import streamlit as st from PyPDF2 import PdfReader from langchain.text_splitter import RecursiveCharacterTextSplitter from langchain.embeddings import HuggingFaceEmbeddings from langchain.vectorstores import FAISS from langchain.chains import RetrievalQA from langchain_groq import ChatGroq from dotenv import load_dotenv load_dotenv() # Initialize the language model llm = ChatGroq(model_name="llama-3.1-8b-instant", temperature=0.7) # Function to read PDF def read_pdf(pdf): raw_text = "" file = PdfReader(pdf) for page in file.pages: content = page.extract_text() if content: raw_text += content return raw_text # Function to process text and create a FAISS index def process_text(raw_text): text_splitter = RecursiveCharacterTextSplitter( chunk_size=200, chunk_overlap=20, length_function=len, separators=["\n"] ) texts = text_splitter.create_documents([raw_text]) modelPath = "sentence-transformers/all-MiniLM-l6-v2" model_kwargs = {'device': 'cpu'} encode_kwargs = {'normalize_embeddings': False} embeddings = HuggingFaceEmbeddings( model_name=modelPath, model_kwargs=model_kwargs, encode_kwargs=encode_kwargs ) db = FAISS.from_documents(texts, embeddings) return db # Function to initialize the QA chain def initialize_chain(db): chain = RetrievalQA.from_chain_type( llm=llm, chain_type="stuff", retriever=db.as_retriever() ) return chain # Main function to run the app def main(): st.title("PDF Question Answering") uploaded_file = st.file_uploader("Choose a PDF file", type="pdf") if uploaded_file is not None: raw_text = read_pdf(uploaded_file) db = process_text(raw_text) chain = initialize_chain(db) question = st.text_input("Ask a question about the PDF") if question: answer = chain.invoke(question) st.write(f"Answer: {answer['result']}") if __name__ == "__main__": main()