()
svm=svm.set_params(**val)
ensembleVotingClassList.append(svm)
elif (key == 'Naive Bayes'):
nb=GaussianNB()
nb=nb.set_params(**val)
ensembleVotingClassList.append(nb)
elif (key == 'K Nearest Neighbors'):
knn=KNeighborsClassifier()
knn=knn.set_params(**val)
ensembleVotingClassList.append(knn)
elif (key == 'Decision Tree'):
dt=DecisionTreeClassifier()
dt=dt.set_params(**val)
ensembleVotingClassList.append(dt)
elif (key == 'Random Forest'):
rf=RandomForestClassifier()
rf=rf.set_params(**val)
ensembleVotingClassList.append(rf)
else:
## Algorithm not found in config, so forming empty alg list. If needs, make list with default alg.
ensembleVotingClassList=[]
pass
return ensembleVotingClassList
''' To make array of voting regression algorithm based on user config list. Not used now, in future if needed similar line with bagging ensemble, please use this. '''
def listEnsembleRegVotingAlgs(self,ensReg_algs_params):
ensembleVotingRegList=list()
for key,val in ensReg_algs_params.items():
if (key == 'Linear Regression'):
lir=LinearRegression()
lir=lir.set_params(**val)
ensembleVotingRegList.append(lir)
elif (key == 'Decision Tree'):
dtr=DecisionTreeRegressor()
dtr=dtr.set_params(**val)
ensembleVotingRegList.append(dtr)
elif (key == 'Ridge'):
ridge=Ridge()
ridge=ridge.set_params(**val)
ensembleVotingRegList.append(ridge)
else:
## Algorithm not found in config, so forming empty alg list. If needs, make list with default alg.
ensembleVotingRegList=[]
return ensembleVotingRegList
def ensemble_voting_classifier(self,X_train,y_train, X_test, y_test,MakeFP0,MakeFN0,modelList):
#bug 12437
status='ERROR'
model=None
estimator=None
score=None
params=None
threshold = -1
precisionscore =-1
recallscore = -1
objClf = aion_matrix()
try:
lr = LogisticRegression(solver='lbfgs',random_state=1,max_iter=200)
rf = RandomForestClassifier(random_state=1)
gnb = GaussianNB()
svc = SVC(probability=True) #Need to keep probability=True, because cross_val_score,predict_proba fn calls
knn=KNeighborsClassifier(n_neighbors=5)
base_estimators = []
if 'Logistic Regression' in modelList:
base_estimators.append(('LogisticRegression', lr))
self.log.info('-------- Ensemble: Logistic Regression-------')
if 'Random Forest' in modelList:
base_estimators.append(('RandomForestClassifier', rf))
self.log.info('-------- Ensemble: Random Forest-------')
if 'Naive Bayes' in modelList:
base_estimators.append(('GaussianNB', gnb))
self.log.info('-------- Ensemble: Naive Bayes-------')
if 'Support Vector Machine' in modelList:
self.log.info('-------- Ensemble: Support Vector Machine-------')
base_estimators.append(('SVC', svc))
if 'K Nearest Neighbors' in modelList:
base_estimators.append(('KNeighborsClassifier', knn))
self.log.info('-------- Ensemble: K Nearest Neighbors-------')
if len(base_estimators) == 0:
self.log.info('-------- Ensemble Voting is only supported for Logistic Regression, Random Forest Classifier, Naive Bayes, SVM and KNN -------')
status = "UNSUPPORTED"
return status, estimator,params,score,model,threshold,precisionscore,recallscore
eclf1 = VotingClassifier(base_estimators, voting='soft')
eclf1.fit(X_train, y_train)
y_predict = eclf1.predict(X_test)
score = objClf.get_score(self.scoreParam,y_test,y_predict)
self.log.info('-------- Ensemble (VoteClassifier) Soft Score:'+str(score))
if MakeFP0:
self.log.info('-------- Ensemble: Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = objClf.check_threshold(eclf1,X_train,y_train,threshold_range,'FP','')
self.log.info('-------- Calculate Threshold for FP End-------')
elif MakeFN0:
self.log.info('-------- Ensemble: Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = objClf.check_threshold(eclf1,X_train,y_train,threshold_range,'FN','')
self.log.info('-------- Calculate Threshold for FN End-------')
if threshold != -1:
predictedData = eclf1.predict_proba(X_test)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold) #bug 12437
score = objClf.get_score(self.scoreParam,y_test,predictedData)
status = 'SUCCESS'
model =eclf1.__class__.__name__
estimator=eclf1
params = estimator.get_params()
#bug 12437 - Removed ensemble hard voting as predict_proba in the later stages will break
except Exception as Inst: #bug 12437
self.log.info('--------- Error in Ensemble Voting ---------\n')
self.log.info(str(Inst))
return status,estimator,params,score,model,threshold,precisionscore,recallscore
def ensemble_voting__regressor(self,X_train,y_train, X_test, y_test,modelList):
scoredetails = ''
vr_predict=None
vr_model=None
try:
lr = LinearRegression()
rfr = RandomForestRegressor(n_estimators=10, random_state=1)
dtr=DecisionTreeRegressor()
base_estimators = []
if 'Linear Regression' in modelList:
base_estimators.append(('LinearRegression', lr))
if 'Decision Tree' in modelList:
base_estimators.append(('DecisionTreeRegressor', dtr))
if 'Random Forest' in modelList:
base_estimators.append(('RandomForestRegressor', rfr))
if len(base_estimators) == 0:
base_estimators = [('LinearRegression', lr), ('RandomForestRegressor', rfr),('DecisionTreeRegressor', dtr)]
voting_reg = VotingRegressor(base_estimators)
vr_model=voting_reg.fit(X_train,y_train)
vr_predict=voting_reg.predict(X_test)
best_vr_alg=voting_reg.__class__.__name__
self.log.info('-----------> Voting regression Model '+str(best_vr_alg))
except Exception as e:
self.log.info("voting regression Exception info: \n")
self.log.info(e)
aion_matrixobj = aion_matrix()
score = aion_matrixobj.get_score(self.scoreParam,y_test,vr_predict)
return voting_reg,voting_reg.get_params(),score,best_vr_alg
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
ensemble_stacking.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import numpy as np
#Classification metrics lib
import logging
import warnings
from sklearn.neighbors import KNeighborsClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.preprocessing import binarize
from sklearn.svm import SVC
from sklearn.ensemble import StackingClassifier
from sklearn.linear_model import LinearRegression
from sklearn.ensemble import RandomForestRegressor
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import StackingRegressor
from sklearn.svm import LinearSVR
from sklearn.linear_model import RidgeCV
from sklearn.linear_model import LassoCV
from learner.aion_matrix import aion_matrix
warnings.filterwarnings('always') # "error", "ignore", "always", "default", "module" or "once"
class ensemble_stacking():
def __init__(self,ensemble_params,scoreParam):
self.ensemble_params = ensemble_params
self.scoreParam=scoreParam
self.final_estimator_r=''
self.final_estimator_c=''
self.log = logging.getLogger('eion')
## Read the aion config "Ensemble-Stacking", parse the algorithm and associated params based on enable or True status.
def getSelected_algs_params(self,problemType,ensembleType,ensembleConfig):
from learner.parameters import parametersDefine
paramObj=parametersDefine()
ensClass_algs_params={}
# algs_status={}
for key,val in ensembleConfig.items():
for s,p in val.items():
if (s == "enable" and p == "True"):
params = val['param']
params_eval = paramObj.paramDefine(params,None)
params_eval = {param_key: param_value[0] for param_key, param_value in params_eval.items()}
ensClass_algs_params[key]=params_eval
else:
pass
return ensClass_algs_params
## To make array of stacking algorithm based on user config list. Not used now, in future if needed similar line with bagging ensemble, please use this.
def listEnsembleClassStackingAlgs(self,ensClass_algs_params):
ensembleBaggingClassList=list()
for key,val in ensClass_algs_params.items():
# print(key)
if (key == 'Logistic Regression'):
lr=LogisticRegression()
lr=lr.set_params(**val)
ensembleBaggingClassList.append(lr)
elif (key == 'Support Vector Machine'):
svm=SVC()
svm=svm.set_params(**val)
ensembleBaggingClassList.append(svm)
elif (key == 'Naive Bayes'):
nb=GaussianNB()
nb=nb.set_params(**val)
ensembleBaggingClassList.append(nb)
elif (key == 'K Nearest Neighbors'):
knn=KNeighborsClassifier()
knn=knn.set_params(**val)
ensembleBaggingClassList.append(knn)
elif (key == 'Decision Tree'):
dt=DecisionTreeClassifier()
dt=dt.set_params(**val)
ensembleBaggingClassList.append(dt)
elif (key == 'Random Forest'):
rf=RandomForestClassifier()
rf=rf.set_params(**val)
ensembleBaggingClassList.append(rf)
else:
ensembleBaggingClassList=[]
pass
return ensembleBaggingClassList
## To make array of stacking regression algorithm based on user config list. Not used now, in future if needed similar line with bagging ensemble, please use this.
def listEnsembleRegStackingAlgs(self,ensReg_algs_params):
ensembleBaggingRegList=list()
for key,val in ensReg_algs_params.items():
if (key == 'LinearSVR'):
lir=LinearSVR()
lir=lir.set_params(**val)
ensembleBaggingRegList.append(lir)
elif (key == 'LinearRegression'):
lr=LinearRegression()
lr=lr.set_params(**val)
ensembleBaggingRegList.append(lr)
elif (key == 'LassoCV'):
lcv=LassoCV()
lcv=lcv.set_params(**val)
ensembleBaggingRegList.append(lcv)
elif (key == 'RandomForestRegressor'):
rfr=RandomForestRegressor()
rfr=rfr.set_params(**val)
ensembleBaggingRegList.append(rfr)
elif (key == 'RidgeCV'):
ridge=RidgeCV()
ridge=ridge.set_params(**val)
ensembleBaggingRegList.append(ridge)
else:
## NO algorithms found in configuration settings, instead of sending empty array,we can add any one of algorithms.
ensembleBaggingRegList=[]
return ensembleBaggingRegList
def extract_params(self,dict):
self.dict=dict
for k,v in self.dict.items():
return k,v
def stacking_params(self):
for k,v in self.ensemble_params.items():
try:
if (k == "max_features_percentage"):
max_features_percentage=float(v)
elif (k == "max_samples"):
max_samples=float(v)
elif (k == "seed"):
seed=int(v)
elif (k == "final_estimator_stack_c"):
final_estimator_c=str(v)
elif (k == "final_estimator_stack_r"):
final_estimator_r=str(v)
else:
self.log.info("Invalid Param in ensemble advanced configuration.\n")
except Exception as e:
self.log.info("\n Ensemble config param parsing error"+str(e))
continue
return final_estimator_c,final_estimator_r,seed,max_samples,max_features_percentage
def ensemble_stacking_classifier(self,X_train,y_train, X_test, y_test,MakeFP0,MakeFN0,modelList):
final_estimator_c,final_estimator_r,seed,max_samples,max_features_percentage= self.stacking_params()
final_estimator_c=""
final_estimator=final_estimator_c
scoredetails=''
lr = LogisticRegression(solver='lbfgs',random_state=1,max_iter=200)
rf = RandomForestClassifier(random_state=2)
gnb = GaussianNB()
svc = SVC(probability=True) #Need to keep probability=True, because of cross_val_score,predict_proba fn calls
knn=KNeighborsClassifier(n_neighbors=5)
try:
if (final_estimator == 'LogisticRegression'):
final_estimator_a=lr
elif (final_estimator == 'RandomForestClassifier'):
final_estimator_a=rf
elif (final_estimator == 'GaussianNB'):
final_estimator_a=gnb
elif (final_estimator == 'SVC'):
final_estimator_a=svc
elif (final_estimator == 'KNeighborsClassifier'):
final_estimator_a=knn
else:
final_estimator_a=lr
except Exception as e:
final_estimator_a=lr
self.log.info("Given stacking regression final estimator algorithm issue, using default one (LogisticRegression) as final_estimator now.\n")
self.log.info(e)
#stacking estimators
base_estimators = []
if 'Logistic Regression' in modelList:
base_estimators.append(('LogisticRegression', lr))
if 'Random Forest' in modelList:
base_estimators.append(('RandomForestClassifier', rf))
if 'Naive Bayes' in modelList:
base_estimators.append(('GaussianNB', gnb))
if 'Support Vector Machine' in modelList:
base_estimators.append(('SVC', svc))
if 'K Nearest Neighbors' in modelList:
base_estimators.append(('KNeighborsClassifier', knn))
if len(base_estimators) == 0:
base_estimators = [('LogisticRegression', lr),('RandomForestClassifier', rf),('GaussianNB', gnb),('SVC', svc),('KNeighborsClassifier', knn)]
stacking_c = StackingClassifier(estimators=base_estimators, final_estimator=final_estimator_a)
stacking_c.fit(X_train, y_train)
y_predict=stacking_c.predict(X_test)
objClf = aion_matrix()
accuracy_score_test = objClf.get_score(self.scoreParam,y_test,y_predict)
MakeFP0 = False
MakeFN0 = False
threshold = -1
recallscore = -1
precisionscore =-1
if MakeFP0:
self.log.info('-------- Ensemble: Calculate Threshold for FP Start-------')
startRange = 0.0
endRange = 1.0
stepsize = 0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = objClf.check_threshold(stacking_c,X_train,y_train,threshold_range,'FP','')
MakeFP0 = True
self.log.info('-------- Calculate Threshold for FP End-------')
elif MakeFN0:
self.log.info('-------- Ensemble: Calculate Threshold for FN Start-------')
startRange = 1.0
endRange = 0.0
stepsize = -0.01
threshold_range = np.arange(startRange,endRange,stepsize)
threshold,precisionscore,recallscore = objClf.check_threshold(stacking_c,X_train,y_train,threshold_range,'FN','')
MakeFN0 = True
self.log.info('-------- Calculate Threshold for FN End-------')
if threshold != -1:
predictedData = stacking_c.predict_proba(X_test)
predictedData = binarize(predictedData[:,1].reshape(-1, 1),threshold=threshold) #bug 12437
accuracy_score_test = objClf.get_score(self.scoreParam,y_test,predictedData)
best_alg_stacking=stacking_c.__class__.__name__
self.log.info('-----------> Best Stacking Classifier Model '+str(best_alg_stacking))
self.log.info('-----------> Best Score '+str(accuracy_score_test))
return stacking_c,stacking_c.get_params(),accuracy_score_test,best_alg_stacking,threshold,precisionscore,recallscore
def ensemble_stacking__regressor(self,X_train,y_train, X_test, y_test,modelList):
final_estimator_c,final_estimator_r,seed,max_samples,max_features_percentage= self.stacking_params()
final_estimator=final_estimator_r
final_estimator_a=None
scoredetails=''
lr=LinearRegression()
rcv=RidgeCV()
svr=LinearSVR()
lcv=LassoCV()
rf=RandomForestRegressor(random_state=42)
try:
if (final_estimator == 'LinearRegression'):
final_estimator_a=lr
if (final_estimator == 'RidgeCV'):
final_estimator_a=rcv
elif (final_estimator == 'LinearSVR'):
final_estimator_a=svr
elif (final_estimator == 'LassoCV'):
final_estimator_a=lcv
elif (final_estimator == 'RandomForestRegressor'):
final_estimator_a=rf
else:
#default is RidgeCV
final_estimator_a=rcv
except Exception as e:
self.log.info("stacking regression Exception info: \n")
self.log.info(e)
final_estimator_a=rcv
base_estimators = []
if 'Linear Regression' in modelList:
base_estimators.append(('LinearRegression', lr))
if 'Ridge' in modelList:
base_estimators.append(('RidgeCV', rcv))
if 'LinearSVR' in modelList:
base_estimators.append(('LinearSVR', svr))
if 'Lasso' in modelList:
base_estimators.append(('LassoCV', lcv))
if 'Random Forest' in modelList:
base_estimators.append(('RandomForestRegressor', rf))
if len(base_estimators) == 0:
base_estimators = [('LinearRegression', lr),('RidgeCV', rcv),('LinearSVR', svr),('LassoCV', lcv),('RandomForestRegressor', rf)]
self.log.info("Stacking Base Alogs :"+str(base_estimators))
self.log.info("Final Estimator :"+final_estimator)
stacking_regressor = StackingRegressor(estimators=base_estimators,final_estimator=final_estimator_a)
stacking_r_model=stacking_regressor.fit(X_train, y_train)
stacking_rpredict=stacking_regressor.predict(X_test)
best_stacking_alg=stacking_regressor.__class__.__name__
#Accuracy
accuracy_score_best=stacking_regressor.score(X_test, y_test)
aion_matrixobj = aion_matrix()
score = aion_matrixobj.get_score(self.scoreParam,y_test,stacking_rpredict)
return stacking_regressor,stacking_regressor.get_params(),score,best_stacking_alg
|
__init__.py | from .genetic_optimization import GeneticOptimizationCV |
genetic_optimization.py | from sklearn.linear_model import LogisticRegression
from sklearn.naive_bayes import GaussianNB
from sklearn.linear_model import SGDClassifier
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
from sklearn.ensemble import GradientBoostingClassifier
from sklearn.neighbors import KNeighborsClassifier
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Lasso
from sklearn.linear_model import Ridge
from sklearn.tree import DecisionTreeRegressor
from sklearn.ensemble import RandomForestRegressor
import warnings
warnings.filterwarnings('ignore')
from genetic_selection import GeneticSelectionCV
def GeneticOptimizationCV(model,genetic_params,n_iter,scoring,cv):
n_generations = n_iter
estimator = model
selector = GeneticSelectionCV(estimator,cv=cv,**genetic_params,n_generations=n_generations,scoring=scoring)
return selector
|
pipelines.py | import itertools
import logging
from typing import Optional, Dict, Union
from nltk import sent_tokenize
import torch
from transformers import(
AutoModelForSeq2SeqLM,
AutoTokenizer,
PreTrainedModel,
PreTrainedTokenizer,
)
logger = logging.getLogger(__name__)
class QGPipeline:
"""Poor man's QG pipeline"""
def __init__(
self,
model: PreTrainedModel,
tokenizer: PreTrainedTokenizer,
ans_model: PreTrainedModel,
ans_tokenizer: PreTrainedTokenizer,
qg_format: str,
use_cuda: bool
):
self.model = model
self.tokenizer = tokenizer
self.ans_model = ans_model
self.ans_tokenizer = ans_tokenizer
self.qg_format = qg_format
self.device = "cuda" if torch.cuda.is_available() and use_cuda else "cpu"
self.model.to(self.device)
if self.ans_model is not self.model:
self.ans_model.to(self.device)
assert self.model.__class__.__name__ in ["T5ForConditionalGeneration", "BartForConditionalGeneration"]
if "T5ForConditionalGeneration" in self.model.__class__.__name__:
self.model_type = "t5"
else:
self.model_type = "bart"
def __call__(self, inputs: str):
inputs = " ".join(inputs.split())
sents, answers = self._extract_answers(inputs)
flat_answers = list(itertools.chain(*answers))
if len(flat_answers) == 0:
return []
if self.qg_format == "prepend":
qg_examples = self._prepare_inputs_for_qg_from_answers_prepend(inputs, answers)
else:
qg_examples = self._prepare_inputs_for_qg_from_answers_hl(sents, answers)
qg_inputs = [example['source_text'] for example in qg_examples]
questions = self._generate_questions(qg_inputs)
output = [{'answer': example['answer'], 'question': que} for example, que in zip(qg_examples, questions)]
return output
def _generate_questions(self, inputs):
inputs = self._tokenize(inputs, padding=True, truncation=True)
outs = self.model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
max_length=32,
num_beams=4,
)
questions = [self.tokenizer.decode(ids, skip_special_tokens=True) for ids in outs]
return questions
def _extract_answers(self, context):
sents, inputs = self._prepare_inputs_for_ans_extraction(context)
inputs = self._tokenize(inputs, padding=True, truncation=True)
outs = self.ans_model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
max_length=32,
)
dec = [self.ans_tokenizer.decode(ids, skip_special_tokens=False) for ids in outs]
answers = [item.split('<sep>') for item in dec]
answers = [i[:-1] for i in answers]
return sents, answers
def _tokenize(self,
inputs,
padding=True,
truncation=True,
add_special_tokens=True,
max_length=512
):
inputs = self.tokenizer.batch_encode_plus(
inputs,
max_length=max_length,
add_special_tokens=add_special_tokens,
truncation=truncation,
padding="max_length" if padding else False,
pad_to_max_length=padding,
return_tensors="pt"
)
return inputs
def _prepare_inputs_for_ans_extraction(self, text):
sents = sent_tokenize(text)
inputs = []
for i in range(len(sents)):
source_text = "extract answers:"
for j, sent in enumerate(sents):
if i == j:
sent = "<hl> %s <hl>" % sent
source_text = "%s %s" % (source_text, sent)
source_text = source_text.strip()
if self.model_type == "t5":
source_text = source_text + " </s>"
inputs.append(source_text)
return sents, inputs
def _prepare_inputs_for_qg_from_answers_hl(self, sents, answers):
inputs = []
for i, answer in enumerate(answers):
if len(answer) == 0: continue
for answer_text in answer:
sent = sents[i]
sents_copy = sents[:]
answer_text = answer_text.strip()
ans_start_idx = 0
# ans_start_idx = sent.index(answer_text)
# if answer_text in sent:
# ans_start_idx = sent.index(answer_text)
# else:
# continue
sent = f"{sent[:ans_start_idx]} <hl> {answer_text} <hl> {sent[ans_start_idx + len(answer_text): ]}"
sents_copy[i] = sent
source_text = " ".join(sents_copy)
source_text = f"generate question: {source_text}"
if self.model_type == "t5":
source_text = source_text + " </s>"
inputs.append({"answer": answer_text, "source_text": source_text})
return inputs
def _prepare_inputs_for_qg_from_answers_prepend(self, context, answers):
flat_answers = list(itertools.chain(*answers))
examples = []
for answer in flat_answers:
source_text = f"answer: {answer} context: {context}"
if self.model_type == "t5":
source_text = source_text + " </s>"
examples.append({"answer": answer, "source_text": source_text})
return examples
class MultiTaskQAQGPipeline(QGPipeline):
def __init__(self, **kwargs):
super().__init__(**kwargs)
def __call__(self, inputs: Union[Dict, str]):
if type(inputs) is str:
# do qg
return super().__call__(inputs)
else:
# do qa
return self._extract_answer(inputs["question"], inputs["context"])
def _prepare_inputs_for_qa(self, question, context):
source_text = f"question: {question} context: {context}"
if self.model_type == "t5":
source_text = source_text + " </s>"
return source_text
def _extract_answer(self, question, context):
source_text = self._prepare_inputs_for_qa(question, context)
inputs = self._tokenize([source_text], padding=False)
outs = self.model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
max_length=16,
)
answer = self.tokenizer.decode(outs[0], skip_special_tokens=True)
return answer
class E2EQGPipeline:
def __init__(
self,
model: PreTrainedModel,
tokenizer: PreTrainedTokenizer,
use_cuda: bool
) :
self.model = model
self.tokenizer = tokenizer
self.device = "cuda" if torch.cuda.is_available() and use_cuda else "cpu"
self.model.to(self.device)
assert self.model.__class__.__name__ in ["T5ForConditionalGeneration", "BartForConditionalGeneration"]
if "T5ForConditionalGeneration" in self.model.__class__.__name__:
self.model_type = "t5"
else:
self.model_type = "bart"
self.default_generate_kwargs = {
"max_length": 256,
"num_beams": 4,
"length_penalty": 1.5,
"no_repeat_ngram_size": 3,
"early_stopping": True,
}
def __call__(self, context: str, **generate_kwargs):
inputs = self._prepare_inputs_for_e2e_qg(context)
# TODO: when overrding default_generate_kwargs all other arguments need to be passsed
# find a better way to do this
if not generate_kwargs:
generate_kwargs = self.default_generate_kwargs
input_length = inputs["input_ids"].shape[-1]
# max_length = generate_kwargs.get("max_length", 256)
# if input_length < max_length:
# logger.warning(
# "Your max_length is set to {}, but you input_length is only {}. You might consider decreasing max_length manually, e.g. summarizer('...', max_length=50)".format(
# max_length, input_length
# )
# )
outs = self.model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
**generate_kwargs
)
prediction = self.tokenizer.decode(outs[0], skip_special_tokens=True)
questions = prediction.split("<sep>")
questions = [question.strip() for question in questions[:-1]]
return questions
def _prepare_inputs_for_e2e_qg(self, context):
source_text = f"generate questions: {context}"
if self.model_type == "t5":
source_text = source_text + " </s>"
inputs = self._tokenize([source_text], padding=False)
return inputs
def _tokenize(
self,
inputs,
padding=True,
truncation=True,
add_special_tokens=True,
max_length=512
):
inputs = self.tokenizer.batch_encode_plus(
inputs,
max_length=max_length,
add_special_tokens=add_special_tokens,
truncation=truncation,
padding="max_length" if padding else False,
pad_to_max_length=padding,
return_tensors="pt"
)
return inputs
SUPPORTED_TASKS = {
"question-generation": {
"impl": QGPipeline,
"default": {
"model": "valhalla/t5-small-qg-hl",
"ans_model": "valhalla/t5-small-qa-qg-hl",
}
},
"multitask-qa-qg": {
"impl": MultiTaskQAQGPipeline,
"default": {
"model": "valhalla/t5-small-qa-qg-hl",
}
},
"e2e-qg": {
"impl": E2EQGPipeline,
"default": {
"model": "valhalla/t5-small-e2e-qg",
}
}
}
def pipeline(
task: str,
model: Optional = None,
tokenizer: Optional[Union[str, PreTrainedTokenizer]] = None,
qg_format: Optional[str] = "highlight",
ans_model: Optional = None,
ans_tokenizer: Optional[Union[str, PreTrainedTokenizer]] = None,
use_cuda: Optional[bool] = True,
**kwargs,
):
# Retrieve the task
if task not in SUPPORTED_TASKS:
raise KeyError("Unknown task {}, available tasks are {}".format(task, list(SUPPORTED_TASKS.keys())))
targeted_task = SUPPORTED_TASKS[task]
task_class = targeted_task["impl"]
# Use default model/config/tokenizer for the task if no model is provided
if model is None:
model = targeted_task["default"]["model"]
# Try to infer tokenizer from model or config name (if provided as str)
if tokenizer is None:
if isinstance(model, str):
tokenizer = model
else:
# Impossible to guest what is the right tokenizer here
raise Exception(
"Impossible to guess which tokenizer to use. "
"Please provided a PretrainedTokenizer class or a path/identifier to a pretrained tokenizer."
