RoadGuard-AI Backend

Complete FastAPI backend — auth, hazard events, sensor inference, admin panel, WebSocket.

Pre-seeded credentials

Role	Email	Password
Admin	admin@roadguard.com	Admin@123
User	user@roadguard.com	User@123

All endpoints

Method	Path	Auth?	Description
GET	/api/health	—	Health check
POST	/api/auth/signup	—	Register new user
POST	/api/auth/login	—	Login → JWT token
GET	/api/auth/me	Bearer	Get current user profile
GET	/api/auth/refresh	Bearer	Refresh token
GET	/api/events	—	All hazard events
GET	/api/events/{label}	—	Events by type (0/1/2)
PATCH	/api/events/{id}/solve	—	Mark solved
PATCH	/api/events/{id}/ignore	—	Mark ignored
POST	/api/hazards/report	Bearer	Submit image + location report
POST	/api/predict	—	Sensor-only inference
POST	/api/predict-multimodal	—	Sensor + vision inference
POST	/api/predict-batch	—	Batch inference
GET	/api/weather	—	Weather for lat/lon
POST	/api/chat	—	AI chatbot response
GET	/api/admin/stats	Admin	Dashboard statistics
GET	/api/admin/users	Admin	List all users
GET	/api/admin/reports	Admin	List hazard reports
PUT	/api/admin/users/{id}/ban	Admin	Ban user
PUT	/api/admin/users/{id}/unban	Admin	Unban user
WS	/ws/live	—	Real-time sensor stream

Request / Response: Login

Request:

POST /api/auth/login
Content-Type: application/json

{
  "email": "admin@roadguard.com",
  "password": "Admin@123"
}

Response (200):

{
  "token": "eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9...",
  "user_id": 1,
  "username": "admin",
  "role": "admin",
  "email": "admin@roadguard.com"
}

Response (401):

{"detail": "Invalid email or password"}

Request / Response: Signup

POST /api/auth/signup
{
  "email": "new@example.com",
  "username": "myname",
  "password": "MyPass@123",
  "role": "user"
}

Deploying to Render (free tier, 5 minutes)

1. Push this folder to a GitHub repo
2. Go to https://render.com → New → Web Service
3. Connect your GitHub repo
4. Settings:
   • Build command: pip install -r requirements.txt
   • Start command: uvicorn main:app --host 0.0.0.0 --port $PORT
   • Python version: 3.11
5. Add environment variable: JWT_SECRET → any long random string
6. Click Deploy
7. Your URL: https://your-service-name.onrender.com
8. Update the mobile app .env:
   EXPO_PUBLIC_BACKEND_URL=https://your-service-name.onrender.com

Running locally

pip install -r requirements.txt
uvicorn main:app --reload --port 8000
# API docs: http://localhost:8000/docs

Notes

• Uses SQLite by default (persists in ./roadguard.db)
• For production PostgreSQL: set DATABASE_URL=postgresql://...
• No ML models required — sensor inference is simulated with signal stats
• The simulation produces realistic hazard classifications (~86% match rate)