Lyo Backend

A production-ready, modular monolith backend for Lyo, an AI-powered multilingual social-learning application.

Features

• Modular Monolith Architecture: Organized by domain functionality with clean separation of concerns
• Multilingual Support: Built-in internationalization for global audience
• AI Integration: Gemma 3 integration for personalized learning experiences
• Multiple Database Support: PostgreSQL for structured data, Firestore for flexible schema needs
• Caching Layer: Redis for performance optimization
• Real-time Capabilities: WebSocket support for instant notifications
• Comprehensive Security: JWT authentication, role-based access controls
• Observability: OpenTelemetry integration for monitoring and tracing

Tech Stack

• Web Framework: FastAPI
• Relational Database: PostgreSQL with SQLModel
• NoSQL Database: Firestore
• Caching: Redis
• AI Platform: Gemma 3
• Authentication: JWT with OAuth2
• Internationalization: Babel
• Testing: Pytest with asyncio
• Containerization: Docker with docker-compose
• Observability: OpenTelemetry

Project Structure

lyo-backend/
├── api/
│   ├── core/         # Core application modules
│   ├── db/           # Database connectors and utilities
│   ├── middlewares/  # Custom middleware components
│   ├── models/       # Data models and entities
│   ├── routers/      # API endpoints and route definitions
│   ├── schemas/      # Pydantic schemas for request/response
│   └── services/     # Business logic and service layer
├── tests/            # Test suite
├── .env.example      # Environment variable template
├── .env.test         # Test environment variables
├── Dockerfile        # Production Docker configuration
├── Dockerfile.test   # Testing Docker configuration
├── docker-compose.yml        # Development environment setup
├── docker-compose.test.yml   # Testing environment setup
├── requirements.txt  # Python dependencies
└── main.py           # Application entry point

Getting Started

Prerequisites

• Python 3.10+
• Docker and docker-compose
• PostgreSQL
• Redis
• Google Cloud account (for Firestore, Pub/Sub, Storage)

Local Development Setup

1. Clone the repository

   git clone https://github.com/yourusername/lyo-backend.git
   cd lyo-backend

2. Create a virtual environment

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

   pip install -r requirements.txt

4. Set up environment variables

   cp .env.example .env
   # Edit .env with your configuration

5. Run the application

   uvicorn main:app --reload

Docker Setup

1. Build and start the containers

   docker-compose up --build

2. The API will be available at http://localhost:8000

Testing

Run the test suite:

# Using pytest directly
pytest

# Using Docker
docker-compose -f docker-compose.test.yml up --build

API Documentation

Once the application is running:

• Swagger UI: http://localhost:8000/api/v1/docs
• ReDoc: http://localhost:8000/api/v1/redoc
• OpenAPI JSON: http://localhost:8000/api/v1/openapi.json

Deployment

Production Considerations

• Update SECRET_KEY and other sensitive environment variables
• Set DEBUG=false
• Configure proper CORS settings
• Set up proper database credentials
• Configure telemetry endpoints
• Enable rate limiting for public endpoints

License

MIT