FastAPI

⚡ FastAPI

FastAPI is a modern, high-performance web framework for building APIs with Python 3.8+ based on standard Python type hints.

Property	Value
Used In	AI Service
Language	Python
Port	3005
Documentation	fastapi.tiangolo.com

⚠️ Note: Do not confuse FastAPI (Python) with Fastify (Node.js).

• FastAPI is used only for the AI microservice.
• Fastify is used for the Gateway, Auth, and Game services.

---

💡 Why FastAPI?

We chose FastAPI for the AI microservice because it bridges the gap between web servers and Data Science tools perfectly.

1. Native Python Support: Our AI model uses PyTorch and Stable Baselines3, which are Python-native. FastAPI allows us to serve these models directly without context switching.
2. High Performance: It is one of the fastest Python frameworks available, running on Uvicorn (an ASGI server). This is critical for the AI, which must reply with a move in <50ms.
3. Automatic Documentation: It automatically generates interactive Swagger UI documentation, allowing us to test the AI moves manually in the browser.
4. Data Validation: It uses Pydantic to validate game coordinates automatically. If the Game Service sends a string instead of a number, FastAPI rejects it instantly.

---

🏗️ Implementation Details

Request Validation (Pydantic)

Instead of manually checking if ball_x is a number, we define a "Model" (Schema). FastAPI guarantees that the function get_move is only called if the data is valid.

from pydantic import BaseModel

# Define the expected data structure
class GameState(BaseModel):
    ball_x: float
    ball_y: float
    paddle_y: float
    opponent_y: float

# The API Endpoint
@app.post("/join-game")
async def get_move(state: GameState):
    # 'state' is already validated and converted to the correct types
    action = model.predict([state.ball_x, state.ball_y, ...])
    return {"action": int(action)}