)
# Instantiate tokenizer if needed
if isinstance(tokenizer, (str, tuple)):
if isinstance(tokenizer, tuple):
# For tuple we have (tokenizer name, {kwargs})
tokenizer = AutoTokenizer.from_pretrained(tokenizer[0], **tokenizer[1])
else:
tokenizer = AutoTokenizer.from_pretrained(tokenizer)
# Instantiate model if needed
if isinstance(model, str):
model = AutoModelForSeq2SeqLM.from_pretrained(model)
if task == "question-generation":
if ans_model is None:
# load default ans model
ans_model = targeted_task["default"]["ans_model"]
ans_tokenizer = AutoTokenizer.from_pretrained(ans_model)
ans_model = AutoModelForSeq2SeqLM.from_pretrained(ans_model)
else:
# Try to infer tokenizer from model or config name (if provided as str)
if ans_tokenizer is None:
if isinstance(ans_model, str):
ans_tokenizer = ans_model
else:
# Impossible to guest what is the right tokenizer here
raise Exception(
"Impossible to guess which tokenizer to use. "
"Please provided a PretrainedTokenizer class or a path/identifier to a pretrained tokenizer."
)
# Instantiate tokenizer if needed
if isinstance(ans_tokenizer, (str, tuple)):
if isinstance(ans_tokenizer, tuple):
# For tuple we have (tokenizer name, {kwargs})
ans_tokenizer = AutoTokenizer.from_pretrained(ans_tokenizer[0], **ans_tokenizer[1])
else:
ans_tokenizer = AutoTokenizer.from_pretrained(ans_tokenizer)
if isinstance(ans_model, str):
ans_model = AutoModelForSeq2SeqLM.from_pretrained(ans_model)
if task == "e2e-qg":
return task_class(model=model, tokenizer=tokenizer, use_cuda=use_cuda)
elif task == "question-generation":
return task_class(model=model, tokenizer=tokenizer, ans_model=ans_model, ans_tokenizer=ans_tokenizer, qg_format=qg_format, use_cuda=use_cuda)
else:
return task_class(model=model, tokenizer=tokenizer, ans_model=model, ans_tokenizer=tokenizer, qg_format=qg_format, use_cuda=use_cuda)
|
DRL_train.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import os
import numpy as np
import pandas as pd
import time
import sys
import logging
from reinforcement.agents.ddqn import TrainDDQN
from reinforcement.agents.dqn import TrainDQN
from reinforcement.utils import rounded_dict
from tensorflow.keras.layers import Dense, Dropout
from sklearn.model_selection import train_test_split
from learner.machinelearning import machinelearning
from learner.aion_matrix import aion_matrix
from reinforcement.metrics import network_predictions
from learner.machinelearning import machinelearning
os.environ["CUDA_VISIBLE_DEVICES"] = "-1" # CPU is faster than GPU on structured data
#def TrainRL(input_csv_file, model_save_path, rl_config, RL_Algo_Name):
class ReinformentLearning():
def __init__(self,rlConfig,scoreParam,modelType):
self.rl_config= rlConfig
self.scoreParam = scoreParam
self.log = logging.getLogger('eion')
self.modelType = modelType
def TrainRL(self,xtrain,ytrain,xtest,ytest,algorithm,deployLocation):
try:
scoredetails = ''
X_train, xval, y_train, yval = train_test_split(xtrain, ytrain, test_size=0.2, stratify=ytrain)
X_train = np.array(X_train)
y_train = np.array(y_train)
xval = np.array(xval)
yval = np.array(yval)
valueCount=ytrain.value_counts()
categoryCountList=valueCount.tolist()
xtest = np.array(xtest)
ytest = np.array(ytest)
objClf = aion_matrix()
episodes = self.rl_config['episodes'] # Total number of episodes
warmup_steps = self.rl_config['warmup_steps'] # Amount of warmup steps to collect data with random policy
memory_length = warmup_steps # Max length of the Replay Memory
batch_size = self.rl_config['batch_size']
collect_steps_per_episode = self.rl_config['collect_steps_per_episode']
collect_every = self.rl_config['collect_every']
target_update_period = self.rl_config['target_update_period'] # Period to overwrite the target Q-network with the default Q-network
target_update_tau = self.rl_config['target_update_tau'] # Soften the target model update
n_step_update = self.rl_config['n_step_update']
learning_rate = self.rl_config['learning_rate'] # Learning rate
gamma = self.rl_config['gamma'] # Discount factor
min_epsilon = self.rl_config['min_epsilon'] # Minimal and final chance of choosing random action
decay_episodes = episodes // 10 # Number of episodes to decay from 1.0 to `min_epsilon``
layers = [Dense(128, activation="relu"), #need modification
Dense(64, activation="relu"),
Dense(32, activation="relu"),
Dense(len(np.unique(y_train)), activation=None)]
logFilePath=os.path.join(deployLocation,'log')
if algorithm == "DQN":
start = time.time()
modelName = "DQN"
model_save_path = os.path.dirname(__file__)
model = TrainDQN(episodes, warmup_steps, learning_rate, gamma, min_epsilon, decay_episodes, target_update_period=target_update_period,target_update_tau=target_update_tau, batch_size=batch_size, collect_steps_per_episode=collect_steps_per_episode,memory_length=memory_length, collect_every=collect_every, n_step_update=n_step_update, model_path=model_save_path,log_dir=logFilePath)
model.compile_model(X_train,y_train,layers)
model.q_net.summary()
model.train(xval,yval)
network = model.get_network()
predictedytrain=network_predictions(network,np.array(xtrain))
predictedytest = network_predictions(network,np.array(xtest))
if "DDQN" == algorithm:
start = time.time()
modelName = "DDQN"
model = TrainDDQN(episodes, warmup_steps, learning_rate, gamma, min_epsilon, decay_episodes, target_update_period=target_update_period,target_update_tau=target_update_tau, batch_size=batch_size, collect_steps_per_episode=collect_steps_per_episode,memory_length=memory_length, collect_every=collect_every, n_step_update=n_step_update,log_dir=logFilePath)
model.compile_model(X_train,y_train,layers)
model.q_net.summary()
model.train(xval,yval)
network = model.get_network()
predictedytrain=network_predictions(network,np.array(xtrain))
predictedytest = network_predictions(network,np.array(xtest))
score = objClf.get_score(self.scoreParam,ytest,predictedytest)
score = round(score,2)
return (network,self.rl_config,score,algorithm,-1,-1,-1)
except Exception as inst:
self.log.info( '\n-----> RL Failed!!!.'+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
|
DRL_predict.py |
import os
import numpy as np
import pandas as pd
import time
from DeepRL.agents.ddqn import TrainDDQN
from DeepRL.agents.dqn import TrainDQN
from DeepRL.dataprocess import get_train_test_val
from DeepRL.utils import rounded_dict
from tensorflow.keras.layers import Dense, Dropout
from sklearn.model_selection import train_test_split
os.environ["CUDA_VISIBLE_DEVICES"] = "-1" # CPU is faster than GPU on structured data
def PredictRL(input_csv_file, model_load_path, RL_hparams_config_file, RL_Algo_Name):
if not (os.path.exists(model_load_path)):
os.makedirs(model_load_path)
episodes = RL_hparams_config_file['DeepRL']['episodes'] # Total number of episodes
warmup_steps = RL_hparams_config_file['DeepRL']['warmup_steps'] # Amount of warmup steps to collect data with random policy
memory_length = warmup_steps # Max length of the Replay Memory
batch_size = RL_hparams_config_file['DeepRL']['batch_size']
collect_steps_per_episode = RL_hparams_config_file['DeepRL']['collect_steps_per_episode']
collect_every = RL_hparams_config_file['DeepRL']['collect_every']
target_update_period = RL_hparams_config_file['DeepRL']['target_update_period'] # Period to overwrite the target Q-network with the default Q-network
target_update_tau = RL_hparams_config_file['DeepRL']['target_update_tau'] # Soften the target model update
n_step_update = RL_hparams_config_file['DeepRL']['n_step_update']
learning_rate = RL_hparams_config_file['DeepRL']['learning_rate'] # Learning rate
gamma = RL_hparams_config_file['DeepRL']['gamma'] # Discount factor
min_epsilon = RL_hparams_config_file['DeepRL']['min_epsilon'] # Minimal and final chance of choosing random action
decay_episodes = episodes // 10 # Number of episodes to decay from 1.0 to `min_epsilon``
#path = '/home/renith/Renith/Project/AION/Reinforcement/RL_Classification/Code/rl_text_classification/telemetry_data.csv'
data = pd.read_csv(input_csv_file)
device5 = data[data['device_id'] == "Device_1"]
device5 = device5.drop(['device_id'], axis = 1)
device5.reset_index(drop=True, inplace=True)
target_value = []
for i in range(device5['device_status'].shape[0]):
if(device5['device_status'][i] == "NORMAL"):
target_value.append(0.0)
else:
target_value.append(1.0)
device5['target'] = target_value
device5 = device5.drop(['device_status'], axis = 1)
X_test = device5.iloc[:,1:-1]
y_test = device5.iloc[:,-1]
X_test = X_test.astype(np.float32)
y_test = y_test.astype(np.int32)
#Normalization
mini, maxi = X_test.min(axis=0), X_test.max(axis=0)
X_test -= mini
X_test /= maxi - mini
min_class = [1] #Minority class
maj_class = [0] #Majority class
#X_train, X_test, y_train, y_test = train_test_split(X_train, y_train, test_size=0.8, stratify=y_train)
#X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.2, stratify=y_train)
#X_train = np.array(X_train)
#y_train = np.array(y_train)
#X_val = np.array(X_val)
#y_val = np.array(y_val)
X_test = np.array(X_test)
y_test = np.array(y_test)
#X_train, y_train, X_test, y_test, X_val, y_val = get_train_test_val(X_train.values, y_train.values, X_test.values, y_test.values,
# min_class, maj_class, val_frac=0.2)
layers = [Dense(128, activation="relu"),
Dense(64, activation="relu"),
Dense(32, activation="relu"),
Dense(2, activation=None)]
if(RL_Algo_Name == "DDQN"):
model = TrainDDQN(episodes, warmup_steps, learning_rate, gamma, min_epsilon, decay_episodes, target_update_period=target_update_period,
target_update_tau=target_update_tau, batch_size=batch_size, collect_steps_per_episode=collect_steps_per_episode,
memory_length=memory_length, collect_every=collect_every, n_step_update=n_step_update, model_path=model_load_path)
elif(RL_Algo_Name == "DQN"):
model = TrainDQN(episodes, warmup_steps, learning_rate, gamma, min_epsilon, decay_episodes, target_update_period=target_update_period,
target_update_tau=target_update_tau, batch_size=batch_size, collect_steps_per_episode=collect_steps_per_episode,
memory_length=memory_length, collect_every=collect_every, n_step_update=n_step_update, model_path=model_load_path)
model.compile_model(X_test, y_test, layers)
model.q_net.summary()
#model.train(X_val, y_val, "F1")
#print("Training Ended !!!!")
stats = model.evaluate(X_test, y_test)
print(rounded_dict(stats))
#stats = model.evaluate(X_train, y_train)
#print(rounded_dict(stats))
|
metrics.py | import matplotlib.pyplot as plt
import numpy as np
import seaborn as sns
from sklearn.metrics import (auc, average_precision_score, confusion_matrix,
f1_score, precision_recall_curve, roc_curve,precision_score,recall_score)
from tensorflow import constant
from tf_agents.trajectories import time_step
def network_predictions(network, X: np.ndarray) -> dict:
"""Computes y_pred using a given network.
Input is array of data entries.
:param network: The network to use to calculate metrics
:type network: (Q)Network
:param X: X data, input to network
:type X: np.ndarray
:return: Numpy array of predicted targets for given X
:rtype: np.ndarray
"""
if not isinstance(X, np.ndarray):
raise ValueError(f"`X` must be of type `np.ndarray` not {type(X)}")
q, _ = network(X, step_type=constant([time_step.StepType.FIRST] * X.shape[0]), training=False)
return np.argmax(q.numpy(), axis=1) # Max action for each x in X
def decision_function(network, X: np.ndarray) -> dict:
"""Computes the score for the predicted class of each x in X using a given network.
Input is array of data entries.
:param network: The network to use to calculate the score per x in X
:type network: (Q)Network
:param X: X data, input to network
:type X: np.ndarray
:return: Numpy array of scores for given X
:rtype: np.ndarray
"""
if not isinstance(X, np.ndarray):
raise ValueError(f"`X` must be of type `np.ndarray` not {type(X)}")
q, _ = network(X, step_type=constant([time_step.StepType.FIRST] * X.shape[0]), training=False)
return np.max(q.numpy(), axis=1) # Value of max action for each x in X
def classification_metrics(y_true: list, y_pred: list) -> dict:
"""Computes metrics using y_true and y_pred.
:param y_true: True labels
:type y_true: np.ndarray
:param y_pred: Predicted labels, corresponding to y_true
:type y_pred: np.ndarray
:return: Dictionairy containing Geometric Mean, F1, Precision, Recall, TP, TN, FP, FN
:rtype: dict
"""
if not isinstance(y_true, (list, tuple, np.ndarray)):
raise ValueError(f"`y_true` must be of type `list` not {type(y_true)}")
if not isinstance(y_pred, (list, tuple, np.ndarray)):
raise ValueError(f"`y_pred` must be of type `list` not {type(y_pred)}")
if len(y_true) != len(y_pred):
raise ValueError("`X` and `y` must be of same length.")
#G_mean = np.sqrt(recall * specificity) # Geometric mean of recall and specificity
F1 = f1_score(y_true, y_pred, average='macro') # Default F-measure
recall = recall_score(y_true,y_pred,average='macro')
precision = precision_score(y_true,y_pred,average='macro')
return {"F1": F1, "Precision": precision, "Recall": recall}
def plot_pr_curve(network, X_test: np.ndarray, y_test: np.ndarray,
X_train: np.ndarray = None, y_train: np.ndarray = None) -> None: # pragma: no cover
"""Plots PR curve of X_test and y_test of given network.
Optionally plots PR curve of X_train and y_train.
Average precision is shown in the legend.
:param network: The network to use to calculate the PR curve
:type network: (Q)Network
:param X_test: X data, input to network
:type X_test: np.ndarray
:param y_test: True labels for `X_test`
:type y_test: np.ndarray
:param X_train: Optional X data to plot validation PR curve
:type X_train: np.ndarray
:param y_train: True labels for `X_val`
:type y_train: np.ndarray
:return: None
:rtype: NoneType
"""
plt.plot((0, 1), (1, 0), color="black", linestyle="--", label="Baseline")
# TODO: Consider changing baseline
if X_train is not None and y_train is not None:
y_val_score = decision_function(network, X_train)
val_precision, val_recall, _ = precision_recall_curve(y_train, y_val_score)
val_AP = average_precision_score(y_train, y_val_score)
plt.plot(val_recall, val_precision, label=f"Train AP: {val_AP:.3f}")
y_test_score = decision_function(network, X_test)
test_precision, test_recall, _ = precision_recall_curve(y_test, y_test_score)
test_AP = average_precision_score(y_test, y_test_score)
plt.plot(test_recall, test_precision, label=f"Test AP: {test_AP:.3f}")
plt.xlim((-0.05, 1.05))
plt.ylim((-0.05, 1.05))
plt.xlabel("Recall")
plt.ylabel("Precision")
plt.title("PR Curve")
plt.gca().set_aspect("equal", adjustable="box")
plt.legend(loc="lower left")
plt.grid(True)
plt.show()
def plot_roc_curve(network, X_test: np.ndarray, y_test: np.ndarray,
X_train: np.ndarray = None, y_train: np.ndarray = None) -> None: # pragma: no cover
"""Plots ROC curve of X_test and y_test of given network.
Optionally plots ROC curve of X_train and y_train.
Average precision is shown in the legend.
:param network: The network to use to calculate the PR curve
:type network: (Q)Network
:param X_test: X data, input to network
:type X_test: np.ndarray
:param y_test: True labels for `X_test`
:type y_test: np.ndarray
:param X_train: Optional X data to plot validation PR curve
:type X_train: np.ndarray
:param y_train: True labels for `X_val`
:type y_train: np.ndarray
:return: None
:rtype: NoneType
"""
plt.plot((0, 1), (0, 1), color="black", linestyle="--", label="Baseline")
# TODO: Consider changing baseline
if X_train is not None and y_train is not None:
y_train_score = decision_function(network, X_train)
fpr_train, tpr_train, _ = roc_curve(y_train, y_train_score)
plt.plot(fpr_train, tpr_train, label=f"Train AUROC: {auc(fpr_train, tpr_train):.2f}")
y_test_score = decision_function(network, X_test)
fpr_test, tpr_test, _ = roc_curve(y_test, y_test_score)
plt.plot(fpr_test, tpr_test, label=f"Test AUROC: {auc(fpr_test, tpr_test):.2f}")
plt.xlim((-0.05, 1.05))
plt.ylim((-0.05, 1.05))
plt.xlabel("False Positive Rate")
plt.ylabel("True Positive Rate")
plt.title("ROC Curve")
plt.gca().set_aspect("equal", adjustable="box")
plt.legend(loc="lower right")
plt.grid(True)
plt.show()
def plot_confusion_matrix(TP: int, FN: int, FP: int, TN: int) -> None: # pragma: no cover
"""Plots confusion matric of given TP, FN, FP, TN.
:param TP: True Positive
:type TP: int
:param FN: False Negative
:type FN: int
:param FP: False Positive
:type FP: int
:param TN: True Negative
:type TN: int
:return: None
:rtype: NoneType
"""
if not all(isinstance(i, (int, np.integer)) for i in (TP, FN, FP, TN)):
raise ValueError("Not all arguments are integers.")
ticklabels = ("Minority", "Majority")
sns.heatmap(((TP, FN), (FP, TN)), annot=True, fmt="_d", cmap="viridis", xticklabels=ticklabels, yticklabels=ticklabels)
plt.title("Confusion matrix")
plt.xlabel("Predicted labels")
plt.ylabel("True labels")
plt.show()
|
utils.py | import os
from typing import List
import numpy as np
import pandas as pd
from sklearn.model_selection import train_test_split
def split_csv(fp: str = "./data/creditcard.csv", fp_dest: str = "./data",
name: str = "credit", test_size: int = 0.5, strat_col: str = "Class") -> None:
"""Splits a csv file in two, in a stratified fashion.
Format for filenames will be `{name}0.csv and `{name}1.csv`.
:param fp: The path at which the csv file is located.
:type fp: str
:param fp_dest: The path to save the train and test files.
:type fp_dest: str
:param name: The prefix for the files.
:type name: str
:param test_size: The fraction of total size for the test file.
:type test_size: float
:param strat_col: The column in the original csv file to stratify.
:return: None, two files located at `fp_dest`.
:rtype: NoneType
"""
if not os.path.isfile(fp):
raise FileNotFoundError(f"File at {fp} does not exist.")
if not os.path.isdir(fp_dest):
raise ValueError(f"Directory at {fp_dest} does not exist.")
if not 0 < test_size < 1:
raise ValueError(f"{test_size} is not in interval 0 < x < 1.")
df = pd.read_csv(fp)
if not (strat_col in df.columns):
raise ValueError(f"Stratify column {strat_col} not found in DataFrame.")
train, test = train_test_split(df, test_size=test_size, stratify=df[strat_col])
train.to_csv(f"{fp_dest}/{name}0.csv", index=False)
test.to_csv(f"{fp_dest}/{name}1.csv", index=False)
def rounded_dict(d: dict, precision: int = 6) -> dict:
"""Rounds all values in a dictionairy to `precision` digits after the decimal point.
:param d: Dictionairy containing only floats or ints as values
:type d: dict
:return: Rounded dictionairy
:rtype: dict
"""
return {k: round(v, precision) for k, v in d.items()}
def imbalance_ratio(y: np.ndarray, min_classes: List[int] = [1], maj_classes: List[int] = [0]) -> float:
"""Calculates imbalance ratio of minority class(es) and majority class(es).
:param y: y-vector with labels.
:type y: np.ndarray
:param min_classes: The labels of the minority classes
:type min_classes: list
:param maj_classes: The labels of the minority classes
:type maj_classes: list
:return: The imbalance ratio
:rtype: float
"""
return np.isin(y, min_classes).sum() / np.isin(y, maj_classes).sum()
|
dataprocess.py | import os
from typing import List, Tuple
import numpy as np
from pandas import read_csv
from sklearn.model_selection import train_test_split
from sklearn.utils import shuffle
from tensorflow.keras.datasets import cifar10, fashion_mnist, imdb, mnist
from tensorflow.keras.preprocessing.sequence import pad_sequences
TrainTestData = Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray]
TrainTestValData = Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray, np.ndarray, np.ndarray]
def load_image(data_source: str) -> TrainTestData:
"""
Loads one of the following image datasets: {mnist, famnist, cifar10}.
Normalizes the data. Returns X and y for both train and test datasets.
Dtypes of X's and y's will be `float32` and `int32` to be compatible with `tf_agents`.
:param data_source: Either mnist, famnist or cifar10
:type data_source: str
:return: Tuple of (X_train, y_train, X_test, y_test) containing original split of train/test
:rtype: tuple
"""
reshape_shape = -1, 28, 28, 1
if data_source == "mnist":
(X_train, y_train), (X_test, y_test) = mnist.load_data()
elif data_source == "famnist":
(X_train, y_train), (X_test, y_test) = fashion_mnist.load_data()
elif data_source == "cifar10":
(X_train, y_train), (X_test, y_test) = cifar10.load_data()
reshape_shape = -1, 32, 32, 3
else:
raise ValueError("No valid `data_source`.")
X_train = X_train.reshape(reshape_shape).astype(np.float32) # Float32 is the expected dtype for the observation spec in the env
X_test = X_test.reshape(reshape_shape).astype(np.float32)
X_train /= 255 # /= is not available when casting int to float: https://stackoverflow.com/a/48948461/10603874
X_test /= 255
y_train = y_train.reshape(y_train.shape[0], ).astype(np.int32)
y_test = y_test.reshape(y_test.shape[0], ).astype(np.int32)
return X_train, y_train, X_test, y_test
def load_csv(fp_train: str, fp_test: str, label_col: str, drop_cols: List[str], normalization: bool = False) -> TrainTestData:
"""
Loads any csv-file from local filepaths. Returns X and y for both train and test datasets.
Option to normalize the data with min-max normalization.
Only csv-files with float32 values for the features and int32 values for the labels supported.
Source for dataset: https://mimic-iv.mit.edu/
:param fp_train: Location of the train csv-file
:type fp_train: str
:param fp_test: Location of the test csv-file
:type fp_test: str
:param label_col: The name of the column containing the labels of the data
:rtype label_col: str
:param drop_cols: List of the names of the columns to be dropped. `label_col` gets dropped automatically
:rtype drop_cols: List of strings
:param normalization: Normalize the data with min-max normalization?
:type normalization: bool
:return: Tuple of (X_train, y_train, X_test, y_test) containing original split of train/test
:rtype: Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray]
"""
if not os.path.isfile(fp_train):
raise FileNotFoundError(f"`fp_train` {fp_train} does not exist.")
if not os.path.isfile(fp_test):
raise FileNotFoundError(f"`fp_test` {fp_test} does not exist.")
if not isinstance(normalization, bool):
raise TypeError(f"`normalization` must be of type `bool`, not {type(normalization)}")
X_train = read_csv(fp_train).astype(np.float32) # DataFrames directly converted to float32
X_test = read_csv(fp_test).astype(np.float32)
y_train = X_train[label_col].astype(np.int32)
y_test = X_test[label_col].astype(np.int32)
X_train.drop(columns=drop_cols + [label_col], inplace=True) # Dropping cols and label column
X_test.drop(columns=drop_cols + [label_col], inplace=True)
# Other data sources are already normalized. RGB values are always in range 0 to 255.
if normalization:
mini, maxi = X_train.min(axis=0), X_train.max(axis=0)
X_train -= mini
X_train /= maxi - mini
X_test -= mini
X_test /= maxi - mini
return X_train.values, y_train.values, X_test.values, y_test.values # Numpy arrays
def load_imdb(config: Tuple[int, int] = (5_000, 500)) -> TrainTestData:
"""Loads the IMDB dataset. Returns X and y for both train and test datasets.
:param config: Tuple of number of most frequent words and max length of each sequence.
:type config: str
:return: Tuple of (X_train, y_train, X_test, y_test) containing original split of train/test
:rtype: tuple
"""
if not isinstance(config, (tuple, list)):
raise TypeError(f"{type(config)} is no valid datatype for `config`.")
if len(config) != 2:
raise ValueError("Tuple length of `config` must be 2.")
if not all(i > 0 for i in config):
raise ValueError("All integers of `config` must be > 0.")
(X_train, y_train), (X_test, y_test) = imdb.load_data(num_words=config[0])
X_train = pad_sequences(X_train, maxlen=config[1])
X_test = pad_sequences(X_test, maxlen=config[1])
y_train = y_train.astype(np.int32)
y_test = y_test.astype(np.int32)
return X_train, y_train, X_test, y_test
def get_train_test_val(X_train: np.ndarray, y_train: np.ndarray, X_test: np.ndarray, y_test: np.ndarray, min_classes: List[int],
maj_classes: List[int], imb_ratio: float = None, imb_test: bool = True, val_frac: float = 0.25,
print_stats: bool = True) -> TrainTestValData:
"""
Imbalances data and divides the data into train, test and validation sets.
The imbalance rate of each individual dataset is approx. the same as the given `imb_ratio`.
:param X_train: The X_train data
:type X_train: np.ndarray
:param y_train: The y_train data
:type y_train: np.ndarray
:param X_test: The X_test data
:type X_test: np.ndarray
:param y_test: The y_test data
:type y_test: np.ndarray
:param min_classes: List of labels of all minority classes
:type min_classes: list
:param maj_classes: List of labels of all majority classes.
:type maj_classes: list
:param imb_ratio: Imbalance ratio for minority to majority class: len(minority datapoints) / len(majority datapoints)
If the `imb_ratio` is None, data will not be imbalanced and will only be relabeled to 1's and 0's.
:type imb_ratio: float
:param imb_test: Imbalance the test dataset?
:type imb_test: bool
:param val_frac: Fraction to take from X_train and y_train for X_val and y_val
:type val_frac: float
:param print_stats: Print the imbalance ratio of the imbalanced data?
:type print_stats: bool
:return: Tuple of (X_train, y_train, X_test, y_test, X_val, y_val)
:rtype: Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray, np.ndarray, np.ndarray]
"""
if not 0 < val_frac < 1:
raise ValueError(f"{val_frac} is not in interval 0 < x < 1.")
if not isinstance(print_stats, bool):
raise TypeError(f"`print_stats` must be of type `bool`, not {type(print_stats)}.")
X_train, y_train = imbalance_data(X_train, y_train, min_classes, maj_classes, imb_ratio=imb_ratio)
# Only imbalance test-data if imb_test is True
X_test, y_test = imbalance_data(X_test, y_test, min_classes, maj_classes, imb_ratio=imb_ratio if imb_test else None)
# stratify=y_train to ensure class balance is kept between train and validation datasets
X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=val_frac, stratify=y_train)
if print_stats:
p_train, p_test, p_val = [((y == 1).sum(), imbalance_ratio(y)) for y in (y_train, y_test, y_val)]
print(f"Imbalance ratio `p`:\n"
f"\ttrain: n={p_train[0]}, p={p_train[1]:.6f}\n"
f"\ttest: n={p_test[0]}, p={p_test[1]:.6f}\n"
f"\tvalidation: n={p_val[0]}, p={p_val[1]:.6f}")
return X_train, y_train, X_test, y_test, X_val, y_val
def imbalance_data(X: np.ndarray, y: np.ndarray, min_class: List[int], maj_class: List[int],
imb_ratio: float = None) -> Tuple[np.ndarray, np.ndarray]:
"""
Split data in minority and majority, only values in {min_class, maj_class} will be kept.
(Possibly) decrease minority rows to match the imbalance rate.
If initial imb_ratio of dataset is lower than given `imb_ratio`, the imb_ratio of the returned data will not be changed.
If the `imb_ratio` is None, data will not be imbalanced and will only be relabeled to 1's and 0's.
"""
if not isinstance(X, np.ndarray):
raise TypeError(f"`X` must be of type `np.ndarray` not {type(X)}")
if not isinstance(y, np.ndarray):
raise TypeError(f"`y` must be of type `np.ndarray` not {type(y)}")
if X.shape[0] != y.shape[0]:
raise ValueError("`X` and `y` must contain the same amount of rows.")
if not isinstance(min_class, (list, tuple)):
raise TypeError("`min_class` must be of type list or tuple.")
if not isinstance(maj_class, (list, tuple)):
raise TypeError("`maj_class` must be of type list or tuple.")
if (imb_ratio is not None) and not (0 < imb_ratio < 1):
raise ValueError(f"{imb_ratio} is not in interval 0 < imb_ratio < 1.")
if imb_ratio is None: # Do not imbalance data if no `imb_ratio` is given
imb_ratio = 1
X_min = X[np.isin(y, min_class)] # Mask the correct indexes
X_maj = X[np.isin(y, maj_class)] # Only keep data/labels for x in {min_class, maj_class} and forget all other
min_len = int(X_maj.shape[0] * imb_ratio) # Amount of rows to select from minority classes to get to correct imbalance ratio
# Keep all majority rows, decrease minority rows to match `imb_ratio`
X_min = X_min[np.random.choice(X_min.shape[0], min(min_len, X_min.shape[0]), replace=False), :]
X_imb = np.concatenate([X_maj, X_min]).astype(np.float32)
y_imb = np.concatenate((np.zeros(X_maj.shape[0]), np.ones(X_min.shape[0]))).astype(np.int32)
X_imb, y_imb = shuffle(X_imb, y_imb)
return X_imb, y_imb
|
dqn.py | import os
import pickle
from datetime import datetime
import numpy as np
import tensorflow as tf
from reinforcement.environments.classifierenv import ClassifierEnv
from reinforcement.metrics import (classification_metrics, decision_function,
network_predictions, plot_pr_curve, plot_roc_curve)
from reinforcement.utils import imbalance_ratio
from tensorflow import data
from tensorflow.keras.optimizers import Adam
#from tf_agents.agents.dqn.dqn_agent import DdqnAgent
from tf_agents.agents import DqnAgent
from tf_agents.drivers.dynamic_step_driver import DynamicStepDriver
from tf_agents.environments.tf_py_environment import TFPyEnvironment
from tf_agents.networks.sequential import Sequential
from tf_agents.policies.random_tf_policy import RandomTFPolicy
from tf_agents.replay_buffers.tf_uniform_replay_buffer import \
TFUniformReplayBuffer
from tf_agents.utils import common
class TrainDQN():
"""Wrapper for DDQN training, validation, saving etc."""
def __init__(self, episodes: int, warmup_steps: int, learning_rate: float, gamma: float, min_epsilon: float, decay_episodes: int,
model_path: str = None, log_dir: str = None, batch_size: int = 64, memory_length: int = None,
collect_steps_per_episode: int = 1, val_every: int = None, target_update_period: int = 1, target_update_tau: float = 1.0,
progressbar: bool = True, n_step_update: int = 1, gradient_clipping: float = 1.0, collect_every: int = 1) -> None:
"""
Wrapper to make training easier.
Code is partly based of https://www.tensorflow.org/agents/tutorials/1_dqn_tutorial
:param episodes: Number of training episodes
:type episodes: int
:param warmup_steps: Number of episodes to fill Replay Buffer with random state-action pairs before training starts
:type warmup_steps: int
:param learning_rate: Learning Rate for the Adam Optimizer
:type learning_rate: float
:param gamma: Discount factor for the Q-values
:type gamma: float
:param min_epsilon: Lowest and final value for epsilon
:type min_epsilon: float
:param decay_episodes: Amount of episodes to decay from 1 to `min_epsilon`
:type decay_episodes: int
:param model_path: Location to save the trained model
:type model_path: str
:param log_dir: Location to save the logs, usefull for TensorBoard
:type log_dir: str
:param batch_size: Number of samples in minibatch to train on each step
:type batch_size: int
:param memory_length: Maximum size of the Replay Buffer
:type memory_length: int
:param collect_steps_per_episode: Amount of data to collect for Replay Buffer each episiode
:type collect_steps_per_episode: int
:param collect_every: Step interval to collect data during training
:type collect_every: int
:param val_every: Validate the model every X episodes using the `collect_metrics()` function
:type val_every: int
:param target_update_period: Update the target Q-network every X episodes
:type target_update_period: int
:param target_update_tau: Parameter for softening the `target_update_period`
:type target_update_tau: float
:param progressbar: Enable or disable the progressbar for collecting data and training
:type progressbar: bool
:return: None
:rtype: NoneType
"""
self.episodes = episodes # Total episodes
self.warmup_steps = warmup_steps # Amount of warmup steps before training
self.batch_size = batch_size # Batch size of Replay Memory
self.collect_steps_per_episode = collect_steps_per_episode # Amount of steps to collect data each episode
self.collect_every = collect_every # Step interval to collect data during training
self.learning_rate = learning_rate # Learning Rate
self.gamma = gamma # Discount factor
self.min_epsilon = min_epsilon # Minimal chance of choosing random action
self.decay_episodes = decay_episodes # Number of episodes to decay from 1.0 to `EPSILON`
self.target_update_period = target_update_period # Period for soft updates
self.target_update_tau = target_update_tau
self.progressbar = progressbar # Enable or disable the progressbar for collecting data and training
self.n_step_update = n_step_update
self.gradient_clipping = gradient_clipping # Clip the loss
self.compiled = False
NOW = "DQN" #datetime.now().strftime("%Y%m%d_%H%M%S")
if memory_length is not None:
self.memory_length = memory_length # Max Replay Memory length
else:
self.memory_length = warmup_steps
if val_every is not None:
self.val_every = val_every # Validate the policy every `val_every` episodes
else:
self.val_every = self.episodes // min(50, self.episodes) # Can't validate the model 50 times if self.episodes < 50
if model_path is not None:
#if os.path.exists(model_path + "/" + NOW + ".pkl"):
# os.remove(model_path + "/" + NOW + ".pkl")
self.model_path = model_path + "/" + NOW + ".pkl"
else:
self.model_path = "./models/" + NOW + ".pkl"
if log_dir is None:
log_dir = "./logs/" + NOW
self.writer = tf.summary.create_file_writer(log_dir)
def compile_model(self, X_train, y_train, layers: list = [], imb_ratio: float = None, loss_fn=common.element_wise_squared_loss) -> None:
"""Initializes the neural networks, DDQN-agent, collect policies and replay buffer.
:param X_train: Training data for the model.
:type X_train: np.ndarray
:param y_train: Labels corresponding to `X_train`. 1 for the positive class, 0 for the negative class.
:param y_train: np.ndarray
:param layers: List of layers to feed into the TF-agents custom Sequential(!) layer.
:type layers: list
:param imb_ratio: The imbalance ratio of the data.
