Python API Development¶

# FastAPI + Uvicorn + testing
make deps  # or:
uv pip install "fastapi>=0.115" "uvicorn[standard]" pytest httpx

# Run the dev server with auto-reload
uv run uvicorn app.main:app --reload

# app/main.py — the minimal viable API
from fastapi import FastAPI

app = FastAPI(title="My API")

@app.get("/health")
def health():
    return {"status": "ok"}

Real talk: - Use --reload only in dev; never in production - uvicorn app.main:app — module:attribute path, no extensions - Interactive docs at /docs, ReDoc at /redoc, spec at /openapi.json

# Pydantic v2 models — validation + docs in one
from pydantic import BaseModel, Field

class ItemCreate(BaseModel):
    name: str = Field(min_length=1, max_length=100)
    price: float = Field(gt=0)
    tags: list[str] = []

@app.post("/items", response_model=ItemCreate, status_code=201)
def create_item(item: ItemCreate):
    return item  # FastAPI validates request body into the model

Why this works: - Pydantic v2 is Rust-backed — 5-50x faster than v1 validation - response_model filters what the client sees (hide password, etc.) - Type hints give you request validation, OpenAPI schema, and docs for free

Tips: - Test with httpx + FastAPI's TestClient — in-process, no server needed - Use dependency injection for DB sessions — FastAPI resolves them per-request - Put routers in app/routers/ and include them with app.include_router()

Gotchas: - Async + sync DB drivers don't mix: pick one and stay consistent - list[str] works in Pydantic v2 (use List[str] only if stuck on v1) - FastAPI's TestClient uses httpx under the hood — don't double-import