app.py

# -*- coding: utf-8 -*-
"""app.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1XblbxoRxB4XOHixjGij789FPD9KjKdhi
"""

import os
import pdfplumber
import gradio as gr
from langchain_groq.chat_models import ChatGroq

# Set Groq API key securely
GROQ_API_KEY = os.getenv("GROQ_API_KEY")  # Fetch from environment variables
if not GROQ_API_KEY:
    raise ValueError("GROQ_API_KEY is not set. Add it in Hugging Face Secrets.")

# Initialize LLM (Mistral-8x7B)
llm = ChatGroq(model_name="mixtral-8x7b-32768")

def extract_text_from_pdf(pdf_file):
    """Extract text from a PDF file using pdfplumber."""
    text = ""
    with pdfplumber.open(pdf_file) as pdf:
        for page in pdf.pages:
            text += page.extract_text() or ""  # Extract text safely
    return text.strip()

def summarize_text(text, length, style):
    """Summarize text with adjustable length and style."""
    prompt = (
        f"""
        Read the following document and summarize it in {style.lower()} format.
        Keep the summary {length.lower()}.
        Follow this step-by-step approach:
        1. Identify key sections.
        2. Extract essential points.
        3. Remove unnecessary details.
        4. Ensure factual accuracy without adding extra information.

        Document:
        {text[:10000]}  # Limit input to 10,000 characters for efficiency
        """
    )
    response = llm.predict(prompt)
    return response.strip()

def process_pdf(file, length, style):
    """Extract text and summarize PDF using Mistral-8x7B with customization."""
    if not file:
        return "No file uploaded."

    text = extract_text_from_pdf(file.name)  # Extract text
    if not text:
        return "Could not extract text from the PDF. Try another file."

    return summarize_text(text, length, style)

# Create Gradio Interface
interface = gr.Interface(
    fn=process_pdf,
    inputs=[
        gr.File(label="Upload a PDF"),
        gr.Radio(["Short", "Medium", "Long"], label="Summary Length", value="Medium"),
        gr.Radio(["Bullets", "Key Takeaways", "Concise Paragraph"], label="Summary Style", value="Key Takeaways"),
    ],
    outputs="text",
    title="📄 Advanced PDF Summarizer",
    description="Upload a PDF file and get a structured summary with customization options."
)

# Run the app
interface.launch()