app.py

import os
import faiss
import streamlit as st
from PyPDF2 import PdfReader
from sentence_transformers import SentenceTransformer
from groq import Groq
from dotenv import load_dotenv
import requests
from io import BytesIO

# Predefined Google Drive links
PDF_LINKS = [
    "https://drive.google.com/uc?id=1JPf0XvDhn8QoDOlZDrxCOpu4WzKFESNz",
    # Add more Google Drive links here
]

# Initialize Groq client
client = Groq(api_key="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")

# Load Sentence Transformer model
model = SentenceTransformer("all-MiniLM-L6-v2")

# Initialize FAISS
dimension = 384  # Embedding size for the Sentence Transformer model
index = faiss.IndexFlatL2(dimension)

# Store chunks globally
stored_chunks = []

# Function to download and extract the PDF content
def download_and_process_pdf(link):
    response = requests.get(link)
    if response.status_code == 200:
        pdf_reader = PdfReader(BytesIO(response.content))
        text = ""
        for page in pdf_reader.pages:
            text += page.extract_text()
        chunks = [text[i:i + 500] for i in range(0, len(text), 500)]  # Chunk into 500-char blocks
        embeddings = model.encode(chunks)
        index.add(embeddings)
        stored_chunks.extend(chunks)
    else:
        print(f"Failed to download PDF from link: {link}")

# Process all predefined links
for link in PDF_LINKS:
    download_and_process_pdf(link)

# Function to query FAISS and generate a response
def query_model(query):
    query_vector = model.encode([query])
    _, indices = index.search(query_vector, k=3)  # Top 3 similar chunks
    response_chunks = [stored_chunks[idx] for idx in indices[0]]
    context = " ".join(response_chunks)

    # Groq API call
    chat_completion = client.chat.completions.create(
        messages=[
            {
                "role": "user",
                "content": f"Context: {context}\n\nQuery: {query}",
            }
        ],
        model="llama3-8b-8192",
    )
    return chat_completion.choices[0].message.content

# Streamlit app
st.title("RAG-based PDF Question Answering")
st.write("Preloaded documents from Google Drive are ready for querying.")

query = st.text_input("Ask a question:")
if query:
    answer = query_model(query)
    st.write("### Answer:")
    st.write(answer)