QCM Generator and Quiz System

System for generating, managing, and delivering multiple-choice questions (QCM - Questionnaire à Choix Multiples) from educational content. The system uses LLaMA for question generation, embeddings for keyword extraction, and provides both API and web interfaces for quiz delivery.

🌟 Features

• Intelligent Question Generation

  • Automatic generation of multiple-choice questions from markdown content
  • Keyword extraction using embedding-based similarity
  • Natural language processing for French content
  • Quality control and validation of generated questions

• Flexible Backend API

  • RESTful endpoints for question retrieval and quiz generation
  • Subject and keyword-based filtering
  • Random question selection
  • Comprehensive statistics

• Modern Web Interface

  • Clean, responsive design using Next.js and Tailwind CSS
  • Subject-based quiz selection
  • Interactive quiz taking experience
  • Immediate feedback and scoring

🛠️ Technology Stack

• Question Generation

  • Python 3.8+
  • LLaMA 3.1 (via Ollama)
  • spaCy (French language model)
  • nomic-embed-text for embeddings

• Backend

  • FastAPI
  • SQLite
  • Pydantic
  • uvicorn

• Frontend

  • Next.js 14
  • React
  • TypeScript
  • Tailwind CSS

📋 Prerequisites

1. Ollama Setup

   # Install Ollama
   curl -fsSL https://ollama.com/install.sh | sh
   
   # Pull required models
   ollama pull llama3.1:latest
   ollama pull nomic-embed-text

2. Python Dependencies

   pip install -r requirements.txt
   python -m spacy download fr_core_news_md

3. Node.js Dependencies

   cd frontend
   npm install

🚀 Installation & Setup

1. Clone the Repository

   git clone https://github.com/yourusername/qcm-system.git
   cd qcm-system

2. Set Up the Database

   # Generate questions from markdown files
   python question_generator/main.py

3. Start the Backend Server

   cd backend
   uvicorn main:app --reload

4. Launch the Frontend

   cd frontend
   npm run dev

🔧 Configuration

Question Generator Settings

# config.py
GENERATOR_CONFIG = {
    "model": "llama3.1:latest",
    "embedding_model": "nomic-embed-text",
    "num_questions_per_file": 5,
    "min_content_length": 200
}

API Configuration

# main.py
app = FastAPI(
    title="QCM API",
    description="API for serving multiple-choice questions",
    version="1.0.0"
)

Frontend Environment Variables

# .env.local
NEXT_PUBLIC_API_URL=http://localhost:8000

🎯 Usage

Generating Questions

1. Prepare your educational content in markdown files:

   # Topic Title
   
   Educational content goes here...

2. Run the question generator:

   python question_generator/main.py --input-dir /path/to/markdown/files

Using the API

1. Get Available Subjects:

   curl http://localhost:8000/subjects

2. Generate a Quiz:

   curl "http://localhost:8000/quiz/generate?num_questions=10&subject=Mathematics"

Taking a Quiz

1. Visit http://localhost:3000 in your browser
2. Select a subject
3. Start the quiz
4. Answer questions and get immediate feedback

📚 API Documentation

Endpoints

Method	Endpoint	Description
GET	/subjects	List all available subjects
GET	/keywords	List all keywords
GET	/questions/random	Get a random question
GET	/quiz/generate	Generate a quiz
GET	/stats	Get database statistics

Example Response

{
  "id": 1,
  "subject": "Mathematics",
  "question_text": "What is 2+2?",
  "choices": {
    "A": "3",
    "B": "4",
    "C": "5",
    "D": "6"
  },
  "answers": ["B"]
}

🧪 Testing

# Backend tests
pytest backend/tests/

# Frontend tests
cd frontend
npm test

👏 Acknowledgments

• LLaMA team for the language model
• Ollama team for the model serving infrastructure
• spaCy team for the French language model

🔮 Future Improvements

• More question types beyond multiple choice
• Enhanced analytics and reporting
• User authentication and progress tracking