smu-ai / ingest.py
inflaton's picture
Duplicate from inflaton/learn-ai
3ac9dae
raw
history blame
4.14 kB
# setting device on GPU if available, else CPU
import os
from timeit import default_timer as timer
from typing import List
from langchain.document_loaders import PyPDFDirectoryLoader
from langchain.embeddings import HuggingFaceInstructEmbeddings
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.vectorstores.base import VectorStore
from langchain.vectorstores.chroma import Chroma
from langchain.vectorstores.faiss import FAISS
from app_modules.init import *
def load_documents(source_pdfs_path, urls) -> List:
loader = PyPDFDirectoryLoader(source_pdfs_path, silent_errors=True)
documents = loader.load()
if urls is not None and len(urls) > 0:
for doc in documents:
source = doc.metadata["source"]
filename = source.split("/")[-1]
for url in urls:
if url.endswith(filename):
doc.metadata["url"] = url
break
return documents
def split_chunks(documents: List, chunk_size, chunk_overlap) -> List:
text_splitter = RecursiveCharacterTextSplitter(
chunk_size=chunk_size, chunk_overlap=chunk_overlap
)
return text_splitter.split_documents(documents)
def generate_index(
chunks: List, embeddings: HuggingFaceInstructEmbeddings
) -> VectorStore:
if using_faiss:
faiss_instructor_embeddings = FAISS.from_documents(
documents=chunks, embedding=embeddings
)
faiss_instructor_embeddings.save_local(index_path)
return faiss_instructor_embeddings
else:
chromadb_instructor_embeddings = Chroma.from_documents(
documents=chunks, embedding=embeddings, persist_directory=index_path
)
chromadb_instructor_embeddings.persist()
return chromadb_instructor_embeddings
# Constants
device_type, hf_pipeline_device_type = get_device_types()
hf_embeddings_model_name = (
os.environ.get("HF_EMBEDDINGS_MODEL_NAME") or "hkunlp/instructor-xl"
)
index_path = os.environ.get("FAISS_INDEX_PATH") or os.environ.get("CHROMADB_INDEX_PATH")
using_faiss = os.environ.get("FAISS_INDEX_PATH") is not None
source_pdfs_path = os.environ.get("SOURCE_PDFS_PATH")
source_urls = os.environ.get("SOURCE_URLS")
chunk_size = os.environ.get("CHUNCK_SIZE")
chunk_overlap = os.environ.get("CHUNK_OVERLAP")
start = timer()
embeddings = HuggingFaceInstructEmbeddings(
model_name=hf_embeddings_model_name, model_kwargs={"device": device_type}
)
end = timer()
print(f"Completed in {end - start:.3f}s")
start = timer()
if not os.path.isdir(index_path):
print(
f"The index persist directory {index_path} is not present. Creating a new one."
)
os.mkdir(index_path)
if source_urls is not None:
# Open the file for reading
file = open(source_urls, "r")
# Read the contents of the file into a list of strings
lines = file.readlines()
# Close the file
file.close()
# Remove the newline characters from each string
source_urls = [line.strip() for line in lines]
print(
f"Loading {'' if source_urls is None else str(len(source_urls)) + ' '}PDF files from {source_pdfs_path}"
)
sources = load_documents(source_pdfs_path, source_urls)
print(f"Splitting {len(sources)} PDF pages in to chunks ...")
chunks = split_chunks(
sources, chunk_size=int(chunk_size), chunk_overlap=int(chunk_overlap)
)
print(f"Generating index for {len(chunks)} chunks ...")
index = generate_index(chunks, embeddings)
else:
print(f"The index persist directory {index_path} is present. Loading index ...")
index = (
FAISS.load_local(index_path, embeddings)
if using_faiss
else Chroma(embedding_function=embeddings, persist_directory=index_path)
)
query = "hi"
print(f"Load relevant documents for standalone question: {query}")
start2 = timer()
docs = index.as_retriever().get_relevant_documents(query)
end = timer()
print(f"Completed in {end - start2:.3f}s")
print(docs)
end = timer()
print(f"Completed in {end - start:.3f}s")