app.py

# app.py

from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.vectorstores import Chroma
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.document_loaders import PyPDFLoader
from langchain.chains import RetrievalQA
from langchain.llms.base import LLM

from typing import List, Optional
from groq import Groq
import gradio as gr
import os
import uuid

from dotenv import load_dotenv
load_dotenv()


# ✅ Groq LLM Wrapper
class GroqLLM(LLM):
    model: str = "llama3-8b-8192"
    api_key: str = os.environ.get("GROQ_API_KEY")  # Use env variable for safety
    temperature: float = 0.0

    def _call(self, prompt: str, stop: Optional[List[str]] = None) -> str:
        client = Groq(api_key=self.api_key)
        messages = [
            {"role": "system", "content": "You are a helpful assistant."},
            {"role": "user", "content": prompt}
        ]
        response = client.chat.completions.create(
            model=self.model,
            messages=messages,
            temperature=self.temperature,
        )
        return response.choices[0].message.content

    @property
    def _llm_type(self) -> str:
        return "groq-llm"

# ✅ Session Store
session_store = {}

# ✅ Process PDF File
def process_pdf_and_setup_chain(pdf_file):
    if not pdf_file:
        return "❌ No PDF uploaded."

    file_path = pdf_file.name
    temp_dir = f"temp_{uuid.uuid4().hex}"
    os.makedirs(temp_dir, exist_ok=True)

    try:
        loader = PyPDFLoader(file_path)
        documents = loader.load()

        splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
        docs = splitter.split_documents(documents)

        embedding = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
        vectorstore = Chroma.from_documents(docs, embedding, persist_directory=os.path.join(temp_dir, "chroma"))

        retriever = vectorstore.as_retriever()
        groq_llm = GroqLLM()

        qa_chain = RetrievalQA.from_chain_type(
            llm=groq_llm,
            retriever=retriever,
            return_source_documents=True
        )

        session_store["qa_chain"] = qa_chain
        session_store["temp_dir"] = temp_dir

        return "✅ PDF processed! You can now ask questions."

    except Exception as e:
        return f"❌ Error: {str(e)}"

# ✅ Answering Function
def answer_question(query):
    qa_chain = session_store.get("qa_chain")
    if not qa_chain:
        return "❌ Please upload and process a PDF first."
    if not query.strip():
        return "❗ Please enter a question."
    try:
        result = qa_chain({"query": query})
        return result["result"]
    except Exception as e:
        return f"❌ Error: {str(e)}"

# ✅ Gradio UI
with gr.Blocks() as demo:
    gr.Markdown("## 📄 PDF Q&A with LangChain + Groq LLaMA3")
    gr.Markdown("Upload a PDF, process it, and ask any question from its content.")

    with gr.Row():
        pdf_input = gr.File(label="📄 Upload PDF", file_types=[".pdf"])
        process_btn = gr.Button("⚙️ Process PDF")

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

    with gr.Row():
        question = gr.Textbox(label="Ask a question", lines=2, placeholder="e.g. What is the document about?")
        ask_btn = gr.Button("🔍 Ask")

    answer = gr.Textbox(label="Answer", interactive=False)

    process_btn.click(fn=process_pdf_and_setup_chain, inputs=pdf_input, outputs=status)
    ask_btn.click(fn=answer_question, inputs=question, outputs=answer)

demo.launch()