import streamlit as st import openai import os from dotenv import load_dotenv from PyPDF2 import PdfReader from langchain.text_splitter import CharacterTextSplitter from langchain.embeddings import OpenAIEmbeddings, HuggingFaceInstructEmbeddings from langchain.vectorstores import FAISS from langchain.memory import ConversationBufferMemory from langchain.chains import ConversationalRetrievalChain from langchain.chat_models import ChatOpenAI from htmlTemplates import css, bot_template, user_template from PIL import Image def get_pdf_text(pdf_docs): text = "" for pdf in pdf_docs: pdf_reader = PdfReader(pdf) for page in pdf_reader.pages: text += page.extract_text() return text # documentation for CharacterTextSplitter: # https://python.langchain.com/en/latest/modules/indexes/text_splitters/examples/character_text_splitter.html def get_text_chunk(text): text_splitter = CharacterTextSplitter( separator="\n", chunk_size = 1000, chunk_overlap = 200, length_function = len ) chunks = text_splitter.split_text(text) return chunks #embedding using openAI embedding. Warn: This will cost you money def get_vectorstore_openAI(text_chunks): embeddings = OpenAIEmbeddings() vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings) return vectorstore #embedding using instructor-xl with your local machine for free #you can find more details at: https://huggingface.co/hkunlp/instructor-xl def get_vectorstore(text_chunks): embeddings = HuggingFaceInstructEmbeddings(model_name="hkunlp/instructor-xl") vectorstore = FAISS.from_texts(texts=text_chunks, embedding=embeddings) return vectorstore def get_conversation_chain(vectorstore): llm = ChatOpenAI() memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True) conversation_chain = ConversationalRetrievalChain.from_llm( llm=llm, retriever=vectorstore.as_retriever(), memory = memory ) return conversation_chain def handle_userinput(user_question): response = st.session_state.conversation({'question': user_question}) st.session_state.chat_history = response['chat_history'] for i, message in enumerate(st.session_state.chat_history): if i%2 == 0: st.write(user_template.replace("{{MSG}}", message.content), unsafe_allow_html=True) else: st.write(bot_template.replace("{{MSG}}", message.content), unsafe_allow_html=True) def main(): ############################################################################## #load openai api_key from .evn # load_dotenv() openai.api_key = os.getenv("OPENAI_API_KEY") ############################################################################## #set up basic page st.set_page_config(page_title="Chat With multiple PDFs", page_icon=":books:") st.write(css, unsafe_allow_html=True) #initial session_state in order to avoid refresh if "conversation" not in st.session_state: st.session_state.conversation = None if "chat_history" not in st.session_state: st.session_state.chat_history = None st.header("Chat based on PDF you provided :books:") user_question = st.text_input("Ask a question about your documents:") if user_question: handle_userinput(user_question) # Define the templates with st.sidebar: st.subheader("Your PDF documents") pdf_docs = st.file_uploader("Upload your pdfs here and click on 'Proces'", accept_multiple_files= True) #if the button is pressed if st.button("Process"): with st.spinner("Processing"): #get pdf text raw_text = get_pdf_text(pdf_docs) print('raw_text is created') #get the text chunks text_chunks = get_text_chunk(raw_text) print('text_chunks are generated') #create vector store vectorstore = get_vectorstore_openAI(text_chunks) print('vectorstore is created') #create converstion chain st.session_state.conversation = get_conversation_chain(vectorstore) print('conversation chain created') # to run this application, you need to run "streamlit run app.py" if __name__ == '__main__': main()