Olive_scrapper / app.py
arkamaldeen's picture
Updated the save button
5403de0
raw
history blame
14.2 kB
import justext
import streamlit as st
from lxml import etree
# import streamlit.components.v1 as components
# File Processing pkgs
from PIL import Image
import requests
# import xml.dom.minidom
from bs4 import BeautifulSoup
# import json
import docx2txt
# import textract
from PyPDF2 import PdfFileReader
import pdfplumber
import os
# ---- LOAD ASSETS ----
img_page_icon = Image.open("./olive_webscrapping.jpg")
# Find more emojis here: https://www.webfx.com/tools/emoji-cheat-sheet/
st.set_page_config(page_title="OdiaGenAI ", page_icon=img_page_icon, layout="wide")
# Load CSS file
def load_css(file_path):
with open(file_path) as f:
st.markdown(f"<style>{f.read()}</style>", unsafe_allow_html=True)
# Load CSS file
load_css('styles2.css')
# ----- FUNCTIONS ----
# function to check whether the url is a sitemap or not
def check_sitemap(url):
# Check the URL's ending
if url.lower().endswith(('sitemap.xml', 'sitemap_index.xml', 'sitemap')):
try:
# Parse the content as XML
response = requests.get(url)
xml_content = etree.fromstring(response.content)
# Check for sitemap-specific elements
if xml_content.tag == 'urlset' or xml_content.tag == 'sitemapindex':
return True
except etree.XMLSyntaxError:
pass
# Additional conditions for identifying sitemaps
if 'sitemap' in url.lower():
# Perform additional checks specific to the website's structure or naming conventions
return True
return False
# function to get urls from the site map and extract those data
def extract_urls_from_sitemaps(xml_url):
# Make a GET request to the URL and extract the xml content
response = requests.get(xml_url)
soup = BeautifulSoup(response.text, 'xml')
extracted_urls = []
# check if the sitemap contains nested sitemaps
sitemap_tags = soup.find_all('sitemap')
if sitemap_tags:
# Process nested sitemaps
for sitemap_tag in sitemap_tags:
print("sitemap_tags:" + sitemap_tag)
nested_url = sitemap_tag.find('loc').text
print('nested_url:', nested_url)
nested_urls = extract_urls_from_sitemaps(nested_url)
extracted_urls.extend(nested_urls)
else:
# Extract URLs from the current sitemap
loc_tags = soup.find_all('loc')
for loc_tag in loc_tags:
# if loc_tag.parent.name != 'image':
url = loc_tag.text
if url.endswith('.pdf') or url.endswith('.jpg') or url.endswith('.jpeg'):
print(f"url skipped because it is a {url.split('.')[-1]}")
else:
print('url:', url)
extracted_urls.append(url)
return extracted_urls
# function to check whether the entered url is valid
def valid_url(url):
try:
# Make a GET request to the URL and extract the text content
response = requests.get(url)
if response.status_code == 200:
return True
except requests.exceptions.RequestException as e:
return False
# function to create a custom stoplist for justext
def custom_stoplist():
odia_stopwords = [
"ଏହି", "ଏକ", "ଏକାଉଣଟ", "ମୁଁ", "ମୋର", "ମୁଁ ନିଜେ", "ଆମେ", "ଆମର", "ଆମର", "ଆମେ ନିଜେ", "ତୁମେ", "ତୁମର", "ତୁମର",
"ନିଜେ", "ନିଜେ", "ସେ", "ତାଙ୍କୁ", "ତାଙ୍କର",
"ନିଜେ", "ସେ", "ତାଙ୍କୁ", "ତାଙ୍କର", "ନିଜେ", "ଏହା", "ଏହାର", "ନିଜେ |", "ସେମାନେ", "ସେଗୁଡିକ", "ସେମାନଙ୍କର",
