fastapi-backend-guidelines

Structure a FastAPI backend using a layered Router → Service → Repository architecture to ensure clean separation of concerns. Organize code by domain with clear boundaries, creating models, repositories, and services for each API domain.

Prevent N+1 query problems in FastAPI by applying DataLoader patterns within your repository layer. This async approach batches database queries, avoiding redundant fetches when accessing related ORM models.

The best way to separate read and write database sessions in FastAPI is through Dependency Injection for asynchronous sessions. This pattern enforces clear boundaries between query and persistence operations within the repository layer.

Use ULID for primary keys in SQLModel by defining reusable domain primitives within your ORM models. This provides sortable, unique identifiers for your entities while maintaining async database operations.

Yes, these FastAPI guidelines support Pydantic v2 for Data Transfer Objects. They specify using field_validator within your DTOs to ensure strict data validation before it reaches the service and repository layers.

Use async/await in FastAPI repository patterns to achieve non-blocking database I/O. This satisfies scalability requirements by ensuring concurrent request handling when the repository layer queries SQLModel or SQLAlchemy ORM models.