Skip to content

📡 API Endpoints (Quick Reference)

🔐 Authentication Summary

Service Endpoints Auth Required
Java Gateway /api/auth/login, /api/traffic/* ✅ JWT (except login & health)
Inference Service All endpoints ❌ Public
LSTM Predictor All endpoints ❌ Public

🔐 Java API Gateway (Port 8080)

POST /api/auth/login

Authenticate and receive JWT bearer token — No authentication required (public endpoint)

Request: 

{
  "username": "admin",
  "password": "admin123"
}

Response (200): 

{
  "tokenType": "Bearer",
  "accessToken": "eyJhbGciOiJIUzI1NiJ9...",
  "expiresIn": 3600,
  "timestamp": 1710000000000
}

Errors: 401 Invalid credentials

GET /api/traffic/action

Demo prediction (random junction, auto-generated observations) — Requires JWT

Headers: Authorization: Bearer <token>

Response (200): 

{
  "junctionId": "300839359",
  "predictedAction": 1,
  "signalState": "YELLOW",
  "timestamp": 1710000000000,
  "status": "success"
}

POST /api/traffic/action

Production prediction (custom observations) — Requires JWT

Headers: Authorization: Bearer <token>

Request: 

{
  "junctionId": "joinedS_265580996_300839357",
  "observations": [0.0, 1.0, 0.0, ...],
  "metadata": "morning-peak"
}

Response (200): Same as demo

Errors: 400 Invalid obs, 401 Invalid token

POST /api/traffic/reset

Reset GRU hidden states — Requires JWT

Headers: Authorization: Bearer <token>

Response (200): 

{
  "status": "ok",
  "message": "Hidden states reset for all junctions"
}

GET /api/traffic/health

Service health check — No authentication required (public endpoint)

Response (200 - Healthy): 

{
  "status": "healthy",
  "inferenceService": "up",
  "timestamp": 1710000000000
}

Response (503 - Degraded): 

{
  "status": "degraded",
  "inferenceService": "down",
  "timestamp": 1710000000000
}

🤖 Python Inference Service (Port 8000)

POST /predict_action

Predict optimal signal action — No authentication

Request: 

{
  "junction_id": "300839359",
  "obs_data": [0.0, 1.0, 1.0, 0.12, 0.33, 0.41, 0.22, 0.55]
}

Response (200): 

{
  "junction_id": "300839359",
  "action": 1,
  "confidence": 0.87
}

Errors: - 404: Unknown junction - 400: Observation size > 19 - 503: Model not loaded

POST /reset_hidden

Reset GRU hidden states (call at start of simulation run) — No authentication

Response (200): 

{
  "status": "ok",
  "message": "Hidden states reset for all junctions"
}

GET /health

Service health check

Response (200): 

{
  "status": "healthy",
  "model_loaded": true,
  "junctions": [
    "joinedS_265580996_300839357",
    "300839359",
    "265580972",
    "1270712555",
    "8541180897"
  ]
}

GET /model_info

Model architecture and junction metadata

Response (200): 

{
  "architecture": "RNNAgent (GRU)",
  "input_shape": 24,
  "hidden_dim": 128,
  "n_actions": 4,
  "n_agents": 5,
  "junctions": {
    "300839359": {
      "agent_index": 1,
      "avail_actions": [1, 1, 0, 0],
      "valid_actions": 2
    }
  }
}

GET /docs

Interactive Swagger UI — Auto-generated OpenAPI documentation

🧠 LSTM Traffic Predictor (Port 8001)

GET /health

Service health check

Response (200): 

{
  "status": "healthy",
  "model_loaded": true,
  "model_version": "1.0",
  "forecast_horizon": 15,
  "timestamp": 1710000000000
}

POST /forecast

Predict vehicle flow 15 minutes ahead

Request: 

{
  "junction_id": "300839359",
  "historical_data": [
    {"timestamp": 1710000000, "vehicle_count": 45, "avg_speed": 35.2, "occupancy": 0.12}
  ]
}

Response (200): 

{
  "junction_id": "300839359",
  "forecast_timestamp": 1710000900,
  "predicted_flow": 52,
  "confidence": 0.92,
  "status": "success",
  "timestamp": 1710000000000
}

GET /model_info

LSTM model metadata

Response (200): 

{
  "model_type": "LSTM",
  "model_version": "1.0",
  "input_features": 3,
  "input_window": 60,
  "output_window": 15,
  "target_accuracy": "MAE < 10%",
  "trained_on_junctions": ["300839359", "265580972", ...]
}

GET /docs

Interactive Swagger UI

For Detailed Specifications

See service-specific documentation: - Java API Gateway Endpoints — Full specs with error handling - Python Inference Service Endpoints — Full specs - LSTM Predictor Endpoints — Full specs