"ସେମାନଙ୍କର", "ନିଜେ |", "କଣ", "ଯାହା", "କିଏ", "କାହାକୁ",
"ଏହା", "ତାହା", "ଏଗୁଡ଼ିକ", "ସେଗୁଡ଼ିକ", "ମୁଁ", "ହେଉଛି", "ହେଉଛି |", "ଥିଲା", "ଥିଲା |", "ହୁଅ", "ହୋଇସାରିଛି |", "ହେବା",
"ଅଛି", "ଅଛି", "ଥିଲା", "ଅଛି", "କର", "କରେ |",
"କରିଛନ୍ତି", "କରିବା", "ଏବଂ", "କିନ୍ତୁ", "ଯଦି", "କିମ୍ବା", "କାରଣ", "ଯେପରି", "ପର୍ଯ୍ୟନ୍ତ", "ଯେତେବେଳେ", "ର", "ପାଇଁ",
"ସହିତ", "ବିଷୟରେ", "ବିପକ୍ଷରେ", "ମଧ୍ୟରେ", "ଭିତରକୁ", "ମାଧ୍ୟମରେ",
"ସମୟରେ", "ପୂର୍ବରୁ", "ପରେ", "ଉପରେ", "ନିମ୍ନରେ |", "କୁ", "ଠାରୁ", "ଅପ୍", "ତଳକୁ", "ଭିତରେ", "ବାହାରେ", "ଉପରେ", "ବନ୍ଦ",
"ସମାପ୍ତ", "ତଳେ |", "ପୁନର୍ବାର", "ଆଗକୁ",
"ତାପରେ", "ଥରେ |", "ଏଠାରେ", "ସେଠାରେ", "କେବେ", "କେଉଁଠାରେ", "କିପରି", "ସମସ୍ତ", "ଉଭୟ", "ପ୍ରତ୍ୟେକ", "ଅଳ୍ପ", "ଅଧିକ",
"ଅଧିକାଂଶ", "ଅନ୍ୟ", "କେତେକ", "ଏହିପରି",
"ନୁହେଁ |", "କେବଳ", "ନିଜର", "ସମାନ", "ତେଣୁ", "ଅପେକ୍ଷା", "ମଧ୍ୟ", "ବହୁତ", "କରିପାରିବେ |", "ଇଚ୍ଛା", "କେବଳ",
"କରିବା ଉଚିତ", "ବର୍ତ୍ତମାନ"
]
return frozenset(odia_stopwords)
# function to extract data from url using justext
def extract_data_from_url_(url):
response = requests.get(url)
response.raise_for_status()
page = response.content
data_url = ""
para = ""
paragraphs = justext.justext(page, custom_stoplist(), 70, 140, 0.0, 0.02, 0.5, 150, False)
for paragraph in paragraphs:
if not paragraph.is_boilerplate:
para = para + '\n' + paragraph.text
data_url = ('\n\nFrom url:' + url + '\n' + para + '\n')
return data_url
sitemap_data = ""
# function to get the text from pdf using PyPDF2
def read_pdf(file):
pdfReader = PdfFileReader(file)
count = pdfReader.numPages
# all_page_text = ""
# for i in range(count):
# page = pdfReader.getPage(i)
# all_page_text += page.extractText()
#
# return all_page_text
return count
# function to run the enter button
def run_function(url, documents):
data = ""
# Check if the user has provided a URL
if url:
if valid_url(url):
data = extract_data_from_url_(url)
st.text_area("Extracted Text", value=data, height=200)
# return extract status, and the data extracted
return True, data
else:
return False, data
# Check if the user has provided a document
elif documents is not None:
for document in documents:
document_details = {
"filename": document.name,
"filetype": document.type,
"filesize": document.size
}
st.write(document_details)
# Extract content from the txt file
if document.type == "text/plain":
# Read as bytes
data += str(document.read(), "utf-8")
# Extract content from the pdf file
elif document.type == "application/pdf":
# using PyPDF2
# data += read_pdf(document)
# using pdfplumber
try:
with pdfplumber.open(document) as pdf:
all_text = ""
for page in pdf.pages:
text = page.extract_text()
all_text += text + "\n"
data += all_text
except requests.exceptions.RequestException as e:
st.write("None")
# Extract content from the docx file
elif document.type == "application/vnd.openxmlformats-officedocument.wordprocessingml.document":
data += docx2txt.process(document)
# Display the extracted text content from file
st.write("attached")
st.text_area("Extracted Text", value=data, height=200)
# return extract status, and the data extracted
return True, data
else:
st.error("Error: An error occurred while fetching content.")
