app.py

import gradio as gr
import os
import fitz  # PyMuPDF for PDF extraction
from pptx import Presentation
from docx import Document
from groq import Groq

# Initialize Groq Client
api_key = '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' # Use environment variable for the API key
client = Groq(api_key=api_key)

# File Extraction Functions
def extract_text_from_pdf(file_path):
    pdf_text = ""
    try:
        pdf_file = fitz.open(file_path)
        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_path):
    ppt_text = ""
    try:
        presentation = Presentation(file_path)
        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_path):
    doc_text = ""
    try:
        document = Document(file_path)
        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 process_files(file_paths):
    text_data = ""
    for file_path in file_paths:
        if file_path.endswith(".pdf"):
            text_data += extract_text_from_pdf(file_path)
        elif file_path.endswith(".pptx"):
            text_data += extract_text_from_ppt(file_path)
        elif file_path.endswith(".docx"):
            text_data += extract_text_from_word(file_path)
        else:
            text_data += f"Unsupported file format: {file_path}\n"
    return text_data

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

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

        question_type_map = {
            "reason": "Generate reasoning-based questions.",
            "short": "Generate short-answer questions.",
            "long": "Generate long-answer questions.",
            "case study": "Generate case study-based questions.",
            "mix": "Generate a mix of question types."
        }

        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. {question_type_map.get(question_type, 'mix')}"

        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 into MCQs and Subjective questions
        mcqs, subjective = "", ""
        is_subjective_section = False

        for line in response.split("\n"):
            if "**Subjective Questions**" in line:
                is_subjective_section = True
            if is_subjective_section:
                subjective += line + "\n"
            else:
                mcqs += line + "\n"

        return mcqs, subjective

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

# Gradio Interface Function
def process_and_generate(file_paths, raw_text, num_mcqs, num_subjective, difficulty_mcqs, difficulty_subjective, question_type):
    combined_text = ""

    # Extract text from uploaded files
    if file_paths:
        extracted_text = process_files(file_paths)
        if extracted_text.strip():
            combined_text += extracted_text
    
    # Add raw text if provided
    if raw_text.strip():
        if combined_text:
            combined_text += "\n" + raw_text
        else:
            combined_text = raw_text
    
    if not combined_text.strip():
        return "No text provided to generate questions.", "No text provided to generate questions."
    
    try:
        mcqs, subjective = generate_questions(
            combined_text, num_mcqs, num_subjective, difficulty_mcqs, difficulty_subjective, question_type
        )
        return mcqs, subjective
    except Exception as e:
        return f"Error generating questions: {e}", f"Error generating questions: {e}"

# Gradio Inputs and Outputs
inputs = [
    gr.File(file_count="multiple", type="filepath", label="Upload Files (.pdf, .pptx, .docx)"),
    gr.Textbox(lines=3, placeholder="Enter raw text here (Optional)...", label="Raw Text"),
    gr.Slider(minimum=2, maximum=40, value=5, step=1, label="Number of MCQs (Max 40)"),
    gr.Slider(minimum=2, 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"),
    gr.Radio(["reason", "short", "long", "case study", "mix"], label="Select Type of Question", value="mix")
]

outputs = [
    gr.Textbox(label="Generated MCQs", lines=10),
    gr.Textbox(label="Generated Subjective Questions", lines=10)
]

# Gradio Interface
gr.Interface(
    fn=process_and_generate,
    inputs=inputs,
    outputs=outputs,
    title="MCQ & Subjective Question Generator",
    live=False
).launch()