chore(skore): Fix typings

skore/pyproject.toml

@@ -159,9 +159,16 @@ convention = "numpy"
 "tests/*" = ["D"]
 
 [tool.mypy]
-ignore_missing_imports = true
-exclude = ["src/skore/_externals/.*", "hatch/*", "tests/*"]
+exclude = ["src/skore/_externals/", "hatch/", "tests/"]
 
 [[tool.mypy.overrides]]
-module = ["sklearn.*"]
 ignore_missing_imports = true
+module = [
+  "ipywidgets.*",
+  "joblib.*",
+  "pandas.*",
+  "plotly.*",
+  "seaborn.*",
+  "sklearn.*",
+  "skrub.*",
+]

skore/src/skore/_sklearn/_comparison/report.py

@@ -342,7 +342,7 @@ def get_predictions(
         ] = "predict",
         X: ArrayLike | None = None,
         pos_label: PositiveLabel | None = _DEFAULT,
-    ) -> list[ArrayLike]:
+    ) -> list[ArrayLike] | list[list[ArrayLike]]:
         """Get predictions from the underlying reports.
 
         This method has the advantage to reload from the cache if the predictions
@@ -406,7 +406,7 @@ def get_predictions(
         >>> print([split_predictions.shape for split_predictions in predictions])
         [(25,), (25,)]
         """
-        return [
+        return [  # type: ignore
             report.get_predictions(
                 data_source=data_source,
                 response_method=response_method,

skore/src/skore/_sklearn/_cross_validation/data_accessor.py

@@ -42,7 +42,7 @@ def _retrieve_data_as_frame(
         y = self._parent.y
 
         if not sbd.is_dataframe(X):
-            X = pd.DataFrame(X, columns=[f"Feature {i}" for i in range(X.shape[1])])
+            X = pd.DataFrame(X, columns=[f"Feature {i}" for i in range(X.shape[1])])  # type: ignore
 
         if with_y:
             if y is None:
@@ -52,10 +52,10 @@ def _retrieve_data_as_frame(
                 name = y.name if y.name is not None else "Target"
                 y = y.to_frame(name=name)
             elif not sbd.is_dataframe(y):
-                if y.ndim == 1:
+                if y.ndim == 1:  # type: ignore
                     columns = ["Target"]
                 else:
-                    columns = [f"Target {i}" for i in range(y.shape[1])]
+                    columns = [f"Target {i}" for i in range(y.shape[1])]  # type: ignore
                 y = pd.DataFrame(y, columns=columns)
 
         return X, y

skore/src/skore/_sklearn/_cross_validation/metrics_accessor.py

@@ -1143,16 +1143,17 @@ def _get_display(
                     X, y, _ = report.metrics._get_X_y_and_data_source_hash(
                         data_source=data_source
                     )
+
                 y_true.append(
                     YPlotData(
                         estimator_name=self._parent.estimator_name_,
                         split=report_idx,
-                        y=y,
+                        y=cast(ArrayLike, y),
                     )
                 )
                 results = _get_cached_response_values(
                     cache=report._cache,
-                    estimator_hash=report._hash,
+                    estimator_hash=int(report._hash),
                     estimator=report._estimator,
                     X=X,
                     response_method=response_method,

skore/src/skore/_sklearn/_estimator/feature_importance_accessor.py

@@ -547,7 +547,7 @@ def _feature_permutation(
             feature_names = (
                 self._parent.estimator_.feature_names_in_
                 if hasattr(self._parent.estimator_, "feature_names_in_")
-                else [f"Feature #{i}" for i in range(X_.shape[1])]
+                else [f"Feature #{i}" for i in range(X_.shape[1])]  # type: ignore
             )
 
             # If there is more than one metric

skore/src/skore/_sklearn/_estimator/metrics_accessor.py

@@ -482,7 +482,7 @@ def _compute_metric_scores(
 
             results = _get_cached_response_values(
                 cache=self._parent._cache,
-                estimator_hash=self._parent._hash,
+                estimator_hash=int(self._parent._hash),
                 estimator=self._parent.estimator_,
                 X=X,
                 response_method=response_method,
@@ -1674,7 +1674,7 @@ def _get_display(
         else:
             results = _get_cached_response_values(
                 cache=self._parent._cache,
-                estimator_hash=self._parent._hash,
+                estimator_hash=int(self._parent._hash),
                 estimator=self._parent.estimator_,
                 X=X,
                 response_method=response_method,

skore/src/skore/_sklearn/_estimator/report.py

@@ -389,7 +389,7 @@ def get_predictions(
 
         results = _get_cached_response_values(
             cache=self._cache,
-            estimator_hash=self._hash,
+            estimator_hash=int(self._hash),
             estimator=self._estimator,
             X=X_,
             response_method=response_method,

