app.py

import os
import streamlit as st
import faiss
import tempfile
from pypdf import PdfReader
from sentence_transformers import SentenceTransformer
from groq import Groq
import numpy as np

# Load environment key
GROQ_API_KEY = os.environ.get(GROQ_API_KEY)

# Load embedding model
embed_model = SentenceTransformer('all-MiniLM-L6-v2')

# Set up Groq client
client = Groq(api_key=GROQ_API_KEY)

# Functions
def extract_text_from_pdf(pdf_file):
    reader = PdfReader(pdf_file)
    return "\n".join(page.extract_text() for page in reader.pages)

def chunk_text(text, chunk_size=500, overlap=50):
    words = text.split()
    chunks = []
    for i in range(0, len(words), chunk_size - overlap):
        chunk = " ".join(words[i:i + chunk_size])
        chunks.append(chunk)
    return chunks

def embed_chunks(chunks):
    return embed_model.encode(chunks)

def create_faiss_index(embeddings):
    dimension = embeddings.shape[1]
    index = faiss.IndexFlatL2(dimension)
    index.add(embeddings)
    return index

def query_groq(query, context):
    messages = [
        {"role": "system", "content": "You are a helpful assistant that answers questions based on the given context."},
        {"role": "user", "content": f"Context: {context}\n\nQuestion: {query}"}
    ]
    chat_completion = client.chat.completions.create(
        messages=messages,
        model="llama3-8b-8192",
    )
    return chat_completion.choices[0].message.content

# Streamlit UI
st.title("RAG PDF Q&A App (Groq + FAISS)")

uploaded_file = st.file_uploader("Upload your PDF", type="pdf")
query = st.text_input("Ask a question about the PDF")

if uploaded_file:
    with tempfile.NamedTemporaryFile(delete=False) as tmp:
        tmp.write(uploaded_file.read())
        text = extract_text_from_pdf(tmp.name)
    
    chunks = chunk_text(text)
    embeddings = embed_chunks(chunks)
    index = create_faiss_index(np.array(embeddings))

    if query:
        query_embedding = embed_model.encode([query])
        distances, indices = index.search(np.array(query_embedding), k=3)
        context = "\n\n".join([chunks[i] for i in indices[0]])
        response = query_groq(query, context)
        st.subheader("Answer:")
        st.write(response)