app.py

import os
import gradio as gr
import tempfile
from dotenv import load_dotenv

from langchain_community.document_loaders import PyPDFLoader
from langchain_text_splitters import RecursiveCharacterTextSplitter
from langchain_huggingface import HuggingFaceEmbeddings
from langchain_community.vectorstores import FAISS

from groq import Groq

# ================== LOAD ENV ==================
load_dotenv()
GROQ_API_KEY = os.getenv("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")

if not GROQ_API_KEY:
    raise ValueError("❌ GROQ_API_KEY not found. Please set it in Hugging Face Secrets.")

client = Groq(api_key=GROQ_API_KEY)

# ================== GLOBAL VECTOR DB ==================
vector_db = None

# ================== LLM FUNCTION ==================
def groq_llm(prompt):
    response = client.chat.completions.create(
        model="llama-3.3-70b-versatile",
        messages=[{"role": "user", "content": prompt}],
    )
    return response.choices[0].message.content


# ================== PDF PROCESSING ==================
def process_pdf(file):
    global vector_db

    if file is None:
        return "❌ Please upload a PDF file."

    # Save uploaded file
    with tempfile.NamedTemporaryFile(delete=False, suffix=".pdf") as tmp:
        tmp.write(file.read())
        pdf_path = tmp.name

    # Load PDF
    loader = PyPDFLoader(pdf_path)
    documents = loader.load()

    # Split text into chunks
    splitter = RecursiveCharacterTextSplitter(
        chunk_size=500,
        chunk_overlap=100
    )
    docs = splitter.split_documents(documents)

    # Create embeddings
    embeddings = HuggingFaceEmbeddings(
        model_name="sentence-transformers/all-MiniLM-L6-v2"
    )

    # Create vector database
    vector_db = FAISS.from_documents(docs, embeddings)

    return f"✅ Document processed successfully! {len(docs)} chunks created."


# ================== QUESTION ANSWERING ==================
def ask_question(question):
    global vector_db

    if vector_db is None:
        return "❌ Please upload and process a document first."

    retriever = vector_db.as_retriever(search_kwargs={"k": 3})
    docs = retriever.invoke(question)

    context = "\n\n".join([doc.page_content for doc in docs])

    prompt = f"""
You are an intelligent assistant.
Use ONLY the context below to answer the question.

Context:
{context}

Question:
{question}

Answer:
"""

    return groq_llm(prompt)


# ================== GRADIO UI ==================
with gr.Blocks(title="📄 RAG PDF Question Answering App") as demo:
    gr.Markdown("# 📄 RAG PDF Question Answering App")
    gr.Markdown("Upload a PDF and ask questions about it.")

    with gr.Row():
        pdf_upload = gr.File(label="Upload PDF", file_types=[".pdf"])
        process_btn = gr.Button("📥 Process Document")

    status = gr.Textbox(label="Status", interactive=False)

    with gr.Row():
        question = gr.Textbox(label="Ask a Question")
        answer = gr.Textbox(label="Answer", interactive=False)

    process_btn.click(process_pdf, inputs=pdf_upload, outputs=status)
    question.submit(ask_question, inputs=question, outputs=answer)

demo.launch()