type-checking-pattern

F821 undefined name errors occur when type hints reference packages that may not be installed. Use TYPE_CHECKING blocks to guard optional imports—wrap them in `if TYPE_CHECKING:` so static type checkers see the types while runtime code never imports unavailable packages. Quote annotations as strings to defer evaluation and prevent NameErrors when dependencies are absent.

TYPE_CHECKING is a Python constant that's True only during static analysis and False at runtime. It lets you import packages solely for type hints without triggering ImportError if they're not installed. Combining TYPE_CHECKING blocks with quoted type annotations preserves IDE autocompletion and ruff linting while keeping code runtime-safe across environments.

Yes. Place optional package imports inside `if TYPE_CHECKING:` blocks and quote your annotations as strings. This pattern is safe for packages like dask, pandas, and cupy—static tools resolve the types, but your code never tries to import them at runtime if they're missing, eliminating F821 errors.

Import TYPE_CHECKING from typing, place optional package imports in `if TYPE_CHECKING:` blocks, and use quoted string annotations for those types in function signatures and variable declarations. This resolves ruff F821 errors and enables IDE autocompletion without requiring optional packages at runtime.

Ruff checks that all names in type hints are defined, but unguarded imports of optional packages fail at runtime if the package is absent. TYPE_CHECKING blocks hide these imports from the runtime interpreter while exposing them to static analysis, so ruff and type checkers see the definitions without causing ImportError.

Yes. TYPE_CHECKING blocks combined with quoted annotations preserve type information for IDEs, mypy, pyright, and ruff. Static tools evaluate the guarded imports during analysis, enabling full autocompletion and type checking without requiring optional dependencies to be installed.