:type imb_ratio: float
:param loss_fn: Callable loss function
:type loss_fn: tf.compat.v1.losses
:return: None
:rtype: NoneType
"""
if imb_ratio is None:
imb_ratio = imbalance_ratio(y_train)
self.train_env = TFPyEnvironment(ClassifierEnv(X_train, y_train, imb_ratio))
self.global_episode = tf.Variable(0, name="global_episode", dtype=np.int64, trainable=False) # Global train episode counter
# Custom epsilon decay: https://github.com/tensorflow/agents/issues/339
epsilon_decay = tf.compat.v1.train.polynomial_decay(
1.0, self.global_episode, self.decay_episodes, end_learning_rate=self.min_epsilon)
self.q_net = Sequential(layers, self.train_env.observation_spec())
self.agent = DqnAgent(self.train_env.time_step_spec(),
self.train_env.action_spec(),
q_network=self.q_net,
optimizer=Adam(learning_rate=self.learning_rate),
td_errors_loss_fn=loss_fn,
train_step_counter=self.global_episode,
target_update_period=self.target_update_period,
target_update_tau=self.target_update_tau,
gamma=self.gamma,
epsilon_greedy=epsilon_decay,
n_step_update=self.n_step_update,
gradient_clipping=self.gradient_clipping)
self.agent.initialize()
self.random_policy = RandomTFPolicy(self.train_env.time_step_spec(), self.train_env.action_spec())
self.replay_buffer = TFUniformReplayBuffer(data_spec=self.agent.collect_data_spec,
batch_size=self.train_env.batch_size,
max_length=self.memory_length)
self.warmup_driver = DynamicStepDriver(self.train_env,
self.random_policy,
observers=[self.replay_buffer.add_batch],
num_steps=self.warmup_steps) # Uses a random policy
self.collect_driver = DynamicStepDriver(self.train_env,
self.agent.collect_policy,
observers=[self.replay_buffer.add_batch],
num_steps=self.collect_steps_per_episode) # Uses the epsilon-greedy policy of the agent
self.agent.train = common.function(self.agent.train) # Optimalization
self.warmup_driver.run = common.function(self.warmup_driver.run)
self.collect_driver.run = common.function(self.collect_driver.run)
self.compiled = True
def train(self, *args) -> None:
"""Starts the training of the model. Includes warmup period, metrics collection and model saving.
:param *args: All arguments will be passed to `collect_metrics()`.
This can be usefull to pass callables, testing environments or validation data.
Overwrite the TrainDQN.collect_metrics() function to use your own *args.
:type *args: Any
:return: None
:rtype: NoneType, last step is saving the model as a side-effect
"""
assert self.compiled, "Model must be compiled with model.compile_model(X_train, y_train, layers) before training."
# Warmup period, fill memory with random actions
if self.progressbar:
print(f"\033[92mCollecting data for {self.warmup_steps:_} steps... This might take a few minutes...\033[0m")
self.warmup_driver.run(time_step=None, policy_state=self.random_policy.get_initial_state(self.train_env.batch_size))
if self.progressbar:
print(f"\033[92m{self.replay_buffer.num_frames():_} frames collected!\033[0m")
dataset = self.replay_buffer.as_dataset(sample_batch_size=self.batch_size, num_steps=self.n_step_update + 1,
num_parallel_calls=data.experimental.AUTOTUNE).prefetch(data.experimental.AUTOTUNE)
iterator = iter(dataset)
def _train():
experiences, _ = next(iterator)
return self.agent.train(experiences).loss
_train = common.function(_train) # Optimalization
ts = None
policy_state = self.agent.collect_policy.get_initial_state(self.train_env.batch_size)
print('Before Collect Metrics')
self.collect_metrics(*args) # Initial collection for step 0
print('After Collect Metrics')
for _ in range(self.episodes):
if not self.global_episode % self.collect_every:
# Collect a few steps using collect_policy and save to `replay_buffer`
if self.collect_steps_per_episode != 0:
ts, policy_state = self.collect_driver.run(time_step=ts, policy_state=policy_state)
# Sample a batch of data from `replay_buffer` and update the agent's network
train_loss = _train()
if not self.global_episode % self.val_every:
with self.writer.as_default():
tf.summary.scalar("train_loss", train_loss, step=self.global_episode)
self.collect_metrics(*args)
def collect_metrics(self, X_val: np.ndarray, y_val: np.ndarray, save_best: str = None):
"""Collects metrics using the trained Q-network.
:param X_val: Features of validation data, same shape as X_train
:type X_val: np.ndarray
:param y_val: Labels of validation data, same shape as y_train
:type y_val: np.ndarra
:param save_best: Saving the best model of all validation runs based on given metric:
Choose one of: {Gmean, F1, Precision, Recall, TP, TN, FP, FN}
This improves stability since the model at the last episode is not guaranteed to be the best model.
:type save_best: str
"""
y_pred = network_predictions(self.agent._target_q_network, X_val)
print('classification_metrics')
stats = classification_metrics(y_val, y_pred)
print('Before AVGQ')
avgQ = np.mean(decision_function(self.agent._target_q_network, X_val)) # Max action for each x in X
print('After AVGQ')
if save_best is not None:
if not hasattr(self, "best_score"): # If no best model yet
self.best_score = 0.0
if stats.get(save_best) >= self.best_score: # Overwrite best model
self.save_network() # Saving directly to avoid shallow copy without trained weights
self.best_score = stats.get(save_best)
with self.writer.as_default():
tf.summary.scalar("AverageQ", avgQ, step=self.global_episode) # Average Q-value for this epoch
for k, v in stats.items():
tf.summary.scalar(k, v, step=self.global_episode)
def evaluate(self, X_test, y_test):
"""
Final evaluation of trained Q-network with X_test and y_test.
Optional PR and ROC curve comparison to X_train, y_train to ensure no overfitting is taking place.
:param X_test: Features of test data, same shape as X_train
:type X_test: np.ndarray
:param y_test: Labels of test data, same shape as y_train
:type y_test: np.ndarray
:param X_train: Features of train data
:type X_train: np.ndarray
:param y_train: Labels of train data
:type y_train: np.ndarray
"""
#if hasattr(self, "best_score"):
# print(f"\033[92mBest score: {self.best_score:6f}!\033[0m")
# network = self.load_network(self.model_path) # Load best saved model
#else:
# network = self.agent._target_q_network # Load latest target model
#network = self.load_network(self.model_path)
#if (X_train is not None) and (y_train is not None):
# plot_pr_curve(network, X_test, y_test, X_train, y_train)
# plot_roc_curve(network, X_test, y_test, X_train, y_train)
y_pred = network_predictions(self.agent._target_q_network, X_test)
return classification_metrics(y_test, y_pred)
def save_network(self):
print('save_network')
"""Saves Q-network as pickle to `model_path`."""
with open(self.model_path, "wb") as f: # Save Q-network as pickle
pickle.dump(self.agent._target_q_network, f)
def get_network(self):
"""Static method to load Q-network pickle from given filepath.
:param fp: Filepath to the saved pickle of the network
:type fp: str
:returns: The network-object loaded from a pickle file.
:rtype: tensorflow.keras.models.Model
"""
return self.agent._target_q_network
|
ddqn.py | import os
import pickle
from datetime import datetime
import numpy as np
import tensorflow as tf
from reinforcement.environments.classifierenv import ClassifierEnv
from reinforcement.metrics import (classification_metrics, decision_function,
network_predictions, plot_pr_curve, plot_roc_curve)
from reinforcement.utils import imbalance_ratio
from tensorflow import data
from tensorflow.keras.optimizers import Adam
from tf_agents.agents.dqn.dqn_agent import DdqnAgent
from tf_agents.drivers.dynamic_step_driver import DynamicStepDriver
from tf_agents.environments.tf_py_environment import TFPyEnvironment
from tf_agents.networks.sequential import Sequential
from tf_agents.policies.random_tf_policy import RandomTFPolicy
from tf_agents.replay_buffers.tf_uniform_replay_buffer import \
TFUniformReplayBuffer
from tf_agents.utils import common
class TrainDDQN():
"""Wrapper for DDQN training, validation, saving etc."""
def __init__(self, episodes: int, warmup_steps: int, learning_rate: float, gamma: float, min_epsilon: float, decay_episodes: int,
model_path: str = None, log_dir: str = None, batch_size: int = 64, memory_length: int = None,
collect_steps_per_episode: int = 1, val_every: int = None, target_update_period: int = 1, target_update_tau: float = 1.0,
progressbar: bool = True, n_step_update: int = 1, gradient_clipping: float = 1.0, collect_every: int = 1) -> None:
"""
Wrapper to make training easier.
Code is partly based of https://www.tensorflow.org/agents/tutorials/1_dqn_tutorial
:param episodes: Number of training episodes
:type episodes: int
:param warmup_steps: Number of episodes to fill Replay Buffer with random state-action pairs before training starts
:type warmup_steps: int
:param learning_rate: Learning Rate for the Adam Optimizer
:type learning_rate: float
:param gamma: Discount factor for the Q-values
:type gamma: float
:param min_epsilon: Lowest and final value for epsilon
:type min_epsilon: float
:param decay_episodes: Amount of episodes to decay from 1 to `min_epsilon`
:type decay_episodes: int
:param model_path: Location to save the trained model
:type model_path: str
:param log_dir: Location to save the logs, usefull for TensorBoard
:type log_dir: str
:param batch_size: Number of samples in minibatch to train on each step
:type batch_size: int
:param memory_length: Maximum size of the Replay Buffer
:type memory_length: int
:param collect_steps_per_episode: Amount of data to collect for Replay Buffer each episiode
:type collect_steps_per_episode: int
:param collect_every: Step interval to collect data during training
:type collect_every: int
:param val_every: Validate the model every X episodes using the `collect_metrics()` function
:type val_every: int
:param target_update_period: Update the target Q-network every X episodes
:type target_update_period: int
:param target_update_tau: Parameter for softening the `target_update_period`
:type target_update_tau: float
:param progressbar: Enable or disable the progressbar for collecting data and training
:type progressbar: bool
:return: None
:rtype: NoneType
"""
self.episodes = episodes # Total episodes
self.warmup_steps = warmup_steps # Amount of warmup steps before training
self.batch_size = batch_size # Batch size of Replay Memory
self.collect_steps_per_episode = collect_steps_per_episode # Amount of steps to collect data each episode
self.collect_every = collect_every # Step interval to collect data during training
self.learning_rate = learning_rate # Learning Rate
self.gamma = gamma # Discount factor
self.min_epsilon = min_epsilon # Minimal chance of choosing random action
self.decay_episodes = decay_episodes # Number of episodes to decay from 1.0 to `EPSILON`
self.target_update_period = target_update_period # Period for soft updates
self.target_update_tau = target_update_tau
self.progressbar = progressbar # Enable or disable the progressbar for collecting data and training
self.n_step_update = n_step_update
self.gradient_clipping = gradient_clipping # Clip the loss
self.compiled = False
NOW = "DDQN" #datetime.now().strftime("%Y%m%d_%H%M%S")
if memory_length is not None:
self.memory_length = memory_length # Max Replay Memory length
else:
self.memory_length = warmup_steps
if val_every is not None:
self.val_every = val_every # Validate the policy every `val_every` episodes
else:
self.val_every = self.episodes // min(50, self.episodes) # Can't validate the model 50 times if self.episodes < 50
if model_path is not None:
#if os.path.exists(model_path + "/" + NOW + ".pkl"):
# os.remove(model_path + "/" + NOW + ".pkl")
self.model_path = model_path + "/" + NOW + ".pkl"
else:
self.model_path = "./models/" + NOW + ".pkl"
if log_dir is None:
log_dir = "./logs/" + NOW
self.writer = tf.summary.create_file_writer(log_dir)
def compile_model(self, X_train, y_train, layers: list = [], imb_ratio: float = None, loss_fn=common.element_wise_squared_loss) -> None:
"""Initializes the neural networks, DDQN-agent, collect policies and replay buffer.
:param X_train: Training data for the model.
:type X_train: np.ndarray
:param y_train: Labels corresponding to `X_train`. 1 for the positive class, 0 for the negative class.
:param y_train: np.ndarray
:param layers: List of layers to feed into the TF-agents custom Sequential(!) layer.
:type layers: list
:param imb_ratio: The imbalance ratio of the data.
:type imb_ratio: float
:param loss_fn: Callable loss function
:type loss_fn: tf.compat.v1.losses
:return: None
:rtype: NoneType
"""
if imb_ratio is None:
imb_ratio = imbalance_ratio(y_train)
self.train_env = TFPyEnvironment(ClassifierEnv(X_train, y_train, imb_ratio))
self.global_episode = tf.Variable(0, name="global_episode", dtype=np.int64, trainable=False) # Global train episode counter
# Custom epsilon decay: https://github.com/tensorflow/agents/issues/339
epsilon_decay = tf.compat.v1.train.polynomial_decay(
1.0, self.global_episode, self.decay_episodes, end_learning_rate=self.min_epsilon)
self.q_net = Sequential(layers, self.train_env.observation_spec())
self.agent = DdqnAgent(self.train_env.time_step_spec(),
self.train_env.action_spec(),
q_network=self.q_net,
optimizer=Adam(learning_rate=self.learning_rate),
td_errors_loss_fn=loss_fn,
train_step_counter=self.global_episode,
target_update_period=self.target_update_period,
target_update_tau=self.target_update_tau,
gamma=self.gamma,
epsilon_greedy=epsilon_decay,
n_step_update=self.n_step_update,
gradient_clipping=self.gradient_clipping)
self.agent.initialize()
self.random_policy = RandomTFPolicy(self.train_env.time_step_spec(), self.train_env.action_spec())
self.replay_buffer = TFUniformReplayBuffer(data_spec=self.agent.collect_data_spec,
batch_size=self.train_env.batch_size,
max_length=self.memory_length)
self.warmup_driver = DynamicStepDriver(self.train_env,
self.random_policy,
observers=[self.replay_buffer.add_batch],
num_steps=self.warmup_steps) # Uses a random policy
self.collect_driver = DynamicStepDriver(self.train_env,
self.agent.collect_policy,
observers=[self.replay_buffer.add_batch],
num_steps=self.collect_steps_per_episode) # Uses the epsilon-greedy policy of the agent
self.agent.train = common.function(self.agent.train) # Optimalization
self.warmup_driver.run = common.function(self.warmup_driver.run)
self.collect_driver.run = common.function(self.collect_driver.run)
self.compiled = True
def train(self, *args) -> None:
"""Starts the training of the model. Includes warmup period, metrics collection and model saving.
:param *args: All arguments will be passed to `collect_metrics()`.
This can be usefull to pass callables, testing environments or validation data.
Overwrite the TrainDDQN.collect_metrics() function to use your own *args.
:type *args: Any
:return: None
:rtype: NoneType, last step is saving the model as a side-effect
"""
assert self.compiled, "Model must be compiled with model.compile_model(X_train, y_train, layers) before training."
# Warmup period, fill memory with random actions
if self.progressbar:
print(f"\033[92mCollecting data for {self.warmup_steps:_} steps... This might take a few minutes...\033[0m")
self.warmup_driver.run(time_step=None, policy_state=self.random_policy.get_initial_state(self.train_env.batch_size))
if self.progressbar:
print(f"\033[92m{self.replay_buffer.num_frames():_} frames collected!\033[0m")
dataset = self.replay_buffer.as_dataset(sample_batch_size=self.batch_size, num_steps=self.n_step_update + 1,
num_parallel_calls=data.experimental.AUTOTUNE).prefetch(data.experimental.AUTOTUNE)
iterator = iter(dataset)
def _train():
experiences, _ = next(iterator)
return self.agent.train(experiences).loss
_train = common.function(_train) # Optimalization
ts = None
policy_state = self.agent.collect_policy.get_initial_state(self.train_env.batch_size)
self.collect_metrics(*args) # Initial collection for step 0
for _ in range(self.episodes):
if not self.global_episode % self.collect_every:
# Collect a few steps using collect_policy and save to `replay_buffer`
if self.collect_steps_per_episode != 0:
ts, policy_state = self.collect_driver.run(time_step=ts, policy_state=policy_state)
# Sample a batch of data from `replay_buffer` and update the agent's network
train_loss = _train()
if not self.global_episode % self.val_every:
with self.writer.as_default():
tf.summary.scalar("train_loss", train_loss, step=self.global_episode)
self.collect_metrics(*args)
def collect_metrics(self, X_val: np.ndarray, y_val: np.ndarray, save_best: str = None):
"""Collects metrics using the trained Q-network.
:param X_val: Features of validation data, same shape as X_train
:type X_val: np.ndarray
:param y_val: Labels of validation data, same shape as y_train
:type y_val: np.ndarray
:param save_best: Saving the best model of all validation runs based on given metric:
Choose one of: {Gmean, F1, Precision, Recall, TP, TN, FP, FN}
This improves stability since the model at the last episode is not guaranteed to be the best model.
:type save_best: str
"""
y_pred = network_predictions(self.agent._target_q_network, X_val)
stats = classification_metrics(y_val, y_pred)
avgQ = np.mean(decision_function(self.agent._target_q_network, X_val)) # Max action for each x in X
if save_best is not None:
if not hasattr(self, "best_score"): # If no best model yet
self.best_score = 0.0
if stats.get(save_best) >= self.best_score: # Overwrite best model
self.save_network() # Saving directly to avoid shallow copy without trained weights
self.best_score = stats.get(save_best)
with self.writer.as_default():
tf.summary.scalar("AverageQ", avgQ, step=self.global_episode) # Average Q-value for this epoch
for k, v in stats.items():
tf.summary.scalar(k, v, step=self.global_episode)
def evaluate(self,X_train,y_train, X_test, y_test):
"""
Final evaluation of trained Q-network with X_test and y_test.
Optional PR and ROC curve comparison to X_train, y_train to ensure no overfitting is taking place.
:param X_test: Features of test data, same shape as X_train
:type X_test: np.ndarray
:param y_test: Labels of test data, same shape as y_train
:type y_test: np.ndarray
:param X_train: Features of train data
:type X_train: np.ndarray
:param y_train: Labels of train data
:type y_train: np.ndarray
"""
#if hasattr(self, "best_score"):
# print(f"\033[92mBest score: {self.best_score:6f}!\033[0m")
# network = self.load_network(self.model_path) # Load best saved model
#else:
# network = self.agent._target_q_network # Load latest target model
#network = self.load_network(self.model_path)
#if (X_train is not None) and (y_train is not None):
# plot_pr_curve(network, X_test, y_test, X_train, y_train)
# plot_roc_curve(network, X_test, y_test, X_train, y_train)
y_pred = network_predictions(self.agent._target_q_network, X_test)
return classification_metrics(y_test, y_pred)
def get_network(self):
#network = self.load_network(self.model_path)
return self.agent._target_q_network
def save_network(self, filename_rl): #usnish
"""Saves Q-network as pickle to `model_path`."""
with open(self.filename_rl, "wb") as f: # Save Q-network as pickle
pickle.dump(self.agent._target_q_network, f)
@staticmethod
def load_network(fp: str):
"""Static method to load Q-network pickle from given filepath.
:param fp: Filepath to the saved pickle of the network
:type fp: str
:returns: The network-object loaded from a pickle file.
:rtype: tensorflow.keras.models.Model
"""
with open(fp, "rb") as f: # Load the Q-network
network = pickle.load(f)
return network
|
classifierenv.py | import numpy as np
from tf_agents.environments.py_environment import PyEnvironment
from tf_agents.specs.array_spec import ArraySpec, BoundedArraySpec
from tf_agents.trajectories import time_step as ts
class ClassifierEnv(PyEnvironment):
"""
Custom `PyEnvironment` environment for imbalanced classification.
Based on https://www.tensorflow.org/agents/tutorials/2_environments_tutorial
"""
def __init__(self, X_train: np.ndarray, y_train: np.ndarray, imb_ratio: float):
"""Initialization of environment with X_train and y_train.
:param X_train: Features shaped: [samples, ..., ]
:type X_train: np.ndarray
:param y_train: Labels shaped: [samples]
:type y_train: np.ndarray
:param imb_ratio: Imbalance ratio of the data
:type imb_ratio: float
:returns: None
:rtype: NoneType
"""
#print('1')
self._action_spec = BoundedArraySpec(shape=(), dtype=np.int32, minimum=0, maximum=(len(np.unique(y_train)) - 1), name="action")
#print(y_train)
self._observation_spec = ArraySpec(shape=X_train.shape[1:], dtype=X_train.dtype, name="observation")
#print('3')
self._episode_ended = False
self.X_train = X_train
self.y_train = y_train
self.imb_ratio = imb_ratio # Imbalance ratio: 0 < imb_ratio < 1
self.id = np.arange(self.X_train.shape[0]) # List of IDs to connect X and y data
self.episode_step = 0 # Episode step, resets every episode
self._state = self.X_train[self.id[self.episode_step]]
def action_spec(self):
"""
Definition of the discrete actionspace.
1 for the positive/minority class, 0 for the negative/majority class.
"""
return self._action_spec
def observation_spec(self):
"""Definition of the continous statespace e.g. the observations in typical RL environments."""
return self._observation_spec
def _reset(self):
"""Shuffles data and returns the first state of the shuffled data to begin training on new episode."""
np.random.shuffle(self.id) # Shuffle the X and y data
self.episode_step = 0 # Reset episode step counter at the end of every episode
self._state = self.X_train[self.id[self.episode_step]]
self._episode_ended = False # Reset terminal condition
return ts.restart(self._state)
def _step(self, action: int):
"""
Take one step in the environment.
If the action is correct, the environment will either return 1 or `imb_ratio` depending on the current class.
If the action is incorrect, the environment will either return -1 or -`imb_ratio` depending on the current class.
