app.py

import gradio as gr
import os
import fitz  # PyMuPDF for PDF extraction
import requests
from bs4 import BeautifulSoup
from pptx import Presentation
from docx import Document  # For Word files
from groq import Groq  # Assuming Groq is available as a Python library

# Initialize Groq Client using environment variable
api_key = os.getenv('MCQs')  # Retrieve API key from environment
if api_key:
    client = Groq(api_key=api_key)
else:
    raise ValueError("API Key not found in environment variables!")

# Step 1: File Extraction Functions
def extract_text_from_pdf(file):
    pdf_text = ""
    try:
        pdf_file = fitz.open(file.name)
        for page_num in range(pdf_file.page_count):
            page = pdf_file.load_page(page_num)
            pdf_text += page.get_text()
    except Exception as e:
        return f"Error reading PDF: {e}"
    return pdf_text

def extract_text_from_ppt(file):
    ppt_text = ""
    try:
        presentation = Presentation(file.name)
        for slide in presentation.slides:
            for shape in slide.shapes:
                if hasattr(shape, 'text'):
                    ppt_text += shape.text + "\n"
    except Exception as e:
        return f"Error reading PPT: {e}"
    return ppt_text

def extract_text_from_word(file):
    doc_text = ""
    try:
        document = Document(file.name)
        for paragraph in document.paragraphs:
            doc_text += paragraph.text + "\n"
    except Exception as e:
        return f"Error reading Word file: {e}"
    return doc_text

def extract_text_from_url(url):
    page_text = ""
    try:
        response = requests.get(url)
        soup = BeautifulSoup(response.content, 'html.parser')
        page_text = soup.get_text(separator="\n", strip=True)
    except Exception as e:
        return f"Error reading URL: {e}"
    return page_text

def process_files(files, url):
    text_data = ""
    for file in files:
        if file.name.endswith(".pdf"):
            text_data += extract_text_from_pdf(file)
        elif file.name.endswith(".pptx"):
            text_data += extract_text_from_ppt(file)
        elif file.name.endswith(".docx"):
            text_data += extract_text_from_word(file)
        else:
            text_data += "Unsupported file format: " + file.name + "\n"  # Inform user of unsupported format
    if url:
        text_data += extract_text_from_url(url)
    return text_data

# Step 2: Generate MCQs and Subjective Questions Using Groq
def generate_questions(text, num_mcqs=5, num_subjective=2, difficulty_mcqs="medium", difficulty_subjective="medium"):
    try:
        num_mcqs = min(num_mcqs, 40)  # Limit MCQs to 40
        num_subjective = min(num_subjective, 20)  # Limit Subjective Questions to 20

        difficulty_levels = {
            "easy": "simple questions with direct answers.",
            "medium": "moderate complexity questions requiring reasoning.",
            "hard": "challenging questions requiring deep understanding."
        }

        prompt = f"Generate {num_mcqs} multiple choice questions and {num_subjective} subjective questions from the following text: {text}. Include the correct answers for each question. The questions should be {difficulty_levels.get(difficulty_mcqs, 'medium')} for MCQs and {difficulty_levels.get(difficulty_subjective, 'medium')} for Subjective questions."

        chat_completion = client.chat.completions.create(
            messages=[{"role": "user", "content": prompt}],
            model="llama3-8b-8192",
        )

        response = chat_completion.choices[0].message.content.strip()

        # Split response based on question types
        mcqs = ""
        subjective = ""

        # Initialize the flag for the subjective section
        is_subjective_section = False

        # Split the MCQs and Subjective questions based on content patterns
        lines = response.split("\n")

        for line in lines:
            if line.strip().startswith("**Subjective Questions**"):  # Identify the start of the subjective section
                is_subjective_section = True
                subjective += line + "\n"  # Add the header to the subjective section as well
            elif is_subjective_section:
                subjective += line + "\n"
            else:
                mcqs += line + "\n"

        return mcqs, subjective

    except Exception as e:
        return f"Error generating questions: {e}", ""


# Step 3: Gradio Interface
def process_and_generate(files, raw_text, url, num_mcqs=5, num_subjective=2, difficulty_mcqs="medium", difficulty_subjective="medium"):
    extracted_text = ""

    # Extract and process all input files
    if files:
        for file in files:
            if file.name.endswith(".pdf"):
                extracted_text += extract_text_from_pdf(file) + "\n"
            elif file.name.endswith(".pptx"):
                extracted_text += extract_text_from_ppt(file) + "\n"
            elif file.name.endswith(".docx"):
                extracted_text += extract_text_from_word(file) + "\n"
            else:
                extracted_text += f"Unsupported file format: {file.name}\n"

    # Append raw text input if provided
    if raw_text:
        extracted_text += raw_text + "\n"

    # Extract text from URL if provided
    if url:
        extracted_text += extract_text_from_url(url) + "\n"

    # Check if there's any extracted text to process
    if not extracted_text.strip():
        return "No valid input provided to generate questions.", ""

    # Generate questions from the extracted text
    mcqs, subjective = generate_questions(extracted_text, num_mcqs, num_subjective, difficulty_mcqs, difficulty_subjective)
    return mcqs, subjective


# Gradio Inputs
inputs = [
    gr.File(file_count="multiple", type="filepath"),  # Corrected to 'filepath'
    gr.Textbox(lines=2, placeholder="Enter raw text here...", label="Raw Text (Optional)"),
    gr.Textbox(lines=1, placeholder="Enter URL here...", label="URL (Optional)"),
    gr.Slider(minimum=1, maximum=40, value=5, step=1, label="Number of MCQs (Max 40)"),
    gr.Slider(minimum=1, maximum=20, value=2, step=1, label="Number of Subjective Questions (Max 20)"),
    gr.Radio(["easy", "medium", "hard"], label="Select Difficulty Level for MCQs", value="medium"),
    gr.Radio(["easy", "medium", "hard"], label="Select Difficulty Level for Subjective Questions", value="medium")
]


# Outputs (Separate outputs for MCQs and Subjective Questions)
mcq_output = gr.Textbox(label="Generated MCQs", lines=10)  # Display MCQs in a larger box
subjective_output = gr.Textbox(label="Generated Subjective Questions", lines=10)  # Display subjective questions in a larger box

# Launch Gradio Interface with separate outputs
gr.Interface(
    fn=process_and_generate,
    inputs=inputs,
    outputs=[mcq_output, subjective_output],  # Only MCQs and Subjective Questions
    live=False,
    title="MCQ & Subjective Question Generator",
    theme="default"
).launch(share=True)  # Share link for easy access