app.py

# import os
# from dotenv import load_dotenv
# load_dotenv()
# from langchain.document_loaders import PyPDFLoader
# from langchain.text_splitter import RecursiveCharacterTextSplitter
# from langchain.embeddings import HuggingFaceEmbeddings
# from langchain.vectorstores import FAISS
# from langchain.chains.question_answering import load_qa_chain
# from langchain_google_genai import ChatGoogleGenerativeAI
# from tkinter import Tk
# from tkinter.filedialog import askopenfilename

# # Hide the main tkinter window
# Tk().withdraw()

# # Open file dialog to select PDF
# pdf_path = askopenfilename(
#     title="Select a PDF File",
#     filetypes=[("PDF Files", "*.pdf")]
# )

# # Print the selected PDF path
# if pdf_path:
#     print("Selected PDF Path:")
#     print(pdf_path)
# else:
#     print("No PDF file selected.")


# # Step 1: Load pdf
# loader = PyPDFLoader(pdf_path)
# documents = loader.load()
# print("PDF successfully loaded....")

# # Step 2: Split into chunks
# text_splitter = RecursiveCharacterTextSplitter(chunk_size=500, chunk_overlap=50)

# docs = text_splitter.split_documents(documents)
# print('Chunks Created', len(docs))


# # Step 3: Create Embeddings
# embeddings = HuggingFaceEmbeddings(model_name='sentence-transformers/all-MiniLM-L6-v2')
# print('Embedding model loaded')

# vectorstore = FAISS.from_documents(docs, embeddings)
# print('Vector Database Created')


# # Step 4: Load Gemini Model
# llm = ChatGoogleGenerativeAI(
#     model = 'gemini-2.5-flash',
#     temperature = 0.3,
#     google_api_key = os.getenv("GOOGLE_API_KEY")
# )
# print("LLM loaded")


# # STep 5: Ask Question
# query = input('Ask Question :')

# matched_docs = vectorstore.similarity_search(query)

# chain = load_qa_chain(llm, chain_type='stuff')

# response = chain.run(input_documents=matched_docs, question=query)

# print('Response :')
# print(response)




import os
from dotenv import load_dotenv
load_dotenv()

import gradio as gr

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 langchain_google_genai import ChatGoogleGenerativeAI

# -----------------------------
# Global Vector Store
# -----------------------------
vectorstore = None

# -----------------------------
# Load Gemini Model
# -----------------------------
llm = ChatGoogleGenerativeAI(
    model="gemini-2.5-flash",
    temperature=0.3,
    google_api_key=os.getenv("GOOGLE_API_KEY")
)

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

# -----------------------------
# Process PDF
# -----------------------------
def process_pdf(pdf_file):
    global vectorstore

    if pdf_file is None:
        return "Please upload a PDF."

    try:
        # Load PDF
        loader = PyPDFLoader(pdf_file.name)
        documents = loader.load()

        # Split Text
        text_splitter = RecursiveCharacterTextSplitter(
            chunk_size=500,
            chunk_overlap=50
        )

        docs = text_splitter.split_documents(documents)

        # Create Vector Store
        vectorstore = FAISS.from_documents(
            docs,
            embeddings
        )

        return f"""
✅ PDF processed successfully

📄 Pages Loaded: {len(documents)}
🧩 Chunks Created: {len(docs)}
"""

    except Exception as e:
        return f"Error processing PDF:\n{str(e)}"

# -----------------------------
# Ask Question
# -----------------------------
def ask_question(query):
    global vectorstore

    if vectorstore is None:
        return "Please upload and process a PDF first."

    try:
        # Retrieve relevant docs
        docs = vectorstore.similarity_search(query, k=3)

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

        # Prompt
        prompt = f"""
Answer the question based only on the context below.

Context:
{context}

Question:
{query}
"""

        # Gemini Response
        response = llm.invoke(prompt)

        return response.content

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

# -----------------------------
# Gradio UI
# -----------------------------
with gr.Blocks(theme=gr.themes.Soft()) as demo:

    gr.Markdown(
        """
# 📚 PDF Question Answering System

Upload a PDF and ask questions from it using Gemini AI.
"""
    )

    with gr.Row():

        pdf_input = gr.File(
            label="Upload PDF",
            file_types=[".pdf"]
        )

        upload_btn = gr.Button(
            "Process PDF",
            variant="primary"
        )

    upload_output = gr.Textbox(
        label="PDF Status",
        lines=4
    )

    upload_btn.click(
        fn=process_pdf,
        inputs=pdf_input,
        outputs=upload_output
    )

    gr.Markdown("## ❓ Ask Questions")

    question_input = gr.Textbox(
        label="Enter your question",
        placeholder="What is this PDF about?"
    )

    ask_btn = gr.Button(
        "Ask Question",
        variant="primary"
    )

    answer_output = gr.Textbox(
        label="Answer",
        lines=10
    )

    ask_btn.click(
        fn=ask_question,
        inputs=question_input,
        outputs=answer_output
    )

# -----------------------------
# Launch
# -----------------------------
if __name__ == "__main__":
    demo.queue()
    demo.launch()