"""
if self._episode_ended:
# The last action ended the episode. Ignore the current action and start a new episode
return self.reset()
env_action = self.y_train[self.id[self.episode_step]] # The label of the current state
self.episode_step += 1
if action == env_action: # Correct action
if env_action: # Minority
reward = 1 # True Positive
else: # Majority
reward = self.imb_ratio # True Negative
else: # Incorrect action
if env_action: # Minority
reward = -1 # False Negative
self._episode_ended = True # Stop episode when minority class is misclassified
else: # Majority
reward = -self.imb_ratio # False Positive
if self.episode_step == self.X_train.shape[0] - 1: # If last step in data
self._episode_ended = True
self._state = self.X_train[self.id[self.episode_step]] # Update state with new datapoint
if self._episode_ended:
return ts.termination(self._state, reward)
else:
return ts.transition(self._state, reward)
|
aionMlopsService.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
# -*- coding: utf-8 -*-
# -*- coding: utf-8 -*-
import logging
logging.getLogger('tensorflow').disabled = True
import json
import mlflow
import mlflow.sklearn
import mlflow.sagemaker as mfs
# from sklearn.ensemble import RandomForestRegressor
from sklearn.metrics import mean_squared_error
from sklearn.model_selection import train_test_split
# from sklearn import datasets
import time
import numpy as np
# Load dataset
# from sklearn.datasets import load_iris
import pickle
# Load the pickled model
# from matplotlib import pyplot
import sys
import os
import boto3
import subprocess
import os.path
from os.path import expanduser
import platform
from pathlib import Path
class aionMlopsService:
def __init__(self,model,mlflowtosagemakerDeploy,mlflowtosagemakerPushOnly,mlflowtosagemakerPushImageName,mlflowtosagemakerdeployModeluri,experiment_name,mlflow_modelname,awsaccesskey_id,awssecretaccess_key,aws_session_token,mlflow_container_name,aws_region,aws_id,iam_sagemakerfullaccess_arn,sm_app_name,sm_deploy_option,delete_ecr_repository,ecrRepositoryName):
try:
self.model=model
self.mlflowtosagemakerDeploy=mlflowtosagemakerDeploy
self.mlflowtosagemakerPushOnly=str(mlflowtosagemakerPushOnly)
self.mlflowtosagemakerPushImageName=str(mlflowtosagemakerPushImageName)
self.mlflowtosagemakerdeployModeluri=str(mlflowtosagemakerdeployModeluri)
self.experiment_name=experiment_name
self.mlflow_modelname=mlflow_modelname
self.awsaccesskey_id=awsaccesskey_id
self.awssecretaccess_key=awssecretaccess_key
self.aws_session_token=aws_session_token
self.mlflow_container_name=mlflow_container_name
self.aws_region=aws_region
self.aws_id=aws_id
self.iam_sagemakerfullaccess_arn=iam_sagemakerfullaccess_arn
self.sm_app_name=sm_app_name
self.sm_deploy_option=sm_deploy_option
self.delete_ecr_repository=delete_ecr_repository
self.ecrRepositoryName=ecrRepositoryName
from appbe.dataPath import LOG_LOCATION
sagemakerLogLocation = LOG_LOCATION
try:
os.makedirs(sagemakerLogLocation)
except OSError as e:
if (os.path.exists(sagemakerLogLocation)):
pass
else:
raise OSError('sagemakerLogLocation error.')
self.sagemakerLogLocation=str(sagemakerLogLocation)
filename_mlops = 'mlopslog_'+str(int(time.time()))
filename_mlops=filename_mlops+'.log'
# filename = 'mlopsLog_'+Time()
filepath = os.path.join(self.sagemakerLogLocation, filename_mlops)
logging.basicConfig(filename=filepath, format='%(message)s',filemode='w')
# logging.basicConfig(filename="uq_logging.log", format='%(asctime)s %(message)s',filemode='w')
# logging.basicConfig(filename="uq_logging.log", format=' %(message)s',filemode='w')
# logging.basicConfig(filename='uq_logging.log', encoding='utf-8', level=logging.INFO)
self.log = logging.getLogger('aionMLOps')
self.log.setLevel(logging.DEBUG)
# mlflow.set_experiment(self.experiment_name)
except Exception as e:
self.log.info('<!------------- mlflow model INIT Error ---------------> '+str(e))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def mlflowSetPath(self,path):
track_dir=os.path.join(path,'mlruns')
uri="file:"+str(Path(track_dir))
return uri
#Currently not used this delete ecr repository option
def ecr_repository_delete(self,rep_name):
# import subprocess
client = boto3.client('ecr')
repositories = client.describe_repositories()
ecr_delete_rep=client.delete_repository(registryId=self.aws_id,repositoryName=self.ecrRepositoryName,force=True)
mlflow_ecr_delete=subprocess.run(['aws', 'ecr', 'delete-repository','--repository-name',rep_name,'||','true'])
self.log.info('Success: deleted aws ecr repository which contains mlops image.')
def check_sm_deploy_status(self,app_name):
sage_client = boto3.client('sagemaker', region_name=self.aws_region)
endpoint_description = sage_client.describe_endpoint(EndpointName=app_name)
endpoint_status = endpoint_description["EndpointStatus"]
try:
failure_reason=endpoint_description["FailureReason"]
self.log.info("sagemaker end point creation failure reason is: "+str(failure_reason))
except:
pass
endpoint_status=str(endpoint_status)
return endpoint_status
def invoke_sm_endpoint(self,app_name, input_json):
client = boto3.session.Session().client("sagemaker-runtime", self.aws_region)
response = client.invoke_endpoint(
EndpointName=app_name,
Body=input_json,
ContentType='application/json; format=pandas-split',
)
# preds = response['Body'].read().decode("ascii")
preds = response['Body'].read().decode("ascii")
preds = json.loads(preds)
# print("preds: {}".format(preds))
return preds
def predict_sm_app_endpoint(self,X_test):
#print(X_test)
import pandas as pd
prediction=None
AWS_ACCESS_KEY_ID=str(self.awsaccesskey_id)
AWS_SECRET_ACCESS_KEY=str(self.awssecretaccess_key)
AWS_SESSION_TOKEN=str(self.aws_session_token)
region = str(self.aws_region)
#Existing model deploy options
# mlflowtosagemakerPushImageName=str(self.mlflowtosagemakerPushImageName)
# mlflowtosagemakerdeployModeluri=str(self.mlflowtosagemakerdeployModeluri)
try:
import subprocess
cmd = 'aws configure set region_name '+region
os.system(cmd)
cmd = 'aws configure set aws_access_key_id '+AWS_ACCESS_KEY_ID
os.system(cmd)
cmd = 'aws configure set aws_secret_access_key '+AWS_SECRET_ACCESS_KEY
os.system(cmd)
'''
aws_region=subprocess.run(['aws', 'configure', 'set','region_name',region])
aws_accesskeyid=subprocess.run(['aws', 'configure', 'set','aws_access_key_id',AWS_ACCESS_KEY_ID])
aws_secretaccesskey=subprocess.run(['aws', 'configure', 'set','aws_secret_access_key',AWS_SECRET_ACCESS_KEY])
'''
except:
pass
#Create a session for aws communication using aws boto3 lib
# s3_client = boto3.client('ecr',aws_access_key_id=AWS_ACCESS_KEY_ID,aws_secret_access_key=AWS_SECRET_ACCESS_KEY,aws_session_token=AWS_SESSION_TOKEN,region_name=region)
# s3 = boto3.resource('ecr', aws_access_key_id=AWS_ACCESS_KEY_ID, aws_secret_access_key= AWS_SECRET_ACCESS_KEY)
session = boto3.Session(aws_access_key_id=AWS_ACCESS_KEY_ID,aws_secret_access_key=AWS_SECRET_ACCESS_KEY,aws_session_token=AWS_SESSION_TOKEN,region_name=region)
#X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.3, random_state=2)
# query_input = pd.DataFrame(X_test).iloc[[1,5]].to_json(orient="split")
try:
query_input = pd.DataFrame(X_test).to_json(orient="split")
#print(query_input)
prediction = self.invoke_sm_endpoint(app_name=self.sm_app_name, input_json=query_input)
# self.log.info("sagemaker end point Prediction: \n"+str(prediction))
except Exception as e:
print(e)
return prediction
def deleteSagemakerApp(self,app_name,region):
# import mlflow.sagemaker as mfs
# region = 'ap-south-1'
# app_name = 'aion-demo-app'
mfs.delete(app_name=app_name,region_name=region, archive=False,synchronous=True, timeout_seconds=300)
# print("AION mlops sagemaker application endpoint is deleted....\n")
self.log.info('AION mlops sagemaker application endpoint is deleted, application name is: '+str(app_name))
def deployModel2sagemaker(self,mlflow_container_name,tag_id,model_path):
region = str(self.aws_region)
aws_id = str(self.aws_id)
iam_sagemakerfullaccess_arn = str(self.iam_sagemakerfullaccess_arn)
app_name = str(self.sm_app_name)
model_uri = str(model_path)
app_status=False
mlflow_root_dir = None
try:
os.chdir(str(self.sagemakerLogLocation))
mlflow_root_dir = os.getcwd()
self.log.info('mlflow root dir: '+str(mlflow_root_dir))
except:
self.log.info("path issue.")
try:
c_status=self.check_sm_deploy_status(app_name)
#if ((c_status == "Failed") or (c_status == "OutOfService")):
if ((c_status == "Failed") or (c_status.lower() == "failed")):
app_status=False
self.log.info("Sagemaker endpoint status: Failed.\n")
mfs.delete(app_name=app_name,region_name=region, archive=False,synchronous=True, timeout_seconds=300)
elif ((c_status.lower() == "inservice") or (c_status == "InService")):
app_status=True
self.log.info("Sagemaker endpoint status: InService. Running sagemaker endpoint name: \n"+str(app_name))
else:
app_status=False
pass
except:
# print("deploy status error.\n")
pass
#aws ecr model app_name should contain only [[a-zA-Z0-9-]]
import re
if app_name:
pattern = re.compile("[A-Za-z0-9-]+")
# if found match (entire string matches pattern)
if pattern.fullmatch(app_name) is not None:
#print("Found match: ")
pass
else:
app_name = 'aion-demo-app'
else:
app_name = 'aion-demo-app'
mlflow_image=mlflow_container_name+':'+tag_id
image_url = aws_id + '.dkr.ecr.' + region + '.amazonaws.com/' + mlflow_image
deploy_option="create"
self.log.info('deploy_option: \n'+str(deploy_option))
if (deploy_option.lower() == "create"):
# Other deploy modes: mlflow.sagemaker.DEPLOYMENT_MODE_ADD,mlflow.sagemaker.DEPLOYMENT_MODE_REPLACE
if not (app_status):
try:
mfs.deploy(app_name=app_name,model_uri=model_uri,region_name=region,mode="create",execution_role_arn=iam_sagemakerfullaccess_arn,image_url=image_url)
self.log.info('sagemaker endpoint created and model deployed. Application name is: \n'+str(app_name))
except:
self.log.info('Creating end point application issue.Please check the connection and aws credentials \n')
else:
self.log.info('Sagemaker application with user endpoint name already running.Please check. Please delete the old endpoint with same name.\n')
elif (deploy_option.lower() == "delete"):
# import mlflow.sagemaker as mfs
# # region = 'ap-south-1'
# # app_name = 'aion-demo-app'
# mfs.delete(app_name=app_name,region_name=region, archive=False,synchronous=True, timeout_seconds=300)
# print("Mlflow sagemaker application endpoint is deleted....\n")
# self.log.info('Mlflow sagemaker application endpoint is deleted, application name is: '+str(app_name))
pass
elif (deploy_option.lower() == "add"):
pass
elif (deploy_option.lower() == "replace"):
pass
else:
pass
return app_status
def mlflow2sagemaker_deploy(self):
self.log.info('<!------------- Inside AION mlops to sagemaker communication and deploy process. ---------------> ')
deploy_status=False
app_name = str(self.sm_app_name)
self.log.info('Sagemaker Application Name: '+str(app_name))
uri_mlflow=self.mlflowSetPath(self.sagemakerLogLocation)
mlflow.set_tracking_uri(uri_mlflow)
mlops_trackuri=mlflow.get_tracking_uri()
mlops_trackuri=str(mlops_trackuri)
self.log.info('mlops tracking uri: '+str(mlops_trackuri))
localhost_deploy=False
try:
#Loading aion model to deploy in sagemaker
mlflow.set_experiment(self.experiment_name)
self.log.info('Endpoint Name: '+str(self.experiment_name))
# Assume, the model already loaded from joblib in aionmlflow2smInterface.py file.
aionmodel2deploy=self.model
# run_id = None
# experiment_id=None
# Use the loaded pickled model to make predictions
# pred = knn_from_pickle.predict(X_test)
with mlflow.start_run(run_name='AIONMLOps') as run:
# aionmodel2deploy.fit(X_train, y_train)
# predictions = aionmodel2deploy.predict(X_test)
mlflow.sklearn.log_model(aionmodel2deploy, self.mlflow_modelname)
run_id = run.info.run_uuid
experiment_id = run.info.experiment_id
self.log.info('AION mlops experiment run_id: '+str(run_id))
self.log.info('AION mlops experiment experiment_id: '+str(experiment_id))
self.log.info('AION mlops experiment model_name: '+str(self.mlflow_modelname))
artifact_uri = {mlflow.get_artifact_uri()}
# print("1.artifact_uri: \n",artifact_uri)
mlflow.end_run()
#If we need, we can check the mlflow experiments.
# try:
# mlflow_client = mlflow.tracking.MlflowClient('./mlruns')
# exp_list = mlflow_client.list_experiments()
# except:
# pass
#print("mlflow exp_list: \n",exp_list)
mlflow_modelname=str(self.mlflow_modelname)
mlops_trackuri=mlops_trackuri.replace('file:','')
mlops_trackuri=str(mlops_trackuri)
# mlflow_root_dir = os.getcwd()
mlflow_root_dir = None
try:
os.chdir(str(self.sagemakerLogLocation))
mlflow_root_dir = os.getcwd()
self.log.info('mlflow root dir: '+str(mlflow_root_dir))
except:
self.log.info("path issue.")
model_path = 'mlruns/%s/%s/artifacts/%s' % (experiment_id, run_id,self.mlflow_modelname)
# model_path=mlops_trackuri+'\\%s\\%s\\artifacts\\%s' % (experiment_id, run_id,mlflow_modelname)
self.log.info("local host aion mlops model_path is: "+str(model_path))
time.sleep(2)
#print("Environment variable setup in the current working dir for aws sagemaker cli connection... \n")
self.log.info('Environment variable setup in the current working dir for aws sagemaker cli connection... \n ')
AWS_ACCESS_KEY_ID=str(self.awsaccesskey_id)
AWS_SECRET_ACCESS_KEY=str(self.awssecretaccess_key)
AWS_SESSION_TOKEN=str(self.aws_session_token)
region = str(self.aws_region)
#Existing model deploy options
mlflowtosagemakerPushImageName=str(self.mlflowtosagemakerPushImageName)
mlflowtosagemakerdeployModeluri=str(self.mlflowtosagemakerdeployModeluri)
import subprocess
cmd = 'aws configure set region_name '+region
os.system(cmd)
cmd = 'aws configure set aws_access_key_id '+AWS_ACCESS_KEY_ID
os.system(cmd)
cmd = 'aws configure set aws_secret_access_key '+AWS_SECRET_ACCESS_KEY
os.system(cmd)
#Create a session for aws communication using aws boto3 lib
# s3_client = boto3.client('ecr',aws_access_key_id=AWS_ACCESS_KEY_ID,aws_secret_access_key=AWS_SECRET_ACCESS_KEY,aws_session_token=AWS_SESSION_TOKEN,region_name=region)
# s3 = boto3.resource('ecr', aws_access_key_id=AWS_ACCESS_KEY_ID, aws_secret_access_key= AWS_SECRET_ACCESS_KEY)
session = boto3.Session(aws_access_key_id=AWS_ACCESS_KEY_ID,aws_secret_access_key=AWS_SECRET_ACCESS_KEY,aws_session_token=AWS_SESSION_TOKEN,region_name=region)
# session = boto3.session.Session(
# aws_access_key_id=AWS_ACCESS_KEY_ID,
# aws_secret_access_key=AWS_SECRET_ACCESS_KEY,
# aws_session_token=AWS_SESSION_TOKEN
# )
# awsclient = session.resource('ecr')
# s3 = session.resource('s3')
self.log.info('aws environment variable setup done... \n')
try:
os.chdir(mlflow_root_dir)
except FileNotFoundError:
self.log.info('Directory does not exist. '+str(mlflow_root_dir))
except NotADirectoryError:
self.log.info('model_path is not a directory. '+str(mlflow_root_dir))
except PermissionError:
self.log.info('Issue in permissions to change to model dir. '+str(mlflow_root_dir))
mlflow_container_name=str(self.mlflow_container_name)
mlflow_version=mlflow.__version__
tag_id=mlflow_version
if (self.mlflowtosagemakerPushOnly.lower() == "true"):
self.log.info('Selected option is <Deploy existing model to sagemaker> \n')
aws_id=str(self.aws_id)
arn=str(self.iam_sagemakerfullaccess_arn)
mlflow_image=mlflow_container_name+':'+tag_id
image_url = aws_id+'.dkr.ecr.'+region+'.amazonaws.com/'+mlflow_image
# print("image_url:========= \n",image_url)
deploy_status=True
try:
model_path=mlflowtosagemakerdeployModeluri
# ##We need to run mlflow docker container command in the artifacts->model directory inside mlruns.
self.log.info('Deploy existing model container-Model path given by user: '+str(model_path))
try:
os.chdir(model_path)
except FileNotFoundError:
self.log.info('Directory does not exist. '+str(model_path))
except NotADirectoryError:
self.log.info('model_path is not a directory. '+str(model_path))
except PermissionError:
self.log.info('Issue in permissions to change to model dir. '+str(model_path))
try:
mfs.push_image_to_ecr(image=mlflowtosagemakerPushImageName)
deploy_status=True
self.log.info('AION mlops pushed the docker container to aws ecr. \n ')
except:
self.log.info("error in pushing existing container to ecr.\n")
deploy_status=False
time.sleep(2)
#Now,change the working dir to root dir,because now deploy needs full mlruns to model name dir.
try:
# print(" Changing directory to mlflow root dir....\n")
os.chdir(mlflow_root_dir)
except FileNotFoundError:
self.log.info('model path is not a directory. '+str(mlflow_root_dir))
except NotADirectoryError:
self.log.info('model path is not a directory. '+str(mlflow_root_dir))
# print("{0} is not a directory".format(mlflow_root_dir))
except PermissionError:
self.log.info('Issue in permissions to change to model dir. '+str(mlflow_root_dir))
# self.deployModel2sagemaker(mlflowtosagemakerPushImageName,tag_id,mlflowtosagemakerdeployModeluri)
try:
if (deploy_status):
self.deployModel2sagemaker(mlflowtosagemakerPushImageName,tag_id,mlflowtosagemakerdeployModeluri)
self.log.info('AION creates docker container and push the container into aws ecr.. ')
time.sleep(2)
except:
self.log.info('AION deploy error.check connection and aws config parameters. ')
deploy_status=False
# self.log.info('model deployed in sagemaker. ')
except Exception as e:
self.log.info('AION mlops failed to push docker container in aws ecr, check configuration parameters. \n'+str(e))
elif (self.mlflowtosagemakerPushOnly.lower() == "false"):
if (self.mlflowtosagemakerDeploy.lower() == "true"):
self.log.info('Selected option is <Create and Deploy model> \n')
deploy_status=True
try:
# ##We need to run mlflow docker container command in the artifacts->model directory inside mlruns.
try:
os.chdir(model_path)
except FileNotFoundError:
self.log.info('Directory does not exist. '+str(model_path))
except NotADirectoryError:
self.log.info('model_path is not a directory. '+str(model_path))
except PermissionError:
self.log.info('Issue in permissions to change to model dir. '+str(model_path))
try:
mlflow_container_push=subprocess.run(['mlflow', 'sagemaker', 'build-and-push-container','--build','--push','--container',mlflow_container_name])
self.log.info('AION mlops creates docker container and push the container into aws ecr.. ')
deploy_status=True
time.sleep(2)
except:
self.log.info('error in pushing aion model container to sagemaker, please check the connection between local host to aws server.')
deploy_status=False
self.log.info('Now deploying the model container to sagemaker starts....\n ')
# Once docker push completes, again going back to mlflow parent dir for deployment
#Now,change the working dir to root dir,because now deploy needs full mlruns to model name dir.
try:
os.chdir(mlflow_root_dir)
except FileNotFoundError:
self.log.info('model_path does not exist. '+str(mlflow_root_dir))
except NotADirectoryError:
self.log.info('model_path is not a directory. '+str(mlflow_root_dir))
except PermissionError:
self.log.info('Issue in permissions to change to model dir. '+str(mlflow_root_dir))
# app_name = str(self.sm_app_name)
try:
if (deploy_status):
self.deployModel2sagemaker(mlflow_container_name,tag_id,model_path)
except:
self.log.info('mlops deploy error.check connection')
deploy_status=False
except Exception as e:
exc = {"status":"FAIL","message":str(e).strip('"')}
out_exc = json.dumps(exc)
self.log.info('mlflow failed to creates docker container please check the aws iam,ecr permission setup, aws id access_key,secret key values for aion.\n')
elif(self.mlflowtosagemakerDeploy.lower() == "false"):
deploy_status=False
localhost_deploy=True
self.log.info('Selected option is <Create AION mlops container in local host .> \n')
self.log.info("User selected create-Deploy sagemaker option as False,")
self.log.info("Creates the AION mlops-sagemaker container locally starting,but doesn't push into aws ecr and deploy in sagemaker. Check the container in docker repository. ")
try:
# ##We need to run AION mlops docker container command in the artifacts->model directory inside mlruns.
try:
os.chdir(model_path)
self.log.info('After change to AION mlops model dir, cwd: '+str(model_path))
except FileNotFoundError:
self.log.info('Directory does not exist. '+str(model_path))
except NotADirectoryError:
self.log.info('model_path is not a directory. '+str(model_path))
except PermissionError:
self.log.info('Issue in permissions to change to model dir. '+str(model_path))
# mlflow_container_local=subprocess.run(['AION mlops', 'sagemaker', 'build-and-push-container','--build','--no-push','--container',mlflow_container_name])
try:
if not (deploy_status):
mlflow_container_local=subprocess.run(['mlflow', 'sagemaker', 'build-and-push-container','--build','--no-push','--container',mlflow_container_name])
self.log.info('AION creates local host bsed docker container and push the container local docker repository. Check with <docker images> command.\n ')
localhost_deploy=True
time.sleep(2)
except:
self.log.info('error in pushing aion model container to sagemaker, please check the connection between local host to aws server.')
deploy_status=False
localhost_deploy=False
# print("AION mlops creates docker container and push the container into aws ecr.\n")
self.log.info('AION mlops creates docker container and stored locally... ')
time.sleep(2)
except Exception as e:
localhost_deploy=False
# print("mlflow failed to creates docker container please check the aws iam,ecr permission setup, aws id access_key,secret key values for aion.\n")
self.log.info('AION mlops failed to creates docker container in local machine.\n'+str(e))
else:
self.log.info('Deploy option not selected, Please check. ')
localhost_deploy=False
deploy_status=False
else:
pass
localhost_container_status="Notdeployed"
mlflow2sm_deploy_status="Notdeployed"
if localhost_deploy:
localhost_container_status="success"
mlflow2sm_deploy_status="Notdeployed"
# print("AION creates local docker container successfully.Please check in docker repository.")
self.log.info("AION creates local docker container successfully.Please check in docker repository.")
# else:
# localhost_container_status="failed"
# # print("AION failed to create local docker container successfully.Please check in docker repository.")
# self.log.info("AION failed to create local docker container successfully.Please check in docker repository.")
if (deploy_status):
# Finally checking whether mlops model is deployed to sagemaker or not.
app_name = str(self.sm_app_name)
deploy_s = self.check_sm_deploy_status(app_name)
if (deploy_s == "InService"):
# print("AION mlops model is deployed at aws sagemaker, use application name(app_name) and region to access.\n")
self.log.info('AION mlops model is deployed at aws sagemaker, use application name(app_name) and region to access.\n'+str(app_name))
mlflow2sm_deploy_status="success"
localhost_container_status="Notdeployed"
else:
# print("AION Mlflow model not able to deploy at aws sagemaker\n")
self.log.info('AION mlops model not able to deploy at aws sagemaker.\n')
mlflow2sm_deploy_status="failed"
localhost_container_status="Notdeployed"
# else:
# mlflow2sm_deploy_status="None"
return mlflow2sm_deploy_status,localhost_container_status
except Exception as inst:
exc = {"status":"FAIL","message":str(inst).strip('"')}
out_exc = json.dumps(exc)
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
baseline.py | import joblib
import pandas as pd
import sys
import math
import time
import pandas as pd
import numpy as np
from sklearn.metrics import confusion_matrix
from sklearn.metrics import precision_score, recall_score, f1_score, accuracy_score
from sklearn.metrics import r2_score,mean_absolute_error,mean_squared_error
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report, confusion_matrix
from sklearn.svm import SVC
from sklearn.linear_model import LinearRegression
import argparse
import json
def mltesting(modelfile,datafile,features,target):
model = joblib.load(modelfile)
ProblemName = model.__class__.__name__
if ProblemName in ['LogisticRegression','SGDClassifier','SVC','DecissionTreeClassifier','RandomForestClassifier','GaussianNB','KNeighborsClassifier','DecisionTreeClassifier','GradientBoostingClassifier','XGBClassifier','LGBMClassifier','CatBoostClassifier']:
Problemtype = 'Classification'
elif ProblemName in ['LinearRegression','Lasso','Ridge','DecisionTreeRegressor','RandomForestRegressor','GradientBoostingRegressor','XGBRegressor','LGBMRegressor','CatBoostRegressor']:
Problemtype = 'Regression'
else:
Problemtype = 'Unknown'
if Problemtype == 'Classification':
Params = model.get_params()
try:
df = pd.read_csv(datafile,encoding='utf-8',skipinitialspace = True)
if ProblemName == 'LogisticRegression' or ProblemName == 'DecisionTreeClassifier' or ProblemName == 'RandomForestClassifier' or ProblemName == 'GaussianNB' or ProblemName == 'KNeighborsClassifier' or ProblemName == 'GradientBoostingClassifier' or ProblemName == 'SVC':
features = model.feature_names_in_
elif ProblemName == 'XGBClassifier':
features = model.get_booster().feature_names
elif ProblemName == 'LGBMClassifier':
features = model.feature_name_
elif ProblemName == 'CatBoostClassifier':
features = model.feature_names_
modelfeatures = features
dfp = df[modelfeatures]
tar = target
target = df[tar]
predic = model.predict(dfp)
output = {}
matrixconfusion = pd.DataFrame(confusion_matrix(predic,target))
matrixconfusion = matrixconfusion.to_json(orient='index')
classificationreport = pd.DataFrame(classification_report(target,predic,output_dict=True)).transpose()
classificationreport = round(classificationreport,2)
classificationreport = classificationreport.to_json(orient='index')
output["Precision"] = "%.2f" % precision_score(target, predic,average='weighted')
output["Recall"] = "%.2f" % recall_score(target, predic,average='weighted')
output["Accuracy"] = "%.2f" % accuracy_score(target, predic)
output["ProblemName"] = ProblemName
output["Status"] = "Success"
output["Params"] = Params
output["Problemtype"] = Problemtype
output["Confusionmatrix"] = matrixconfusion
output["classificationreport"] = classificationreport
# import statistics
# timearray = []
# for i in range(0,5):
# start = time.time()
# predic1 = model.predict(dfp.head(1))
# end = time.time()
# timetaken = (round((end - start) * 1000,2),'Seconds')
# timearray.append(timetaken)
# print(timearray)
start = time.time()
for i in range(0,5):
predic1 = model.predict(dfp.head(1))
end = time.time()
timetaken = (round((end - start) * 1000,2),'Seconds')
# print(timetaken)
start1 = time.time()
for i in range(0,5):
predic2 = model.predict(dfp.head(10))
end1 = time.time()
timetaken1 = (round((end1 - start1) * 1000,2) ,'Seconds')
# print(timetaken1)
start2 = time.time()
for i in range(0,5):
predic3 = model.predict(dfp.head(100))
end2 = time.time()
timetaken2 = (round((end2 - start2) * 1000,2) ,'Seconds')
# print(timetaken2)
output["onerecord"] = timetaken
output["tenrecords"] = timetaken1
output["hundrecords"] = timetaken2
print(json.dumps(output))
except Exception as e:
output = {}
output['Problemtype']='Classification'
output['Status']= "Fail"
output["ProblemName"] = ProblemName
output["Msg"] = 'Detected Model : {} \\n Problem Type : Classification \\n Error : {}'.format(ProblemName, str(e).replace('"','//"').replace('\n', '\\n'))
print(output["Msg"])
print(json.dumps(output))
elif Problemtype == 'Regression':
Params = model.get_params()
try:
df = pd.read_csv(datafile,encoding='utf-8',skipinitialspace = True)
if ProblemName == 'LinearRegression' or ProblemName == 'Lasso' or ProblemName == 'Ridge' or ProblemName == 'DecisionTreeRegressor' or ProblemName == 'RandomForestRegressor' or ProblemName == 'GaussianNB' or ProblemName == 'KNeighborsRegressor' or ProblemName == 'GradientBoostingRegressor':
features = model.feature_names_in_
elif ProblemName == 'XGBRegressor':
features = model.get_booster().feature_names
elif ProblemName == 'LGBMRegressor':
features = model.feature_name_
elif ProblemName == 'CatBoostRegressor':
features = model.feature_names_
modelfeatures = features
dfp = df[modelfeatures]
tar = target
target = df[tar]
predict = model.predict(dfp)
mse = mean_squared_error(target, predict)
mae = mean_absolute_error(target, predict)
rmse = math.sqrt(mse)
r2 = r2_score(target,predict,multioutput='variance_weighted')
output = {}
output["MSE"] = "%.2f" % mean_squared_error(target, predict)
output["MAE"] = "%.2f" % mean_absolute_error(target, predict)
output["RMSE"] = "%.2f" % math.sqrt(mse)
output["R2"] = "%.2f" %r2_score(target,predict,multioutput='variance_weighted')
output["ProblemName"] = ProblemName
output["Problemtype"] = Problemtype
output["Params"] = Params
output['Status']='Success'
start = time.time()
predic1 = model.predict(dfp.head(1))
end = time.time()
timetaken = (round((end - start) * 1000,2) ,'Seconds')
# print(timetaken)
start1 = time.time()
predic2 = model.predict(dfp.head(10))
end1 = time.time()
timetaken1 = (round((end1 - start1) * 1000,2),'Seconds')
# print(timetaken1)
start2 = time.time()
predic3 = model.predict(dfp.head(100))
end2 = time.time()
timetaken2 = (round((end2 - start2) * 1000,2) ,'Seconds')
# print(timetaken2)
output["onerecord"] = timetaken
output["tenrecords"] = timetaken1
output["hundrecords"] = timetaken2
print(json.dumps(output))
except Exception as e:
output = {}
output['Problemtype']='Regression'
output['Status']='Fail'
output["ProblemName"] = ProblemName
output["Msg"] = 'Detected Model : {} \\n Problem Type : Regression \\n Error : {}'.format(ProblemName, str(e).replace('"','//"').replace('\n', '\\n'))
print(json.dumps(output))
else:
output = {}
output['Problemtype']='Unknown'
output['Status']='Fail'
output['Params'] = ''
output["ProblemName"] = ProblemName
output["Msg"] = 'Detected Model : {} \\n Error : {}'.format(ProblemName, 'Model not supported')
print(json.dumps(output))
return(json.dumps(output))
def baseline_testing(modelFile,csvFile,features,target):
features = [x.strip() for x in features.split(',')]
return mltesting(modelFile,csvFile,features,target) |
item_rating.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import pandas as pd
import numpy as np
import os
import datetime, time, timeit
from sklearn.model_selection import KFold
from sklearn.metrics import confusion_matrix
from sklearn.metrics import accuracy_score
from sklearn.metrics import classification_report
import pickle
import logging
class recommendersystem():
def __init__(self,features,svd_params):
self.features = features
self.svd_input = svd_params
self.log = logging.getLogger('eion')
print ("recommendersystem starts \n")
#To extract dict key,values
def extract_params(self,dict):
self.dict=dict
for k,v in self.dict.items():
return k,v
def recommender_model(self,df,outputfile):
from sklearn.metrics.pairwise import cosine_similarity
from utils.file_ops import save_csv
USER_ITEM_MATRIX = 'user_item_matrix'
ITEM_SIMILARITY_MATRIX = 'item_similarity_matrix'
selectedColumns = self.features.split(',')
data = pd.DataFrame()
for i in range(0,len(selectedColumns)):
data[selectedColumns[i]] = df[selectedColumns[i]]
dataset = data
self.log.info('-------> Top(5) Rows')
self.log.info(data.head(5))
start = time.time()
self.log.info('\n----------- Recommender System Training Starts -----------')
#--------------- Task 11190:recommender system changes Start ---Usnish------------------#
# selectedColumns = ['userId', 'movieId', 'rating']
df_eda = df.groupby(selectedColumns[1]).agg(mean_rating=(selectedColumns[2], 'mean'),number_of_ratings=(selectedColumns[2], 'count')).reset_index()
self.log.info('-------> Top 10 most rated Items:')
self.log.info(df_eda.sort_values(by='number_of_ratings', ascending=False).head(10))
matrix = data.pivot_table(index=selectedColumns[1], columns=selectedColumns[0], values=selectedColumns[2])
relative_file = os.path.join(outputfile, 'data', USER_ITEM_MATRIX + '.csv')
matrix.to_csv(relative_file)
item_similarity_cosine = cosine_similarity(matrix.fillna(0))
item_similarity_cosine = pd.DataFrame(item_similarity_cosine,columns=pd.Series([i + 1 for i in range(item_similarity_cosine.shape[0])],name='ItemId'),index=pd.Series([i + 1 for i in range(item_similarity_cosine.shape[0])],name='ItemId'))
self.log.info('---------> Item-Item Similarity matrix created:')
self.log.info(item_similarity_cosine.head(5))
relative_file = os.path.join(outputfile, 'data', ITEM_SIMILARITY_MATRIX + '.csv')
save_csv(item_similarity_cosine,relative_file)
# --------------- recommender system changes End ---Usnish------------------#
executionTime=time.time() - start
self.log.info("------->Execution Time: "+str(executionTime))
self.log.info('----------- Recommender System Training End -----------\n')
return "filename",matrix,"NA","","" |
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
text_similarity.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import numpy as np
import pickle
import pandas as pd
import sys
import time
import os
from os.path import expanduser
import platform
from sklearn.preprocessing import binarize
import logging
import tensorflow as tf
from sklearn.model_selection import train_test_split
from tensorflow.keras import preprocessing
from sklearn.metrics import roc_auc_score
from sklearn.metrics import accuracy_score
from tensorflow.keras.preprocessing.sequence import pad_sequences
from tensorflow.keras.layers import Input, Embedding, LSTM, Lambda
import tensorflow.keras.backend as K
from tensorflow.keras.models import Model
from tensorflow.keras.optimizers import Adam
from tensorflow.keras.layers import Concatenate
from tensorflow.keras.layers import Input, Dense, Flatten, GlobalMaxPool2D, GlobalAvgPool2D, Concatenate, Multiply, Dropout, Subtract, Add, Conv2D
from sklearn.metrics.pairwise import cosine_similarity, cosine_distances
import tensorflow.keras.backend as K
from tensorflow.keras.models import Model, Sequential
from tensorflow.keras import layers, utils, callbacks, optimizers, regularizers
## Keras subclassing based siamese network
class siameseNetwork(Model):
def __init__(self, activation,inputShape, num_iterations):
self.activation=activation
self.log = logging.getLogger('eion')
super(siameseNetwork, self).__init__()
i1 = layers.Input(shape=inputShape)
i2 = layers.Input(shape=inputShape)
featureExtractor = self.build_feature_extractor(inputShape, num_iterations)
f1 = featureExtractor(i1)
f2 = featureExtractor(i2)
#distance vect
distance = layers.Concatenate()([f1, f2])
cosine_loss = tf.keras.losses.CosineSimilarity(axis=1)
c_loss=cosine_loss(f1, f2)
similarity = tf.keras.layers.Dot(axes=1,normalize=True)([f1,f2])
outputs = layers.Dense(1, activation="sigmoid")(distance)
self.model = Model(inputs=[i1, i2], outputs=outputs)
##Build dense sequential layers
def build_feature_extractor(self, inputShape, num_iterations):
layers_config = [layers.Input(inputShape)]
for i, n_units in enumerate(num_iterations):
layers_config.append(layers.Dense(n_units))
layers_config.append(layers.Dropout(0.2))
layers_config.append(layers.BatchNormalization())
layers_config.append(layers.Activation(self.activation))
model = Sequential(layers_config, name='feature_extractor')
return model
def call(self, x):
return self.model(x)
def euclidean_distance(vectors):
(f1, f2) = vectors
sumSquared = K.sum(K.square(f1 - f2), axis=1, keepdims=True)
return K.sqrt(K.maximum(sumSquared, K.epsilon()))
def cosine_similarity(vectors):
(f1, f2) = vectors
f1 = K.l2_normalize(f1, axis=-1)
f2 = K.l2_normalize(f2, axis=-1)
return K.mean(f1 * f2, axis=-1, keepdims=True)
def cos_dist_output_shape(shapes):
shape1, shape2 = shapes
return (shape1[0],1)
class eion_similarity_siamese:
def __init__(self):
self.log = logging.getLogger('eion')
def siamese_model(self,df,col1,col2,targetColumn,conf,pipe,deployLocation,iterName,iterVersion,testPercentage,predicted_data_file):
try:
self.log.info('-------> Read Embedded File')
home = expanduser("~")
if platform.system() == 'Windows':
modelsPath = os.path.join(home,'AppData','Local','HCLT','AION','PreTrainedModels','TextSimilarity')
else:
modelsPath = os.path.join(home,'HCLT','AION','PreTrainedModels','TextSimilarity')
if os.path.isdir(modelsPath) == False:
os.makedirs(modelsPath)
embedding_file_path = os.path.join(modelsPath,'glove.6B.100d.txt')
if not os.path.exists(embedding_file_path):
from pathlib import Path
import urllib.request
import zipfile
location = modelsPath
local_file_path = os.path.join(location,"glove.6B.zip")
file_test, header_test = urllib.request.urlretrieve('http://nlp.stanford.edu/data/wordvecs/glove.6B.zip', local_file_path)
with zipfile.ZipFile(local_file_path, 'r') as zip_ref:
zip_ref.extractall(location)
os.unlink(os.path.join(location,"glove.6B.zip"))
if os.path.isfile(os.path.join(location,"glove.6B.50d.txt")):
os.unlink(os.path.join(location,"glove.6B.50d.txt"))
if os.path.isfile(os.path.join(location,"glove.6B.300d.txt")):
os.unlink(os.path.join(location,"glove.6B.300d.txt"))
if os.path.isfile(os.path.join(location,"glove.6B.200d.txt")):
os.unlink(os.path.join(location,"glove.6B.200d.txt"))
X = df[[col1,col2]]
Y = df[targetColumn]
testPercentage = testPercentage
self.log.info('\n-------------- Test Train Split ----------------')
if testPercentage == 0:
xtrain=X
ytrain=Y
xtest=X
ytest=Y
else:
testSize=testPercentage/100
self.log.info('-------> Split Type: Random Split')
self.log.info('-------> Train Percentage: '+str(testSize))
X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=testSize)
self.log.info('-------> Train Data Shape: '+str(X_train.shape)+' ---------->')
self.log.info('-------> Test Data Shape: '+str(X_test.shape)+' ---------->')
self.log.info('-------------- Test Train Split End ----------------\n')
self.log.info('\n-------------- Train Validate Split ----------------')
X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.20, random_state=42)
self.log.info('-------> Train Data Shape: '+str(X_train.shape)+' ---------->')
self.log.info('-------> Validate Data Shape: '+str(X_val.shape)+' ---------->')
self.log.info('-------------- Train Validate Split End----------------\n')
self.log.info('Status:- |... Train / test split done: '+str(100-testPercentage)+'% train,'+str(testPercentage)+'% test')
train_sentence1 = pipe.texts_to_sequences(X_train[col1].values)
train_sentence2 = pipe.texts_to_sequences(X_train[col2].values)
val_sentence1 = pipe.texts_to_sequences(X_val[col1].values)
val_sentence2 = pipe.texts_to_sequences(X_val[col2].values)
len_vec = [len(sent_vec) for sent_vec in train_sentence1]
max_len = np.max(len_vec)
len_vec = [len(sent_vec) for sent_vec in train_sentence2]
if (max_len < np.max(len_vec)):
max_len = np.max(len_vec)
train_sentence1 = pad_sequences(train_sentence1, maxlen=max_len, padding='post')
train_sentence2 = pad_sequences(train_sentence2, maxlen=max_len, padding='post')
val_sentence1 = pad_sequences(val_sentence1, maxlen=max_len, padding='post')
val_sentence2 = pad_sequences(val_sentence2, maxlen=max_len, padding='post')
y_train = y_train.values
y_val = y_val.values
activation = str(conf['activation'])
model = siameseNetwork(activation,inputShape=train_sentence1.shape[1], num_iterations=[10])
model.compile(
loss="binary_crossentropy",
optimizer=optimizers.Adam(learning_rate=0.0001),
metrics=["accuracy"])
es = callbacks.EarlyStopping(monitor='val_loss', patience=5, verbose=1, restore_best_weights=True)
rlp = callbacks.ReduceLROnPlateau(
monitor='val_loss', factor=0.1, patience=2, min_lr=1e-10, mode='min', verbose=1
)
x_valid=X_val
y_valid=y_val
n_epoch = int(conf['num_epochs'])
batch_size = int(conf['batch_size'])
similarityIndex = conf['similarityIndex']
model.fit([train_sentence1,train_sentence2],y_train.reshape(-1,1), epochs = n_epoch,batch_size=batch_size,
validation_data=([val_sentence1, val_sentence2],y_val.reshape(-1,1)),callbacks=[es, rlp])
scores = model.evaluate([val_sentence1, val_sentence2], y_val.reshape(-1,1), verbose=0)
self.log.info('-------> Model Score Matrix: Accuracy')
self.log.info('-------> Model Score (Validate Data) : '+str(scores[1]))
self.log.info('Status:- |... Algorithm applied: SIAMESE')
test_sentence1 = pipe.texts_to_sequences(X_test[col1].values)
test_sentence2 = pipe.texts_to_sequences(X_test[col2].values)
test_sentence1 = pad_sequences(test_sentence1, maxlen=max_len, padding='post')
test_sentence2 = pad_sequences(test_sentence2, maxlen=max_len, padding='post')
prediction = model.predict([test_sentence1, test_sentence2 ])
n_epoch = conf['num_epochs']
batch_size = conf['batch_size']
activation = conf['activation']
similarityIndex = conf['similarityIndex']
self.log.info('-------> similarityIndex : '+str(similarityIndex))
prediction = np.where(prediction > similarityIndex,1,0)
rocauc_sco = roc_auc_score(y_test,prediction)
acc_sco = accuracy_score(y_test, prediction)
predict_df = pd.DataFrame()
predict_df['actual'] = y_test
predict_df['predict'] = prediction
predict_df.to_csv(predicted_data_file)
self.log.info('-------> Model Score Matrix: Accuracy')
self.log.info('-------> Model Score (Validate Data) : '+str(scores[1]))
self.log.info('Status:- |... Algorithm applied: SIAMESE')
test_sentence1 = pipe.texts_to_sequences(X_test[col1].values)
test_sentence2 = pipe.texts_to_sequences(X_test[col2].values)
test_sentence1 = pad_sequences(test_sentence1, maxlen=max_len, padding='post')
test_sentence2 = pad_sequences(test_sentence2, maxlen=max_len, padding='post')
prediction = model.predict([test_sentence1, test_sentence2 ])
prediction = np.where(prediction > similarityIndex,1,0)
rocauc_sco = roc_auc_score(y_test,prediction)
acc_sco = accuracy_score(y_test, prediction)
predict_df = pd.DataFrame()
predict_df['actual'] = y_test
predict_df['predict'] = prediction
predict_df.to_csv(predicted_data_file)
self.log.info("predict_df: \n"+str(predict_df))
sco = acc_sco
self.log.info('-------> Test Data Accuracy Score : '+str(acc_sco))
self.log.info('Status:- |... Testing Score: '+str(acc_sco))
self.log.info('-------> Test Data ROC AUC Score : '+str(rocauc_sco))
matrix = '"Accuracy":'+str(acc_sco)+',"ROC AUC":'+str(rocauc_sco)
prediction = model.predict([train_sentence1, train_sentence2])
prediction = np.where(prediction > similarityIndex,1,0)
train_rocauc_sco = roc_auc_score(y_train,prediction)
train_acc_sco = accuracy_score(y_train, prediction)
self.log.info('-------> Train Data Accuracy Score : '+str(train_acc_sco))
self.log.info('-------> Train Data ROC AUC Score : '+str(train_rocauc_sco))
trainmatrix = '"Accuracy":'+str(train_acc_sco)+',"ROC AUC":'+str(train_rocauc_sco)
model_tried = '{"Model":"SIAMESE","Score":'+str(sco)+'}'
