Fix mypy errors. (#8444)

python-package/xgboost/callback.py

@@ -135,7 +135,7 @@ class CallbackContainer:
     def __init__(
         self,
         callbacks: Sequence[TrainingCallback],
-        metric: Callable = None,
+        metric: Optional[Callable] = None,
         output_margin: bool = True,
         is_cv: bool = False
     ) -> None:
@@ -391,8 +391,6 @@ class EarlyStopping(TrainingCallback):
         else:
             improve_op = minimize
 
-        assert improve_op
-
         if not self.stopping_history:  # First round
             self.current_rounds = 0
             self.stopping_history[name] = {}

python-package/xgboost/dask.py

@@ -288,10 +288,10 @@ class DaskDMatrix:
         *,
         weight: Optional[_DaskCollection] = None,
         base_margin: Optional[_DaskCollection] = None,
-        missing: float = None,
+        missing: Optional[float] = None,
         silent: bool = False,  # pylint: disable=unused-argument
         feature_names: Optional[FeatureNames] = None,
-        feature_types: FeatureTypes = None,
+        feature_types: Optional[FeatureTypes] = None,
         group: Optional[_DaskCollection] = None,
         qid: Optional[_DaskCollection] = None,
         label_lower_bound: Optional[_DaskCollection] = None,
@@ -304,7 +304,7 @@ class DaskDMatrix:
 
         self.feature_names = feature_names
         self.feature_types = feature_types
-        self.missing = missing
+        self.missing = missing if missing is not None else numpy.nan
         self.enable_categorical = enable_categorical
 
         if qid is not None and weight is not None:
@@ -651,7 +651,7 @@ class DaskQuantileDMatrix(DaskDMatrix):
         *,
         weight: Optional[_DaskCollection] = None,
         base_margin: Optional[_DaskCollection] = None,
-        missing: float = None,
+        missing: Optional[float] = None,
         silent: bool = False,  # disable=unused-argument
         feature_names: Optional[FeatureNames] = None,
         feature_types: Optional[Union[Any, List[Any]]] = None,
@@ -2129,7 +2129,7 @@ class DaskXGBRanker(DaskScikitLearnBase, XGBRankerMixIn):
         eval_group: Optional[Sequence[_DaskCollection]] = None,
         eval_qid: Optional[Sequence[_DaskCollection]] = None,
         eval_metric: Optional[Union[str, Sequence[str], Callable]] = None,
-        early_stopping_rounds: int = None,
+        early_stopping_rounds: Optional[int] = None,
         verbose: Union[int, bool] = False,
         xgb_model: Optional[Union[XGBModel, Booster]] = None,
         sample_weight_eval_set: Optional[Sequence[_DaskCollection]] = None,

python-package/xgboost/rabit.py

@@ -152,7 +152,7 @@ def version_number() -> int:
 class RabitContext:
     """A context controlling rabit initialization and finalization."""
 
-    def __init__(self, args: List[bytes] = None) -> None:
+    def __init__(self, args: Optional[List[bytes]] = None) -> None:
         if args is None:
             args = []
         self.args = args

python-package/xgboost/sklearn.py

@@ -233,7 +233,7 @@ __model_doc = f"""
         should be used to specify categorical data type.  Also, JSON/UBJSON
         serialization format is required.
 
-    feature_types : FeatureTypes
+    feature_types : Optional[FeatureTypes]
 
         .. versionadded:: 1.7.0
 
@@ -572,7 +572,7 @@ class XGBModel(XGBModelBase):
         validate_parameters: Optional[bool] = None,
         predictor: Optional[str] = None,
         enable_categorical: bool = False,
-        feature_types: FeatureTypes = None,
+        feature_types: Optional[FeatureTypes] = None,
         max_cat_to_onehot: Optional[int] = None,
         max_cat_threshold: Optional[int] = None,
         eval_metric: Optional[Union[str, List[str], Callable]] = None,

python-package/xgboost/training.py

@@ -1,4 +1,3 @@
-# coding: utf-8
 # pylint: disable=too-many-locals, too-many-arguments, invalid-name
 # pylint: disable=too-many-branches, too-many-statements
 """Training Library containing training routines."""
@@ -71,7 +70,7 @@ def train(
     feval: Optional[Metric] = None,
     maximize: Optional[bool] = None,
     early_stopping_rounds: Optional[int] = None,
-    evals_result: TrainingCallback.EvalsLog = None,
+    evals_result: Optional[TrainingCallback.EvalsLog] = None,
     verbose_eval: Optional[Union[bool, int]] = True,
     xgb_model: Optional[Union[str, os.PathLike, Booster, bytearray]] = None,
     callbacks: Optional[Sequence[TrainingCallback]] = None,
@@ -285,15 +284,20 @@ def groups_to_rows(groups: List[np.ndarray], boundaries: np.ndarray) -> np.ndarr
     return np.concatenate([np.arange(boundaries[g], boundaries[g+1]) for g in groups])
 
 
-def mkgroupfold(dall: DMatrix, nfold: int, param: BoosterParam,
+def mkgroupfold(
-                evals: Sequence[str] = (), fpreproc: FPreProcCallable = None,
+    dall: DMatrix,
-                shuffle: bool = True) -> List[CVPack]:
+    nfold: int,
+    param: BoosterParam,
+    evals: Sequence[str] = (),
+    fpreproc: Optional[FPreProcCallable] = None,
+    shuffle: bool = True,
+) -> List[CVPack]:
     """
     Make n folds for cross-validation maintaining groups
     :return: cross-validation folds
     """
     # we have groups for pairwise ranking... get a list of the group indexes
-    group_boundaries = dall.get_uint_info('group_ptr')
+    group_boundaries = dall.get_uint_info("group_ptr")
     group_sizes = np.diff(group_boundaries)
 
     if shuffle is True:
@@ -327,10 +331,17 @@ def mkgroupfold(dall: DMatrix, nfold: int, param: BoosterParam,
     return ret
 
 
-def mknfold(dall: DMatrix, nfold: int, param: BoosterParam, seed: int,
+def mknfold(
-            evals: Sequence[str] = (), fpreproc: FPreProcCallable = None,
+    dall: DMatrix,
-            stratified: bool = False, folds: XGBStratifiedKFold = None, shuffle: bool = True
+    nfold: int,
-            ) -> List[CVPack]:
+    param: BoosterParam,
+    seed: int,
+    evals: Sequence[str] = (),
+    fpreproc: Optional[FPreProcCallable] = None,
+    stratified: Optional[bool] = False,
+    folds: Optional[XGBStratifiedKFold] = None,
+    shuffle: bool = True,
+) -> List[CVPack]:
     """
     Make an n-fold list of CVPack from random indices.
     """
@@ -393,14 +404,14 @@ def cv(
     metrics: Sequence[str] = (),
     obj: Optional[Objective] = None,
     feval: Optional[Metric] = None,
-    maximize: bool = None,
+    maximize: Optional[bool] = None,
-    early_stopping_rounds: int = None,
+    early_stopping_rounds: Optional[int] = None,
-    fpreproc: FPreProcCallable = None,
+    fpreproc: Optional[FPreProcCallable] = None,
     as_pandas: bool = True,
     verbose_eval: Optional[Union[int, bool]] = None,
     show_stdv: bool = True,
     seed: int = 0,
-    callbacks: Sequence[TrainingCallback] = None,
+    callbacks: Optional[Sequence[TrainingCallback]] = None,
     shuffle: bool = True,
     custom_metric: Optional[Metric] = None,
 ) -> Union[Dict[str, float], DataFrame]: