Fastapi Patterns

FastAPI Patterns

Patterns and conventions for building FastAPI applications. Follow the principle of explicit dependency injection, leverage async where it matters, and use Pydantic models as the single source of truth for request/response schemas.

Core Principles

• Dependency injection everywhere -- Use Depends() for database sessions, auth, config, and shared logic instead of global state or imports
• Pydantic as contract -- Define request and response models explicitly with Pydantic; avoid raw dicts in route signatures
• Async when beneficial -- Use async def for I/O-bound routes with async drivers; use plain def for CPU-bound or sync-only code
• Router organization -- Split routes into APIRouter modules by domain, compose them in the main app with prefixes and tags

When to Use

Apply these patterns when working on any project that lists fastapi in its dependencies. Detect this by checking pyproject.toml, requirements.txt, or Pipfile for FastAPI.

Type Hints

Use | None instead of Optional for all type annotations:

from fastapi import Query

async def list_items(
    category: str | None = Query(None, description="Filter by category"),
) -> list[ItemResponse]:
    ...

Logging

Use emojis as prefixes in log messages:

logger.info("✅ Request processed for %s", endpoint)
logger.warning("⚠️ Rate limit approaching for client %s", client_id)
logger.error("❌ Failed to connect to database: %s", exc)

Project Structure

src/
├── main.py              # App factory, lifespan, router composition
├── config.py            # Settings with pydantic-settings
├── dependencies.py      # Shared dependencies (DB session, current user)
├── routes/
│   ├── __init__.py
│   ├── users.py         # APIRouter for user endpoints
│   └── items.py         # APIRouter for item endpoints
├── models/              # SQLAlchemy or ORM models
├── schemas/             # Pydantic request/response models
├── services/            # Business logic layer
└── middleware/           # Custom middleware

Reference Documents

• routes-di.md -- Routes, dependency injection, middleware, error handling, OpenAPI customization, lifespan events
• async-patterns.md -- Async/await patterns, background tasks, WebSockets, streaming, connection pooling, testing
• auth.md -- Authentication and authorization: OAuth2, JWT, API keys, RBAC, scopes, password hashing, CORS for auth

Source: weorbitant/compound-engineering-feat-python-plugin — distributed by TomeVault.