skore/src/skore/_sklearn/_plot/metrics/precision_recall_curve.py

@@ -865,17 +865,19 @@ def _compute_data_for_display(
             ):
                 label_binarizer = LabelBinarizer().fit(est.classes_)
                 y_true_onehot_i: NDArray = label_binarizer.transform(y_true_i.y)
+                y_pred_i_y = cast(NDArray, y_pred_i.y)
+
                 for class_idx, class_ in enumerate(est.classes_):
                     precision_class_i, recall_class_i, thresholds_class_i = (
                         precision_recall_curve(
                             y_true_onehot_i[:, class_idx],
-                            y_pred_i.y[:, class_idx],
+                            y_pred_i_y[:, class_idx],
                             pos_label=None,
                             drop_intermediate=drop_intermediate,
                         )
                     )
                     average_precision_class_i = average_precision_score(
-                        y_true_onehot_i[:, class_idx], y_pred_i.y[:, class_idx]
+                        y_true_onehot_i[:, class_idx], y_pred_i_y[:, class_idx]
                     )
 
                     for precision, recall, threshold in zip(

skore/src/skore/_sklearn/_plot/metrics/prediction_error.py

@@ -1,6 +1,6 @@
 import numbers
 from collections import namedtuple
-from typing import Any, Literal
+from typing import Any, Literal, cast
 
 import matplotlib.pyplot as plt
 import numpy as np
@@ -265,7 +265,7 @@ def _plot_single_estimator(
         self.ax_.legend(handles, labels, loc="lower right")
         self.ax_.set_title(f"Prediction Error for {estimator_name}")
 
-        return scatter
+        return cast(list[Artist], scatter)
 
     def _plot_cross_validated_estimator(
         self,
@@ -352,7 +352,7 @@ def _plot_cross_validated_estimator(
             self.ax_.legend(handles, labels, loc="lower right", title=legend_title)
         self.ax_.set_title(f"Prediction Error for {estimator_name}")
 
-        return scatter
+        return cast(list[Artist], scatter)
 
     def _plot_comparison_estimator(
         self,
@@ -435,7 +435,7 @@ def _plot_comparison_estimator(
             self.ax_.legend(handles, labels, loc="lower right", title=legend_title)
         self.ax_.set_title("Prediction Error")
 
-        return scatter
+        return cast(list[Artist], scatter)
 
     def _plot_comparison_cross_validation(
         self,
@@ -518,7 +518,7 @@ def _plot_comparison_cross_validation(
             self.ax_.legend(handles, labels, loc="lower right", title=legend_title)
         self.ax_.set_title("Prediction Error")
 
-        return scatter
+        return cast(list[Artist], scatter)
 
     @DisplayMixin.style_plot
     def plot(
@@ -824,9 +824,9 @@ def _compute_data_for_display(
                         }
                     )
             else:
-                y_true_sample = y_true_i.y
-                y_pred_sample = y_pred_i.y
-                residuals_sample = y_true_i.y - y_pred_i.y
+                y_true_sample = cast(np.typing.NDArray, y_true_i.y)
+                y_pred_sample = cast(np.typing.NDArray, y_pred_i.y)
+                residuals_sample = y_true_sample - y_pred_sample
 
                 for y_true_sample_i, y_pred_sample_i, residuals_sample_i in zip(
                     y_true_sample, y_pred_sample, residuals_sample, strict=False

skore/src/skore/_sklearn/_plot/metrics/roc_curve.py

@@ -148,6 +148,8 @@ def __init__(
         self.ml_task = ml_task
         self.report_type = report_type
 
+        self.chance_level_: Line2D | list[Line2D] | None
+
     def _plot_single_estimator(
         self,
         *,
@@ -947,10 +949,12 @@ def _compute_data_for_display(
             ):
                 label_binarizer = LabelBinarizer().fit(est.classes_)
                 y_true_onehot_i: NDArray = label_binarizer.transform(y_true_i.y)
+                y_pred_i_y = cast(NDArray, y_pred_i.y)
+
                 for class_idx, class_ in enumerate(est.classes_):
                     fpr_class_i, tpr_class_i, thresholds_class_i = roc_curve(
                         y_true_onehot_i[:, class_idx],
-                        y_pred_i.y[:, class_idx],
+                        y_pred_i_y[:, class_idx],
                         pos_label=None,
                         drop_intermediate=drop_intermediate,
                     )

skore/src/skore/_sklearn/train_test_split/train_test_split.py

@@ -207,6 +207,8 @@ class labels.
     if y is None and len(arrays) >= 2:
         y = arrays[-1]
 
+    y_labels: np.ndarray | None
+
     if y is not None:
         y_labels = np.unique(y)
         y_test = (

skore/src/skore/project/widget.py

@@ -207,7 +207,7 @@ def _filter_dataframe(self, ml_task: str, report_type: str) -> pd.DataFrame:
             df.columns = [col.removesuffix("_mean") for col in df.columns]
         return df
 
-    def _get_datasets(self, ml_task: str, report_type: str) -> np.ndarray:
+    def _get_datasets(self, ml_task: str, report_type: str) -> list[str]:
         """Get the unique datasets from the filtered dataframe.
 
         Parameters
@@ -219,7 +219,7 @@ def _get_datasets(self, ml_task: str, report_type: str) -> np.ndarray:
 
         Returns
         -------
-        np.ndarray
+        list[str]
             The unique datasets.
         """
         return self._filter_dataframe(ml_task, report_type)["dataset"].unique()