FastAPI Patterns

Build production-ready async APIs with FastAPI's dependency injection and type system.

Project Layout

src/
├── app/
│   ├── __init__.py
│   ├── main.py           # FastAPI app factory
│   ├── config.py          # Pydantic Settings
│   ├── dependencies.py    # Shared DI providers
│   ├── middleware.py       # Custom middleware
│   ├── routers/
│   │   ├── __init__.py
│   │   ├── skills.py
│   │   └── registry.py
│   ├── models/
│   │   ├── __init__.py
│   │   ├── skill.py       # Pydantic models
│   │   └── registry.py
│   ├── services/
│   │   ├── __init__.py
│   │   ├── skill_service.py
│   │   └── registry_service.py
│   └── db/
│       ├── __init__.py
│       └── session.py
└── tests/
    ├── conftest.py
    └── test_skills.py

App Factory

from fastapi import FastAPI
from contextlib import asynccontextmanager

@asynccontextmanager
async def lifespan(app: FastAPI):
    # Startup
    await init_db()
    await init_cache()
    yield
    # Shutdown
    await close_db()
    await close_cache()

def create_app() -> FastAPI:
    app = FastAPI(
        title="ORGANVM Skills API",
        version="1.0.0",
        lifespan=lifespan,
    )
    app.include_router(skills_router, prefix="/api/skills", tags=["skills"])
    app.include_router(registry_router, prefix="/api/registry", tags=["registry"])
    app.add_middleware(CorrelationMiddleware)
    return app

app = create_app()

Dependency Injection

from fastapi import Depends
from typing import Annotated

async def get_db():
    async with async_session() as session:
        yield session

async def get_current_user(token: str = Depends(oauth2_scheme)) -> User:  # allow-secret
    user = await verify_token(token)
    if not user:
        raise HTTPException(status_code=401, detail="Invalid token")
    return user

# Type aliases for clean signatures
DB = Annotated[AsyncSession, Depends(get_db)]
CurrentUser = Annotated[User, Depends(get_current_user)]

@router.get("/skills/{skill_id}")
async def get_skill(skill_id: str, db: DB, user: CurrentUser) -> SkillResponse:
    skill = await db.get(Skill, skill_id)
    if not skill:
        raise HTTPException(status_code=404, detail="Skill not found")
    return SkillResponse.model_validate(skill)

Pydantic Models

from pydantic import BaseModel, Field
from datetime import datetime

class SkillBase(BaseModel):
    name: str = Field(pattern=r"^[a-z][a-z0-9-]*$", min_length=2, max_length=64)
    description: str = Field(min_length=20, max_length=600)
    category: str
    tags: list[str] = []

class SkillCreate(SkillBase):
    pass

class SkillResponse(SkillBase):
    id: str
    created_at: datetime
    updated_at: datetime

    model_config = {"from_attributes": True}

class SkillList(BaseModel):
    items: list[SkillResponse]
    total: int
    page: int
    per_page: int

Router Patterns

from fastapi import APIRouter, Query

router = APIRouter()

@router.get("/", response_model=SkillList)
async def list_skills(
    db: DB,
    category: str | None = None,
    tag: str | None = None,
    page: int = Query(default=1, ge=1),
    per_page: int = Query(default=20, ge=1, le=100),
) -> SkillList:
    query = select(Skill)
    if category:
        query = query.where(Skill.category == category)
    if tag:
        query = query.where(Skill.tags.contains([tag]))
    total = await db.scalar(select(func.count()).select_from(query.subquery()))
    skills = await db.scalars(query.offset((page - 1) * per_page).limit(per_page))
    return SkillList(items=skills.all(), total=total, page=page, per_page=per_page)

@router.post("/", response_model=SkillResponse, status_code=201)
async def create_skill(data: SkillCreate, db: DB, user: CurrentUser) -> SkillResponse:
    skill = Skill(**data.model_dump())
    db.add(skill)
    await db.commit()
    await db.refresh(skill)
    return SkillResponse.model_validate(skill)

Background Tasks

from fastapi import BackgroundTasks

@router.post("/skills/{skill_id}/validate")
async def validate_skill(
    skill_id: str,
    background_tasks: BackgroundTasks,
    db: DB,
):
    skill = await db.get(Skill, skill_id)
    background_tasks.add_task(run_validation, skill_id)
    return {"status": "validation_queued", "skill_id": skill_id}

async def run_validation(skill_id: str):
    # Long-running validation logic
    async with async_session() as db:
        skill = await db.get(Skill, skill_id)
        result = await validate_frontmatter(skill)
        skill.validation_status = result.status
        await db.commit()

Error Handling

from fastapi import Request
from fastapi.responses import JSONResponse

class AppError(Exception):
    def __init__(self, message: str, code: str, status: int):
        self.message = message
        self.code = code
        self.status = status

@app.exception_handler(AppError)
async def app_error_handler(request: Request, exc: AppError):
    return JSONResponse(
        status_code=exc.status,
        content={"error": {"code": exc.code, "message": exc.message}},
    )

Testing

import pytest
from httpx import AsyncClient, ASGITransport

@pytest.fixture
async def client():
    async with AsyncClient(
        transport=ASGITransport(app=app),
        base_url="http://test",
    ) as client:
        yield client

@pytest.mark.asyncio
async def test_list_skills(client: AsyncClient):
    response = await client.get("/api/skills/")
    assert response.status_code == 200
    data = response.json()
    assert "items" in data
    assert "total" in data

@pytest.mark.asyncio
async def test_create_skill(client: AsyncClient):
    response = await client.post("/api/skills/", json={
        "name": "test-skill",
        "description": "A test skill for unit testing purposes",
        "category": "development",
    })
    assert response.status_code == 201
    assert response.json()["name"] == "test-skill"

Anti-Patterns

• Sync database calls — Always use async drivers (asyncpg, aiosqlite)
• Business logic in routes — Extract to service layer for testability
• No response models — Always define response_model for API documentation
• Global state — Use dependency injection, not module-level singletons
• Missing validation — Pydantic models should validate all inputs
• No lifespan management — Use the lifespan context manager for startup/shutdown