app.py

import os
from typing import List, Optional

import gradio as gr
from langchain_community.embeddings import HuggingFaceEmbeddings
from langchain_community.vectorstores import Chroma
from langchain.text_splitter import RecursiveCharacterTextSplitter
from langchain.document_loaders import TextLoader
from langchain.docstore.document import Document
from langchain.chains import RetrievalQA
from langchain.llms.base import LLM
from groq import Groq
import pypdf  # PyMuPDF


# --- Custom LLM class using Groq ---
class GroqLLM(LLM):
    model: str = "llama3-8b-8192"
    api_key: str = "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"  # <-- PUT YOUR GROQ API KEY HERE
    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"


# --- RAG Setup ---
retriever = None
qa_chain = None


def extract_text_from_pdf(file_path: str) -> str:
    doc = fitz.open(file_path)
    text = ""
    for page in doc:
        text += page.get_text()
    doc.close()
    return text


def process_file(file_obj):
    global retriever, qa_chain

    ext = os.path.splitext(file_obj.name)[1].lower()
    try:
        # Load content
        if ext == ".pdf":
            text = extract_text_from_pdf(file_obj.name)
        elif ext == ".txt":
            with open(file_obj.name, "r", encoding="utf-8") as f:
                text = f.read()
        else:
            return "❌ Unsupported file format. Please upload a .txt or .pdf file."

        # Create document chunks
        document = Document(page_content=text)
        splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)
        docs = splitter.split_documents([document])

        # Vectorstore with HuggingFace embeddings
        embeddings = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
        vectorstore = Chroma.from_documents(docs, embedding=embeddings, persist_directory="rag_store")

        retriever = vectorstore.as_retriever()
        qa_chain = RetrievalQA.from_chain_type(
            llm=GroqLLM(),
            retriever=retriever,
            return_source_documents=True
        )

        return "✅ File processed successfully. You can now ask questions."

    except Exception as e:
        return f"❌ Error processing file: {e}"


def ask_question(query):
    if qa_chain is None:
        return "⚠ Please upload a file first."
    result = qa_chain({"query": query})
    return result["result"]


# --- Gradio UI ---
with gr.Blocks(title="RAG PDF & Text Chatbot") as demo:
    gr.Markdown("## 🧠 RAG-powered Q&A Chatbot (Groq + LangChain)")
    gr.Markdown("Upload a .pdf or .txt file and ask questions based on its content.")

    file_input = gr.File(label="Upload PDF or Text File", file_types=[".pdf", ".txt"])
    upload_status = gr.Textbox(label="Status", interactive=False)

    file_input.change(fn=process_file, inputs=file_input, outputs=upload_status)

    question_box = gr.Textbox(label="Ask your question")
    answer_box = gr.Textbox(label="Answer", interactive=False)

    submit_btn = gr.Button("Get Answer")
    submit_btn.click(fn=ask_question, inputs=question_box, outputs=answer_box)

demo.launch()