saved_model = 'textsimilarity_'+iterName+'_'+iterVersion
# filename = os.path.join(deployLocation,'model','textsimilarity_'+iterName+'_'+iterVersion+'.sav')
# filename = os.path.join(deployLocation,'model','textsimilarity_'+iterName+'_'+iterVersion+'.h5')
## Because we are using subclassing layer api, please use dir (as below) to store deep learn model instead of .h5 model.
filename = os.path.join(deployLocation,'model','textsimilarity_'+iterName+'_'+iterVersion)
model.save(filename)
# model.save_weights(filename)
model_name = 'SIAMESE MODEL'
return(model_name,scores[1],matrix,trainmatrix,model_tried,saved_model,filename,max_len,similarityIndex)
except Exception as inst:
self.log.info("SIAMESE failed " + str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)[1]
self.log.info(str(exc_type) + ' ' + str(fname) + ' ' + str(exc_tb.tb_lineno)) |
pipelines.py | import itertools
import logging
from typing import Optional, Dict, Union
from nltk import sent_tokenize
import torch
from transformers import(
AutoModelForSeq2SeqLM,
AutoTokenizer,
PreTrainedModel,
PreTrainedTokenizer,
)
logger = logging.getLogger(__name__)
class QGPipeline:
"""Poor man's QG pipeline"""
def __init__(
self,
model: PreTrainedModel,
tokenizer: PreTrainedTokenizer,
ans_model: PreTrainedModel,
ans_tokenizer: PreTrainedTokenizer,
qg_format: str,
use_cuda: bool
):
self.model = model
self.tokenizer = tokenizer
self.ans_model = ans_model
self.ans_tokenizer = ans_tokenizer
self.qg_format = qg_format
self.device = "cuda" if torch.cuda.is_available() and use_cuda else "cpu"
self.model.to(self.device)
if self.ans_model is not self.model:
self.ans_model.to(self.device)
assert self.model.__class__.__name__ in ["T5ForConditionalGeneration", "BartForConditionalGeneration"]
if "T5ForConditionalGeneration" in self.model.__class__.__name__:
self.model_type = "t5"
else:
self.model_type = "bart"
def __call__(self, inputs: str):
inputs = " ".join(inputs.split())
sents, answers = self._extract_answers(inputs)
flat_answers = list(itertools.chain(*answers))
if len(flat_answers) == 0:
return []
if self.qg_format == "prepend":
qg_examples = self._prepare_inputs_for_qg_from_answers_prepend(inputs, answers)
else:
qg_examples = self._prepare_inputs_for_qg_from_answers_hl(sents, answers)
qg_inputs = [example['source_text'] for example in qg_examples]
questions = self._generate_questions(qg_inputs)
output = [{'answer': example['answer'], 'question': que} for example, que in zip(qg_examples, questions)]
return output
def _generate_questions(self, inputs):
inputs = self._tokenize(inputs, padding=True, truncation=True)
outs = self.model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
max_length=32,
num_beams=4,
)
questions = [self.tokenizer.decode(ids, skip_special_tokens=True) for ids in outs]
return questions
def _extract_answers(self, context):
sents, inputs = self._prepare_inputs_for_ans_extraction(context)
inputs = self._tokenize(inputs, padding=True, truncation=True)
outs = self.ans_model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
max_length=32,
)
dec = [self.ans_tokenizer.decode(ids, skip_special_tokens=False) for ids in outs]
answers = [item.split('<sep>') for item in dec]
answers = [i[:-1] for i in answers]
return sents, answers
def _tokenize(self,
inputs,
padding=True,
truncation=True,
add_special_tokens=True,
max_length=512
):
inputs = self.tokenizer.batch_encode_plus(
inputs,
max_length=max_length,
add_special_tokens=add_special_tokens,
truncation=truncation,
padding="max_length" if padding else False,
pad_to_max_length=padding,
return_tensors="pt"
)
return inputs
def _prepare_inputs_for_ans_extraction(self, text):
sents = sent_tokenize(text)
inputs = []
for i in range(len(sents)):
source_text = "extract answers:"
for j, sent in enumerate(sents):
if i == j:
sent = "<hl> %s <hl>" % sent
source_text = "%s %s" % (source_text, sent)
source_text = source_text.strip()
if self.model_type == "t5":
source_text = source_text + " </s>"
inputs.append(source_text)
return sents, inputs
def _prepare_inputs_for_qg_from_answers_hl(self, sents, answers):
inputs = []
for i, answer in enumerate(answers):
if len(answer) == 0: continue
for answer_text in answer:
sent = sents[i]
sents_copy = sents[:]
answer_text = answer_text.strip()
ans_start_idx = 0
# ans_start_idx = sent.index(answer_text)
# if answer_text in sent:
# ans_start_idx = sent.index(answer_text)
# else:
# continue
sent = f"{sent[:ans_start_idx]} <hl> {answer_text} <hl> {sent[ans_start_idx + len(answer_text): ]}"
sents_copy[i] = sent
source_text = " ".join(sents_copy)
source_text = f"generate question: {source_text}"
if self.model_type == "t5":
source_text = source_text + " </s>"
inputs.append({"answer": answer_text, "source_text": source_text})
return inputs
def _prepare_inputs_for_qg_from_answers_prepend(self, context, answers):
flat_answers = list(itertools.chain(*answers))
examples = []
for answer in flat_answers:
source_text = f"answer: {answer} context: {context}"
if self.model_type == "t5":
source_text = source_text + " </s>"
examples.append({"answer": answer, "source_text": source_text})
return examples
class MultiTaskQAQGPipeline(QGPipeline):
def __init__(self, **kwargs):
super().__init__(**kwargs)
def __call__(self, inputs: Union[Dict, str]):
if type(inputs) is str:
# do qg
return super().__call__(inputs)
else:
# do qa
return self._extract_answer(inputs["question"], inputs["context"])
def _prepare_inputs_for_qa(self, question, context):
source_text = f"question: {question} context: {context}"
if self.model_type == "t5":
source_text = source_text + " </s>"
return source_text
def _extract_answer(self, question, context):
source_text = self._prepare_inputs_for_qa(question, context)
inputs = self._tokenize([source_text], padding=False)
outs = self.model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
max_length=16,
)
answer = self.tokenizer.decode(outs[0], skip_special_tokens=True)
return answer
class E2EQGPipeline:
def __init__(
self,
model: PreTrainedModel,
tokenizer: PreTrainedTokenizer,
use_cuda: bool
) :
self.model = model
self.tokenizer = tokenizer
self.device = "cuda" if torch.cuda.is_available() and use_cuda else "cpu"
self.model.to(self.device)
assert self.model.__class__.__name__ in ["T5ForConditionalGeneration", "BartForConditionalGeneration"]
if "T5ForConditionalGeneration" in self.model.__class__.__name__:
self.model_type = "t5"
else:
self.model_type = "bart"
self.default_generate_kwargs = {
"max_length": 256,
"num_beams": 4,
"length_penalty": 1.5,
"no_repeat_ngram_size": 3,
"early_stopping": True,
}
def __call__(self, context: str, **generate_kwargs):
inputs = self._prepare_inputs_for_e2e_qg(context)
# TODO: when overrding default_generate_kwargs all other arguments need to be passsed
# find a better way to do this
if not generate_kwargs:
generate_kwargs = self.default_generate_kwargs
input_length = inputs["input_ids"].shape[-1]
# max_length = generate_kwargs.get("max_length", 256)
# if input_length < max_length:
# logger.warning(
# "Your max_length is set to {}, but you input_length is only {}. You might consider decreasing max_length manually, e.g. summarizer('...', max_length=50)".format(
# max_length, input_length
# )
# )
outs = self.model.generate(
input_ids=inputs['input_ids'].to(self.device),
attention_mask=inputs['attention_mask'].to(self.device),
**generate_kwargs
)
prediction = self.tokenizer.decode(outs[0], skip_special_tokens=True)
questions = prediction.split("<sep>")
questions = [question.strip() for question in questions[:-1]]
return questions
def _prepare_inputs_for_e2e_qg(self, context):
source_text = f"generate questions: {context}"
if self.model_type == "t5":
source_text = source_text + " </s>"
inputs = self._tokenize([source_text], padding=False)
return inputs
def _tokenize(
self,
inputs,
padding=True,
truncation=True,
add_special_tokens=True,
max_length=512
):
inputs = self.tokenizer.batch_encode_plus(
inputs,
max_length=max_length,
add_special_tokens=add_special_tokens,
truncation=truncation,
padding="max_length" if padding else False,
pad_to_max_length=padding,
return_tensors="pt"
)
return inputs
SUPPORTED_TASKS = {
"question-generation": {
"impl": QGPipeline,
"default": {
"model": "valhalla/t5-small-qg-hl",
"ans_model": "valhalla/t5-small-qa-qg-hl",
}
},
"multitask-qa-qg": {
"impl": MultiTaskQAQGPipeline,
"default": {
"model": "valhalla/t5-small-qa-qg-hl",
}
},
"e2e-qg": {
"impl": E2EQGPipeline,
"default": {
"model": "valhalla/t5-small-e2e-qg",
}
}
}
def pipeline(
task: str,
model: Optional = None,
tokenizer: Optional[Union[str, PreTrainedTokenizer]] = None,
qg_format: Optional[str] = "highlight",
ans_model: Optional = None,
ans_tokenizer: Optional[Union[str, PreTrainedTokenizer]] = None,
use_cuda: Optional[bool] = True,
**kwargs,
):
# Retrieve the task
if task not in SUPPORTED_TASKS:
raise KeyError("Unknown task {}, available tasks are {}".format(task, list(SUPPORTED_TASKS.keys())))
targeted_task = SUPPORTED_TASKS[task]
task_class = targeted_task["impl"]
# Use default model/config/tokenizer for the task if no model is provided
if model is None:
model = targeted_task["default"]["model"]
# Try to infer tokenizer from model or config name (if provided as str)
if tokenizer is None:
if isinstance(model, str):
tokenizer = model
else:
# Impossible to guest what is the right tokenizer here
raise Exception(
"Impossible to guess which tokenizer to use. "
"Please provided a PretrainedTokenizer class or a path/identifier to a pretrained tokenizer."
)
# Instantiate tokenizer if needed
if isinstance(tokenizer, (str, tuple)):
if isinstance(tokenizer, tuple):
# For tuple we have (tokenizer name, {kwargs})
tokenizer = AutoTokenizer.from_pretrained(tokenizer[0], **tokenizer[1])
else:
tokenizer = AutoTokenizer.from_pretrained(tokenizer)
# Instantiate model if needed
if isinstance(model, str):
model = AutoModelForSeq2SeqLM.from_pretrained(model)
if task == "question-generation":
if ans_model is None:
# load default ans model
ans_model = targeted_task["default"]["ans_model"]
ans_tokenizer = AutoTokenizer.from_pretrained(ans_model)
ans_model = AutoModelForSeq2SeqLM.from_pretrained(ans_model)
else:
# Try to infer tokenizer from model or config name (if provided as str)
if ans_tokenizer is None:
if isinstance(ans_model, str):
ans_tokenizer = ans_model
else:
# Impossible to guest what is the right tokenizer here
raise Exception(
"Impossible to guess which tokenizer to use. "
"Please provided a PretrainedTokenizer class or a path/identifier to a pretrained tokenizer."
)
# Instantiate tokenizer if needed
if isinstance(ans_tokenizer, (str, tuple)):
if isinstance(ans_tokenizer, tuple):
# For tuple we have (tokenizer name, {kwargs})
ans_tokenizer = AutoTokenizer.from_pretrained(ans_tokenizer[0], **ans_tokenizer[1])
else:
ans_tokenizer = AutoTokenizer.from_pretrained(ans_tokenizer)
if isinstance(ans_model, str):
ans_model = AutoModelForSeq2SeqLM.from_pretrained(ans_model)
if task == "e2e-qg":
return task_class(model=model, tokenizer=tokenizer, use_cuda=use_cuda)
elif task == "question-generation":
return task_class(model=model, tokenizer=tokenizer, ans_model=ans_model, ans_tokenizer=ans_tokenizer, qg_format=qg_format, use_cuda=use_cuda)
else:
return task_class(model=model, tokenizer=tokenizer, ans_model=model, ans_tokenizer=tokenizer, qg_format=qg_format, use_cuda=use_cuda)
|
__init__.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
''' |
aionNAS.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import logging
logging.getLogger('tensorflow').disabled = True
import json
#from nltk.corpus import stopwords
from collections import Counter
from numpy import mean
from numpy import std
from pandas import read_csv
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import OneHotEncoder
from sklearn.compose import ColumnTransformer
from learner.machinelearning import machinelearning
# from sklearn.dummy import DummyClassifier
# create histograms of numeric input variables
import sys
import os
import re
import pandas as pd
import numpy as np
from learner.aion_matrix import aion_matrix
import tensorflow as tf
tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)
import autokeras as ak
# load the sonar dataset
from sklearn.model_selection import train_test_split
# from sklearn.metrics import cohen_kappa_score
# from sklearn.metrics import roc_auc_score
# from sklearn.metrics import confusion_matrix
from sklearn.metrics import roc_curve
from math import sqrt
from sklearn.metrics import mean_squared_error, explained_variance_score,mean_absolute_error
from sklearn import metrics
class aionNAS:
def __init__(self,nas_class,nas_params,xtrain1,xtest1,ytrain1,ytest1,deployLocation):
try:
self.dfFeatures=None
self.nas_class=nas_class
self.nas_params=nas_params
self.targetFeature=None
self.log = logging.getLogger('eion')
self.n_models=int(self.nas_params['n_models'])
self.n_epochs=int(self.nas_params['n_epochs'])
self.optimizer=self.nas_params['optimizer']
self.metrics=self.nas_params['metrics']
self.tuner=self.nas_params['tuner']
self.seed=int(self.nas_params['seed'])
self.xtrain = xtrain1
self.xtest = xtest1
self.ytrain = ytrain1
self.ytest = ytest1
#self.labelMaps = labelMaps
self.deployLocation=deployLocation
except Exception as e:
self.log.info('<!------------- NAS INIT Error ---------------> ')
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
def paramCheck(self):
try:
if not (self.nas_class):
self.log.info('<!------------- NAS class input Error ---------------> ')
if not (self.nas_params):
self.log.info('<!------------- NAS model hyperparameter input Error ---------------> ')
if not (self.targetFeature):
self.log.info('<!------------- NAS model targetFeature input Error ---------------> ')
if (self.n_models < 1):
self.n_models=1
if not (self.dfFeatures):
self.log.info('<!------------- NAS model features Error ---------------> ')
if (self.n_epochs < 1):
self.n_models=1
if not (self.optimizer):
self.optimizer="adam"
if not (self.tuner):
self.tuner="greedy"
if (self.seed < 1):
self.seed=0
if not (self.metrics):
self.metrics=None
except ValueError:
self.log.info('<------------------ NAS config file error. --------------->')
def recall_m(self,y_true, y_pred):
true_positives = tf.keras.metrics.Sum(tf.keras.backend.round(tf.keras.backend.clip(y_true * y_pred, 0, 1)))
possible_positives = tf.keras.metrics.Sum(tf.keras.backend.round(tf.keras.backend.clip(y_true, 0, 1)))
recall = true_positives / (possible_positives + tf.keras.backend.epsilon())
return recall
def precision_m(self,y_true, y_pred):
true_positives = tf.keras.metrics.Sum(tf.keras.backend.round(tf.keras.backend.clip(y_true * y_pred, 0, 1)))
predicted_positives = tf.keras.metrics.Sum(tf.keras.backend.round(tf.keras.backend.clip(y_pred, 0, 1)))
precision = true_positives / (predicted_positives + tf.keras.backend.epsilon())
return precision
def f1_score(self,y_true, y_pred):
precision = self.precision_m(y_true, y_pred)
recall = self.recall_m(y_true, y_pred)
return 2*((precision*recall)/(precision+recall+tf.keras.backend.epsilon()))
def nasStructdataPreprocess(self):
df=self.data
self.paramCheck()
target=df[self.targetFeature].values
counter = Counter(target)
for k,v in counter.items():
per = v / len(target) * 100
self.log.info('autokeras struct Class=%d, Count=%d, Percentage=%.3f%%' % (k, v, per))
# select columns with numerical data types
num_ix = df.select_dtypes(include=['int64', 'float64']).columns
subset = df[num_ix]
last_ix = len(df.columns) - 1
y=df[self.targetFeature]
X = df.drop(self.targetFeature, axis=1)
#Using Pearson Correlation
# plt.figure(figsize=(12,10))
# cor = df.corr()
# sns.heatmap(cor, annot=True, cmap=plt.cm.Reds)
# plt.show()
# select categorical features
cat_ix = X.select_dtypes(include=['object', 'bool']).columns
# one hot encode cat features only
ct = ColumnTransformer([('o',OneHotEncoder(),cat_ix)], remainder='passthrough')
X = X.reset_index()
X=X.replace(to_replace="NULL",value=0)
X = X.dropna(how='any',axis=0)
X = ct.fit_transform(X)
from sklearn.preprocessing import scale
X = scale(X)
# label encode the target variable to have the classes 0 and 1
y = LabelEncoder().fit_transform(y)
# separate into train and test sets
X_train, X_test, y_train, y_test=train_test_split(X,y,test_size=self.test_size,random_state=1)
return X_train, X_test, y_train, y_test
def nasStructClassification(self,scoreParam):
try:
objClf = aion_matrix()
X_train, X_test, y_train, y_test= self.xtrain, self.xtest, self.ytrain, self.ytest
modelName="nas_structdata_classifier"
self.log.info("Processing structured data block...\n")
