python-type-safety

Annotate all public function parameters, return types, and class attributes first, then run mypy or pyright to find gaps. For existing projects, enable strict mode incrementally using per-module overrides in pyproject.toml rather than converting everything at once.

Protocols define structural interfaces through duck typing, so a class satisfies them without explicit inheritance. Abstract base classes require explicit subclassing. Use Protocol when you want type-safe interfaces for classes you do not control.

Python 3.10+ supports the cleaner `User | None` union syntax, which is preferred for new code. The older `Optional[User]` from the typing module remains valid and is required when targeting Python 3.9 or earlier.

Yes, through incremental adoption. Enable strict mode per-module using `# mypy: strict` comments or per-module overrides in pyproject.toml, then expand coverage gradually instead of fixing the entire codebase at once.

Use bounded TypeVars like `TypeVar("ModelT", bound=BaseModel)` when a generic function or class only works with specific types. This lets the type checker reject invalid arguments, such as passing str where a BaseModel subclass is required.