app.py

import os
import gradio as gr
import faiss
import numpy as np
from PyPDF2 import PdfReader
from sentence_transformers import SentenceTransformer
from groq import Groq

# ✅ API Key from environment (set via Hugging Face "Secrets")
GROQ_API_KEY = os.getenv("GROQ_API_KEY")
if not GROQ_API_KEY:
    raise ValueError("GROQ_API_KEY not found in environment variables. Set it in the Space 'Secrets' section.")

client = Groq(api_key=GROQ_API_KEY)

# ✅ Global state
embedding_model = SentenceTransformer("all-MiniLM-L6-v2")
faiss_index = None
text_chunks = []

# ✅ Helper functions
def extract_text_from_pdf(pdf_file_obj):
    reader = PdfReader(pdf_file_obj)
    text = ""
    for page in reader.pages:
        page_text = page.extract_text()
        if page_text:
            text += page_text + "\n"
    return text

def split_into_chunks(text, chunk_size=500):
    words = text.split()
    return [" ".join(words[i:i+chunk_size]) for i in range(0, len(words), chunk_size)]

def process_pdf_and_query(pdf_file, question):
    global faiss_index, text_chunks

    # Step 1: Process PDF
    raw_text = extract_text_from_pdf(pdf_file)
    text_chunks = split_into_chunks(raw_text)
    embeddings = embedding_model.encode(text_chunks)
    dimension = embeddings.shape[1]
    faiss_index = faiss.IndexFlatL2(dimension)
    faiss_index.add(np.array(embeddings))

    # Step 2: Search FAISS index
    if not question.strip():
        return "PDF processed. Please enter a question."

    query_vector = embedding_model.encode([question])
    distances, indices = faiss_index.search(np.array(query_vector), k=3)
    context = "\n\n".join([text_chunks[i] for i in indices[0]])

    # Step 3: Use Groq API to get response
    response = client.chat.completions.create(
        model="llama3-8b-8192",
        messages=[
            {"role": "system", "content": "You are an assistant that summarizes and answers questions from documents."},
            {"role": "user", "content": f"{context}\n\nQuestion: {question}"}
        ]
    )

    return response.choices[0].message.content

# ✅ Gradio interface
iface = gr.Interface(
    fn=process_pdf_and_query,
    inputs=[
        gr.File(label="Upload PDF"),
        gr.Textbox(label="Ask a question about the PDF")
    ],
    outputs=gr.Textbox(label="Answer"),
    title="Document Q&A with Groq",
    description="Upload a PDF and ask a question. Powered by Groq + LLaMA 3."
)

iface.launch()