s_in = ak.StructuredDataInput()
#s_in = Flatten()(s_in)
s_out = ak.StructuredDataBlock(categorical_encoding=True)(s_in)
self.log.info("Data pipe via autokeras Classification Dense layers ...\n")
s_out = ak.ClassificationHead()(s_out)
self.log.info("applying autokeras automodel to run different neural models...\n")
try:
tuner = str(self.tuner).lower()
except UnicodeEncodeError:
tuner = (self.tuner.encode('utf8')).lower()
nasclf = ak.AutoModel(
inputs=s_in,
outputs=s_out,
overwrite=True,
tuner=tuner,
max_trials=self.n_models,
seed=self.seed)
# compile the model
#nasclf.compile(optimizer='adam', loss='binary_crossentropy', metrics=['acc',self.f1_score,self.precision_m, self.recall_m])
nasclf.fit(X_train, y_train, epochs=self.n_epochs)
best_model = nasclf.export_model()
mpredict=best_model.predict(X_test)
mtpredict=best_model.predict(X_train)
#loss, accuracy, f1_score, precision, recall = nasclf.evaluate(X_test, y_test, verbose=0)
#from sklearn.metrics import classification_report
#Classification report
y_pred_bool = np.argmax(mpredict, axis=1)
y_train_pred_bool = np.argmax(mtpredict, axis=1)
score = objClf.get_score(scoreParam,y_test, y_pred_bool)
#best_model = nasclf.export_model()
best_model_summary=best_model.summary()
filename = os.path.join(self.deployLocation,'log','summary.txt')
with open(filename,'w') as f:
best_model.summary(print_fn=lambda x: f.write(x + '\n'))
f.close()
#self.log.info("==========")
#self.log.info(best_model_summary)
self.log.info("NAS struct data classification, best model summary: \n"+str(best_model.summary(print_fn=self.log.info)))
#self.log.info("==========")
#Save and load model
# # #try:
# try:
# best_model.save("model_class_autokeras", save_format="tf")
# except Exception:
# best_model.save("model_class_autokeras.h5")
# loaded_model = load_model("model_class_autokeras", custom_objects=ak.CUSTOM_OBJECTS)
# loadedmodel_predict=loaded_model.predict(X_test)
loss,accuracy_m=nasclf.evaluate(X_test, y_test)
#mpredict_classes = mpredict.argmax(axis=-1)
#accuracy = accuracy_score(y_test.astype(int), mpredict.astype(int))
# precision tp / (tp + fp)
#precision = precision_score(y_test.astype(int), mpredict.astype(int),average='macro')
# recall: tp / (tp + fn)
#recall = recall_score(y_test.astype(int), mpredict.astype(int),average='macro')
#f1score=f1_score(y_test.astype(int), mpredict.astype(int) , average="macro")
self.log.info("Autokeras struct data classification metrics: \n")
except Exception as inst:
self.log.info("Error: NAS failed "+str(inst))
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
self.log.info(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
print(inst)
return modelName,nasclf,score
def nasStructRegressor(self,scoreParam):
objClf = aion_matrix()
modelName="nas_struct_regressor"
#self.paramCheck()
X_train, X_test, y_train, y_test= self.xtrain, self.xtest, self.ytrain, self.ytest
# Autokeras alg
s_in = ak.StructuredDataInput()
#tf.keras.layers.GlobalMaxPooling2D()(s_in)
s_out = ak.StructuredDataBlock(categorical_encoding=True)(s_in)
self.log.info("Data pipe via autokeras Regression Dense layers ...\n")
s_out = ak.RegressionHead(loss='mse', metrics=['mae'])(s_out)
self.log.info("applying autokeras automodel to evaluate different neural models...\n")
try:
tuner = str(self.tuner).lower()
except UnicodeEncodeError:
tuner = (self.tuner.encode('utf8')).lower()
nas_reg = ak.AutoModel(
inputs=s_in,
outputs=s_out,
overwrite=True,
tuner=tuner,
max_trials=self.n_models)
nas_reg.fit(X_train, y_train, epochs=self.n_epochs)
best_model = nas_reg.export_model()
self.log.info("NAS struct data regression best model summary: \n")
best_model_summary=best_model.summary(print_fn=self.log.info)
self.log.info(best_model_summary)
predictm=best_model.predict(X_test)
mtpredict=best_model.predict(X_train)
score = objClf.get_score(scoreParam,y_test, predictm)
self.log.info("Autokeras struct data regression metrics: \n")
return modelName,nas_reg,score
def nasMain(self,scoreParam):
modelName = ""
nasclf=None
nas_reg=None
#text_reg_model=None
mse_value=0
reg_rmse=0
mape_reg=0
huber_loss_reg=0
accuracy=0
precision=0
recall=0
#Dummy values to return main for classification problems
dummy_score_1=int(0)
#dummy_score_2=int(0)
try:
if ((self.nas_class.lower() == "classification")):
modelName,nasclf,score=self.nasStructClassification(scoreParam)
self.log.info('NAS Struct Classification score: '+str(score))
best_model_nas = nasclf.export_model()
scoredetails = '{"Model":"NAS","Score":'+str(round(score,2))+'}'
return best_model_nas,self.nas_params,round(score,2),'NAS',-1,-1,-1
elif (self.nas_class.lower() == "regression"):
modelName,nas_reg,score =self.nasStructRegressor(scoreParam)
self.log.info('NAS Struct Regression score: '+str(score))
best_model_nas = nas_reg.export_model()
'''
filename = os.path.join(self.deployLocation,'model','autoKerasModel')
best_model_nas = nas_reg.export_model()
try:
best_model_nas.save(filename, save_format="tf")
modelName = 'autoKerasModel'
except Exception:
filename = os.path.join(self.deployLocation,'model','autoKerasModel.h5')
best_model_nas.save(filename)
modelName = 'autoKerasModel.h5'
'''
scoredetails = '{"Model":"NAS","Score":'+str(round(score,2))+'}'
'''
error_matrix = '"MSE":"'+str(round(mse_value,2))+'","RMSE":"'+str(round(reg_rmse,2))+'","MAPE":"'+str(round(mape_reg,2))+'","MSLE":"'+str(round(msle_reg,2))+'"'
'''
return best_model_nas,self.nas_params,score,'NAS'
else:
pass
except Exception as inst:
print(inst)
exc_type, exc_obj, exc_tb = sys.exc_info()
fname = os.path.split(exc_tb.tb_frame.f_code.co_filename)
print(str(exc_type)+' '+str(fname)+' '+str(exc_tb.tb_lineno))
output = {"status":"FAIL","message":str(inst).strip('"')}
output = json.dumps(output)
|
pushrecords.py | '''
*
* =============================================================================
* COPYRIGHT NOTICE
* =============================================================================
* @ Copyright HCL Technologies Ltd. 2021, 2022,2023
* Proprietary and confidential. All information contained herein is, and
* remains the property of HCL Technologies Limited. Copying or reproducing the
* contents of this file, via any medium is strictly prohibited unless prior
* written permission is obtained from HCL Technologies Limited.
*
'''
import socket
import os
import rsa
from os.path import expanduser
from pathlib import Path
import requests
import platform
from appbe.dataPath import DATA_DIR
import socket
import getmac
import subprocess
import sys
import json
from datetime import datetime
import binascii
computername = socket.getfqdn()
global_key = '''
-----BEGIN RSA PUBLIC KEY-----
MIIBCgKCAQEAzJcxqRiUpp7CzViyqNlYaeyceDh5y6Ib4SoxoyNkN3+k0q+cr1lb
k0KdWTtHIVqH1wsLYofYjpB7X2RN0KYTv8VfwmfQNrpFEbiRz4gcAeuxGCPgGaue
N1ttujQMWHWCcY+UH5Voh8YUfkW8P+T3zxvr1d30D+kVBJC59y/31JvTzr3Bw/T+
NYv6xiienYiEYtm9d5ATioEwZOXaQBrtVvRmqcod5A1h4kn1ZauLX2Ph8H4TAuit
NLtw6xUCJNumphP7xdU+ca6P6a6eaLprgKhvky+nz16u9/AC2AazRQHKWf8orS6b
fw16JDCRs0zU4mTQLCjkUUt0edOaRhUtcQIDAQAB
-----END RSA PUBLIC KEY-----
'''
quarter_key = '''
-----BEGIN RSA PUBLIC KEY-----
MIIBCgKCAQEAmKzOJxVEV9ulA+cjfxguAduLMD47OWjLcEAEmEuK8vR4O5f6e2h1
08NniGC+nkwqmM00U7JTVBkqnt9S/JgE3pAH2xwfWda2OvXNWisWmOQdqB0+XRHh
NXsIG3yRk/sMlDpe7MJIyM5ADSu01PLn9FZTfmMq7lEp32tAf71cuUE/dwuWSvEQ
WK2hn1L4D97O43XCd7FHtMSHfgtjdcCFgX9IRgWLKC8Bm3q5qcqF4v3cHuYTj3V9
njxPtRqPg6HJFiJrm9AX5bUEHAvbTcw4wAmsNTRQHPvVB+Lc+yGh5x8crhKjNB01
gdB5I3a4mPO7dKvadR6Mr28trr0Ff5t2HQIDAQAB
-----END RSA PUBLIC KEY-----
'''
halfYear_key='''
-----BEGIN RSA PUBLIC KEY-----
MIIBCgKCAQEAgrGNwl8CNYQmVxi8/GEgPjfL5aEmyPkDyaJb9h4hZDSZCeeKd7Rv
wwhuRTdBBfOp0bQ7QS7NYMg38Xlc3x85I9RnxdQdDKn2nRuvG0hG3wMBFy/DCSXF
tXbDjJkLijAhqcBNu8m+a2Gtn14ShC7TbcfY4iVXho3WFUrn0xq6S5ducqWCsLJh
R+TNImCaMICqfoAzEDGC3ojO5Hi3vJmmyK5CVp6bt4wLRATQjcp1ujGW4Uv4kEgp
7TR077c226v1KOdKdyZPHJzT1MKwZrG2Gdluk3/Y1apbwyGzYqFdTCOAB+mE73Dn
wFXURgDJQmaU2oxxaA13WRcELpnirm+aIwIDAQAB
-----END RSA PUBLIC KEY-----
'''
oneYear_key='''
-----BEGIN RSA PUBLIC KEY-----
MIIBCgKCAQEA3GLqn+vkKn3fTNH3Bbb3Lq60pCoe+mn0KPz74Bp7p5OkZAUe14pP
Tcf/UqdPwiENhSCseWtfZmfKDK8qYRHJ5xW02+AhHPPdiacS45X504/lGG3q/4SG
ZgaFhMDvX+IH/ZH+qqbU3dRQhXJCCrAVAa7MonzM6yPiVeS2SdpMkNg1VDR1oTLB
Pn+qSV6CnkK1cYtWCRQ23GH2Ru7fc09r7m8hVcifKJze84orpHC5FX0WScQuR8h/
fs1IbGkxTOxP8vplUj/cd4JjUxgd+w+8R4kcoPhdGZF5UGeZA8xMERzQLvh+4Ui0
KIvz5/iyKB/ozaeSG0OMwDAk3WDEnb1WqQIDAQAB
-----END RSA PUBLIC KEY-----
'''
full_key='''
-----BEGIN RSA PUBLIC KEY-----
MIIBCgKCAQEAhqfNMuYYLdVrePhkO9rU/qT6FgolzI0YyzIJ2OeJE+++JioYm6nn
ohQU32iiE0DZlCCLrHJXOOIAz2Op80goX0lxtngyxVUPsiB5CI77sAC7x6K3anJ0
elpnQCC0+xV2ZL5eIMNQHLe+X6wJl/HGWqkUlxKpWr4/kBEB4EisW60OePfhntIN
4OUJ7iEq+sDdOM5WazJIXeNV1cig4i6057GE3k5ITcQUmw17DZu2+dqkIscckaG+
t5SF7Qnvt4IY8IeQp2htx3yD+CJCV0u2uKwoSFMGJn3OWdaixC3+eojyMXmfAWtQ
Ee9NLNNaTCMIvQ8BeItJLQs2Htw3bZNMvwIDAQAB
-----END RSA PUBLIC KEY-----
'''
def validate_key_Pair(privatepath,publickey):
with open(privatepath, 'rb') as privatefile:
keydata = privatefile.read()
privatefile.close()
try:
privkey = rsa.PrivateKey.load_pkcs1(keydata,'PEM')
data = 'Validate Global License'
signature = rsa.sign(data.encode('utf-8'), privkey, 'SHA-1')
pubkey = rsa.PublicKey.load_pkcs1(publickey)
except:
return False
try:
rsa.verify(data.encode('utf-8'), signature, pubkey)
return True
except Exception as e:
return False
def updateDRecord(licensepath):
domain_license_path = os.path.join(DATA_DIR,'License','license_domain.lic')
if(os.path.isfile(licensepath)):
with open(licensepath, 'rb') as f:
licensekey = f.read()
f.close()
with open(domain_license_path, 'wb') as f:
f.write(licensekey)
f.close()
if(validate_key_Pair(domain_license_path,global_key)):
return True,'Valid Domain License'
else:
return False,'Invalid Domain License'
else:
return False,'File Not Exists'
def generateLicenseKey(userKey):
record = {'UserKey':userKey}
record = json.dumps(record)
status = 'Error'
url = 'https://qw7e33htlk.execute-api.ap-south-1.amazonaws.com/default/aion_license'
try:
response = requests.post(url, data=record,headers={"x-api-key":"3cQKRkKA4S57pYrkFp1Dd9jRXt4xnFoB9iqhAQRM","Content-Type":"application/json",})
if response.status_code == 200:
outputStr=response.content
outputStr = outputStr.decode('utf-8','ignore')
outputStr = outputStr.strip()
license_dict = json.loads(str(outputStr))
if license_dict['status'] == 'success':
status = 'Success'
licenseKey = license_dict['msg']
else:
status = 'Error'
licenseKey = ''
else:
status = 'Error'
licenseKey = ''
except Exception as inst:
print(inst)
status = 'Error'
licenseKey = ''
msg = {'status':status,'key':userKey,'licenseKey':licenseKey,'link':''}
return msg
def updateRecord(licensepath):
currentDirectory = os.path.dirname(os.path.abspath(__file__))
license_path = os.path.join(currentDirectory,'..','lic','license.lic')
if(os.path.isfile(licensepath)):
with open(licensepath, 'rb') as f:
licensekey = f.read()
f.close()
with open(license_path, 'wb') as f:
f.write(licensekey)
f.close()
status,msg = check_domain_license()
if status:
status,msg = getdaysfromstartdate()
if status:
status,msg = check_days_license(int(msg))
return status,msg
else:
return False,'File Not Exists'
def check_domain_license():
if 'CORP.HCL.IN' in computername:
return True,'HCL Domain'
else:
return True,'HCL Domain'
def diff_month(d1, d2):
return (d1.year - d2.year) * 12 + d1.month - d2.month
def getdaysfromstartdate():
currentDirectory = os.path.dirname(os.path.abspath(__file__))
startdatePath = os.path.join(currentDirectory,'..','lic','startdate.txt')
if(os.path.isfile(startdatePath)):
with open(startdatePath, "rb") as fl:
encrypted_message = fl.read()
fl.close()
privkey = '''-----BEGIN RSA PRIVATE KEY-----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=
-----END RSA PRIVATE KEY-----
'''
privkey = rsa.PrivateKey.load_pkcs1(privkey,'PEM')
decrypted_message = rsa.decrypt(encrypted_message, privkey)
decrypted_message = decrypted_message.decode()
import datetime
start_time = datetime.datetime.strptime(decrypted_message, '%Y-%m-%d')
current_date = datetime.datetime.today().strftime('%Y-%m-%d')
current_date = datetime.datetime.strptime(current_date, '%Y-%m-%d')
Months = diff_month(current_date,start_time)
return True,Months
else:
return False,'Start Date Not Exists'
def check_days_license(months):
currentDirectory = os.path.dirname(os.path.abspath(__file__))
license_path = os.path.join(currentDirectory,'..','lic','license.lic')
if(os.path.isfile(license_path)):
if(validate_key_Pair(license_path,full_key)):
return True,'Valid License'
elif(validate_key_Pair(license_path,oneYear_key)):
if months <= 12:
return True,'Valid License'
else:
return False,'License for AI.ON has expired. Please contact ERS Research for renewal.'
elif(validate_key_Pair(license_path,halfYear_key)):
if months <= 6:
return True,'Valid License'
else:
return False,'License for AI.ON has expired. Please contact ERS Research for renewal.'
elif(validate_key_Pair(license_path,quarter_key)):
if months <= 3:
return True,'Valid License'
else:
return False,'License for AI.ON has expired. Please contact ERS Research for renewal.'
else:
return False,'Invalid License'
else:
return False,'License Not exists.Please contact ERS Research for renewal.'
def checklicense():
import binascii
license_path = os.path.join(DATA_DIR,'License','license.lic')
if(os.path.isfile(license_path)):
try:
with open(license_path, 'r') as privatefile:
license_key = privatefile.read()
privatefile.close()
encrypted_message = binascii.unhexlify(license_key.encode())
privkey = '''-----BEGIN RSA PRIVATE KEY-----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-----END RSA PRIVATE KEY-----
'''
privkey = rsa.PrivateKey.load_pkcs1(privkey,'PEM')
decrypted_message = rsa.decrypt(encrypted_message, privkey)
msg = decrypted_message.decode().split('####')
product = msg[0]
computernameLicense = msg[1]
computername = socket.getfqdn()
licenseValid = False
if product.lower() == 'aion':
if computernameLicense == computername:
uuidlicense = msg[3]
uuid = guid()
if uuidlicense == uuid:
current_date = datetime.now()
license_expiry_date = msg[5]
license_expiry_date = datetime.strptime(license_expiry_date,'%Y-%m-%d %H:%M:%S')
if current_date > license_expiry_date:
return False,'License Expire'
else:
return True,''
return False,'License Error'
except Exception as e:
print(e)
return False,'License Error'
else:
return False,'Generate License'
def generate_record_key(product,version):
computername = socket.getfqdn()
macaddress = getmac.get_mac_address()
license_date = datetime.today().strftime('%Y-%m-%d %H:%M:%S')
try:
user = os.getlogin()
except:
user = 'NA'
uuid = guid()
msg = product+'###'+version+'###'+computername+'###'+macaddress+'###'+user+'###'+sys.platform+'###'+uuid+'###'+license_date
pkeydata='''-----BEGIN RSA PUBLIC KEY-----
MIIBCgKCAQEAm75ZwaepuxGJjU1Slk1+IUO2E49Hy8i9dym5FUaBRyTRH6R+GTF1
kcpd+1QinIZDMIdsmAc95Y8pTufxY30QxCkOhVASitSQWHS/IiWQHmsTJwdr38lq
ZnQQloOt/iPlhcavbxu/yKFzwBmp+nM+ErDTnCBh6EGCGrw1xWF30T2IBpmpWwME
oqZsFV69RzwQAw39KG1KCxi5uscrB62YPgUdlT2b4Yaa90egQhGLLVdnKvhPORiG
T9omCH90Dkm1oMMQ0Y2JBLezgXa/bunSqtTBxEwzlwUAX2JJcanFYrzKy2OLxzwN
RlWUXilZ4R/1RHAgUdNyKbYxZqc24MApoQIDAQAB
-----END RSA PUBLIC KEY-----
'''
pubkey = rsa.PublicKey.load_pkcs1(pkeydata)
encrypted_message = rsa.encrypt(msg.encode(), pubkey)
encrypted_message = binascii.hexlify(encrypted_message).decode()
return(encrypted_message)
def run(cmd):
try:
return subprocess.run(cmd, shell=True, capture_output=True, check=True, encoding="utf-8").stdout.strip()
except Exception as e:
print(e)
return None
def guid():
if sys.platform == 'darwin':
return run(
"ioreg -d2 -c IOPlatformExpertDevice | awk -F\\\" '/IOPlatformUUID/{print $(NF-1)}'",
)
if sys.platform == 'win32' or sys.platform == 'cygwin' or sys.platform == 'msys':
return run('wmic csproduct get uuid').split('\n')[2].strip()
if sys.platform.startswith('linux'):
return run('cat /var/lib/dbus/machine-id') or \
run('cat /etc/machine-id')
if sys.platform.startswith('openbsd') or sys.platform.startswith('freebsd'):
return run('cat /etc/hostid') or \
run('kenv -q smbios.system.uuid')
def updateLicense(licensekey):
license_folder = os.path.join(DATA_DIR,'License')
license_folder = Path(license_folder)
license_folder.mkdir(parents=True, exist_ok=True)
license_file = license_folder/'license.lic'
with open(license_file, "w") as fl:
fl.write(licensekey)
fl.close()
def enterRecord(version):
validLicense,msg = checklicense()
if not validLicense:
key = generate_record_key('AION',version)
msg = {'status':msg,'key':key,'licenseKey':'','link':''}
return validLicense,msg
|