app.py

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

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

checkpoint = "LaMini-T5-738M"

@st.cache_resource
def load_llm():
    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 validate_index_file(index_path):
    try:
        if os.path.getsize(index_path) == 0:
            st.error(f"Index file '{index_path}' is empty.")
            return False
        with open(index_path, 'rb') as f:
            data = f.read(100)
        logger.info(f"Successfully read {len(data)} bytes from the index file")
        return True
    except Exception as e:
        logger.error(f"Error validating index file: {e}")
        return False

def load_faiss_index():
    index_path = "faiss_index/index.faiss"
    
    if not os.path.exists(index_path) or not validate_index_file(index_path):
        st.warning("Index file is missing or corrupted. Creating a new one...")
        if os.path.exists(index_path):
            os.remove(index_path)
            st.info("Deleted the corrupted index file.")
        create_faiss_index()

    if not os.path.exists(index_path):
        st.error("Failed to create the FAISS index. Please check the 'docs' directory and try again.")
        raise RuntimeError("FAISS index creation failed.")

    try:
        index = faiss.read_index(index_path)
        if index is None:
            raise ValueError("Failed to read FAISS index.")
        
        embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
        db = FAISS.load_local("faiss_index", embeddings)
        if db.index is None or db.index_to_docstore_id is None:
            raise ValueError("FAISS index or docstore_id mapping is None.")

        return db.as_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):
    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 diagnose_faiss_index():
    index_path = "faiss_index/index.faiss"
    if os.path.exists(index_path):
        st.write(f"Index file size: {os.path.getsize(index_path)} bytes")
        st.write(f"Index file permissions: {oct(os.stat(index_path).st_mode)[-3:]}")
        st.write(f"Index file owner: {os.stat(index_path).st_uid}")
        st.write(f"Current process user ID: {os.getuid()}")
        validate_index_file(index_path)
    else:
        st.warning("Index file does not exist.")

def main():
    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.
            """
        )

    diagnose_faiss_index()

    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()