Dataclasses

Why this matters

FastAPI request models, Pydantic v1 internals, and half of modern config layers use @dataclass or its stricter cousin attrs to kill boilerplate. Staff engineers reach for frozen=True + slots=True when they need hashable, memory-lean value objects for caching keys or event payloads.

Common pitfalls

• Using a mutable default like field=[] — raises ValueError; use field(default_factory=list) instead.
• Assuming frozen=True means deeply immutable — nested lists still mutate; wrap in tuples or use MappingProxyType.
• Skipping slots=True on hot-path classes — costs ~40% more memory per instance versus a slotted dataclass.