[ci] [python-package] fix mypy errors in sklearn.py

[jameslamb]

Contributes to #3756.
Contributes to #3867.

Fixes the following errors from mypy.

python-package/lightgbm/sklearn.py:751: error: List item 0 has incompatible type "Union[str, Callable[[Optional[Any], Any], Tuple[str, float, bool]], Callable[[Optional[Any], Any], List[Tuple[str, float, bool]]], Callable[[Optional[Any], Any, Optional[Any]], Tuple[str, float, bool]], Callable[[Optional[Any], Any, Optional[Any]], List[Tuple[str, float, bool]]], Callable[[Optional[Any], Any, Optional[Any], Optional[Any]], Tuple[str, float, bool]], Callable[[Optional[Any], Any, Optional[Any], Optional[Any]], List[Tuple[str, float, bool]]], None]"; expected "Union[str, Callable[[Optional[Any], Any], Tuple[str, float, bool]], Callable[[Optional[Any], Any], List[Tuple[str, float, bool]]], Callable[[Optional[Any], Any, Optional[Any]], Tuple[str, float, bool]], Callable[[Optional[Any], Any, Optional[Any]], List[Tuple[str, float, bool]]], Callable[[Optional[Any], Any, Optional[Any], Optional[Any]], Tuple[str, float, bool]], Callable[[Optional[Any], Any, Optional[Any], Optional[Any]], List[Tuple[str, float, bool]]]]" [list-item]

python-package/lightgbm/sklearn.py:959: error: Incompatible return value type (got "Union[str, Union[Callable[[Optional[Any], Any], Tuple[Any, Any]], Callable[[Optional[Any], Any, Optional[Any]], Tuple[Any, Any]], Callable[[Optional[Any], Any, Optional[Any], Optional[Any]], Tuple[Any, Any]]], None]", expected "Union[str, Union[Callable[[Optional[Any], Any], Tuple[Any, Any]], Callable[[Optional[Any], Any, Optional[Any]], Tuple[Any, Any]], Callable[[Optional[Any], Any, Optional[Any], Optional[Any]], Tuple[Any, Any]]]]") [return-value]

[jameslamb]

This is the same as what's described in #5788 .

self._objective is initialized to None.

LightGBM/python-package/lightgbm/sklearn.py

Line 402 in 0957ab7

objective: Optional[Union[str, _LGBM_ScikitCustomObjectiveFunction]] = None,

And then overwritten with a string like "regression" or a custom objective function during training.

But mypy doesn't understand that it'll always be non-None if self.__sklearn_is_fitted__() is True (in other words, that it's not possible under normal operation for a user to ever call model.objective_ and get None back).

[jameslamb]

This refactoring adds an explicit case for eval_metric is None. mypy can't tell, I guess, that isinstance(x, list) would must mean something isn't None.

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed.
To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues
including a reference to this.