🎛️ Backend Architecture

The backend is built on AdonisJS, a robust Node.js web framework that follows the MVC (Model-View-Controller) pattern. We rigorously apply the Service-Repository pattern to keep controllers thin and logic testable.

Directory Structure

app/
├── Controllers/      # Request Handling (Input Validation, HTTP Responses)
├── Models/           # Lucid ORM Definitions (Database Schema)
├── Services/         # Business Logic (The core intelligence)
├── Validators/       # Data validation rules
├── Websockets/       # Real-time event handlers
└── Tasks/            # Planned tasks (Cron jobs)

Core Design Patterns

1. Service Layer

We do not put business logic in Controllers. Instead, we use dedicated Services.

• CameraService: Manages camera provisioning, connection status checks, and configuration updates.
• AlertService: Handles the lifecycle of an alert (Trigger -> Notify Users -> Acknowledge -> Archive).
• StatsService: Aggregates time-series data for dashboard graphs (e.g., "Activity Over Time").

2. Real-Time Communication

We use Websockets to obtain sub-second latency for critical alerts.

Channel	Event	Payload	Description
cameras	status:update	{ id, status }	Camera going Online/offline
alerts	new:alert	{ type, severity }	Critical user notification
stream	frame	Binary Blob	Live preview (MJPEG over socket)

Database Schema (MySQL)

The database is normalized to ensure data integrity.

• Users 1:N Stables
• Stables 1:N Boxes
• Boxes 1:1 Cameras
• Cameras 1:N Alerts

Note: We use Lucid ORM Migrations to manage schema changes versioning. Always run node ace migration:run after pulling updates.