app.py

import os
import logging
import faiss
import streamlit as st
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM, pipeline
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain.vectorstores import FAISS
from langchain_community.llms import HuggingFacePipeline
from langchain.chains import RetrievalQA

# Set up logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# HuggingFace model checkpoint
checkpoint = "LaMini-T5-738M"

@st.cache_resource
def load_llm():
    """Load the language model for text generation."""
    tokenizer = AutoTokenizer.from_pretrained(checkpoint)
    model = AutoModelForSeq2SeqLM.from_pretrained(checkpoint)
    pipe = pipeline(
        'text2text-generation',
        model=model,
        tokenizer=tokenizer,
        max_length=256,
        do_sample=True,
        temperature=0.3,
        top_p=0.95
    )
    return HuggingFacePipeline(pipeline=pipe)

def load_faiss_index():
    """Load the FAISS index for vector search."""
    index_path = "faiss_index/index.faiss"
    if not os.path.exists(index_path):
        st.error(f"FAISS index not found at {index_path}. Please ensure the file exists.")
        raise RuntimeError(f"FAISS index not found at {index_path}.")

    try:
        index = faiss.read_index(index_path)
        logger.info(f"FAISS index loaded successfully from {index_path}")
        embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
        retriever = FAISS(index, embeddings)
        return retriever
    except Exception as e:
        st.error(f"Failed to load FAISS index: {e}")
        logger.exception("Exception in load_faiss_index")
        raise

def process_answer(instruction):
    """Process the user's question using the QA system."""
    try:
        retriever = load_faiss_index()
        llm = load_llm()
        qa = RetrievalQA.from_chain_type(
            llm=llm,
            chain_type="stuff",
            retriever=retriever,
            return_source_documents=True
        )
        generated_text = qa.invoke(instruction)
        answer = generated_text['result']
        return answer, generated_text
    except Exception as e:
        st.error(f"An error occurred while processing the answer: {e}")
        logger.exception("Exception in process_answer")
        return "An error occurred while processing your request.", {}

def main():
    """Main function to run the Streamlit application."""
    st.title("Search Your PDF 📚📝")
    
    with st.expander("About the App"):
        st.markdown(
            """
            This is a Generative AI powered Question and Answering app that responds to questions about your PDF File.
            """
        )

    question = st.text_area("Enter your Question")
    
    if st.button("Ask"):
        st.info("Your Question: " + question)
        st.info("Your Answer")
        try:
            answer, metadata = process_answer(question)
            st.write(answer)
            st.write(metadata)
        except Exception as e:
            st.error(f"An unexpected error occurred: {e}")
            logger.exception("Unexpected error in main function")

if __name__ == '__main__':
    main()