# return extract status, and the data extracted
return False, data
def main():
# ---- HEADER SECTION ----
with st.container():
st.subheader("Hi!! :wave:")
st.write("##")
st.markdown("<h5 class='text'>OdiaGenAI is a collaborative initiative that conducts research on </h5>",
unsafe_allow_html=True)
st.markdown("<h5>Generative AI and LLM for the Odia Language.</h5>", unsafe_allow_html=True)
# st.title("Odia Generative AI")
st.markdown("<h1 class='title'>Odia Generative AI</h1>", unsafe_allow_html=True)
# ---- BODY SECTION ----
with st.container():
st.subheader("Collecting monolingual data (Odia or any Indic Languages)")
# dividing the body section into 3 columns for url, attach button and enter button
col1, col2, col3 = st.columns([0.6, 0.2, 0.2])
# url/xml
with col1:
url_or_xml = st.text_input(label='', placeholder="Enter URL")
is_a_sitemap = check_sitemap(url_or_xml)
# attached files
with col2:
documents = st.file_uploader("", type=["pdf", "txt", "docx"], accept_multiple_files=True)
if not documents:
documents = None
else:
for doc in documents:
if doc.name.split(".")[-1].lower() not in ["pdf", "txt", "docx"]:
# if documents is not the relevant type
st.error("Unsupported file: " + doc.name)
# Initialize state of button Enter
with col3:
st.write('##')
if "button_enter" not in st.session_state:
st.session_state.button_enter = False
if st.button("Enter"):
st.session_state.button_enter = True
# st.write("session state true")
if "extracted" not in st.session_state:
st.session_state.extracted = False
data = ""
# the enter button
if st.session_state.button_enter:
# check if it is a sitemap or not
if is_a_sitemap:
if "Initial" not in st.session_state:
st.session_state.Initial = True
# check whether its the initial state
if st.session_state.Initial == True:
# print("\n\n\n\n1)Initial State", st.session_state.Initial, "\n\n\n\n\n")
xml = url_or_xml
st.write("It is a sitemap")
stored_sitemap_urls = extract_urls_from_sitemaps(xml)
print('\nno. of urls: ', len(stored_sitemap_urls))
if stored_sitemap_urls:
print(stored_sitemap_urls)
for sitemap_url in stored_sitemap_urls:
if valid_url(sitemap_url):
print(sitemap_url)
# using justext to extract data
data = data + extract_data_from_url_(sitemap_url)
else:
st.error("Couldnt extract data from " + sitemap_url)
if "sitemap_data" not in st.session_state:
st.session_state.sitemap_data = data
# print("\n\n\nst.session.data ", st.session_state.sitemap_data)
# print("\n\n\n\nRUNNING \n\n\n\n")
st.session_state.Initial = False
print("\n\n\n\n2)Initial State", st.session_state.Initial, "\n\n\n\n\n")
st.session_state.extracted = True
# st.text_area("Extracted Text", value=st.session_state.sitemap_data, height=300)
else:
st.error("Error: Invalid sitemap.")
else:
url = url_or_xml
st.session_state.extracted, data = run_function(url, documents)
if st.session_state.extracted:
if is_a_sitemap:
st.text_area("Extracted Text", value=st.session_state.sitemap_data, height=300)
col1, col2 = st.columns([0.5, 0.5])
with col1:
saved_button = False
if is_a_sitemap:
saved_data = st.session_state.sitemap_data
if st.download_button(
label="Save",
data=saved_data
):
saved_button = True
else:
if st.download_button(
label="Save",
data=data
):
saved_button = True
with col2:
if st.button("Clear"):
st.session_state.button_enter = False
st.session_state.Initial = True
st.session_state.extracted = False
if 'sitemap_data' in st.session_state:
del st.session_state['sitemap_data']
st.session_state.button_enter = False
st.experimental_rerun()
if saved_button:
# Confirmation message
st.success(f"File saved as {file_name} in the current directory.")
else:
st.warning("Data not extracted")
if __name__ == "__main__":
main()