chore(all): Enforce PERF and PYI rules (#2101)

DimitriPapadopoulos · web-flow · commit ffb9b1367352 · 2025-10-30T09:21:18.000+01:00
Fixes #2099.
diff --git a/skore-hub-project/pyproject.toml b/skore-hub-project/pyproject.toml
@@ -83,7 +83,23 @@ exclude_also = [
 ]
 
 [tool.ruff.lint]
-select = ["E", "F", "UP", "B", "C4", "SIM", "T", "PT003", "PT014", "I", "D", "RUF010", "RUF015"]
+select = [
+    "B",
+    "C4",
+    "T10",
+    "T20",
+    "PYI055", "PYI059",
+    "PT003", "PT014",
+    "SIM",
+    "I",
+    "PERF",
+    "E",
+    "D",
+    "F",
+    "UP",
+    "RUF010", "RUF015",
+]
+ignore = ["PERF203"]
 
 [tool.ruff.lint.pydocstyle]
 convention = "numpy"
diff --git a/skore-hub-project/src/skore_hub_project/artifact/media/media.py b/skore-hub-project/src/skore_hub_project/artifact/media/media.py
@@ -11,7 +11,7 @@
 Report = TypeVar("Report", bound=(EstimatorReport | CrossValidationReport))
 
 
-class Media(Artifact, Generic[Report], ABC):
+class Media(Artifact, ABC, Generic[Report]):
     """
     Payload used to associate a media with the report.
 
diff --git a/skore-hub-project/src/skore_hub_project/metric/metric.py b/skore-hub-project/src/skore_hub_project/metric/metric.py
@@ -25,7 +25,7 @@ def cast_to_float(value: Any) -> float | None:
     return None
 
 
-class Metric(BaseModel, Generic[Report], ABC):
+class Metric(BaseModel, ABC, Generic[Report]):
     """
     Payload used to send a metric to ``hub``.
 
diff --git a/skore-hub-project/src/skore_hub_project/report/report.py b/skore-hub-project/src/skore_hub_project/report/report.py
@@ -15,7 +15,7 @@
 Report = TypeVar("Report", bound=(EstimatorReport | CrossValidationReport))
 
 
-class ReportPayload(BaseModel, Generic[Report], ABC):
+class ReportPayload(BaseModel, ABC, Generic[Report]):
     """
     Payload used to send a report to ``hub``.
 
diff --git a/skore-local-project/pyproject.toml b/skore-local-project/pyproject.toml
@@ -74,7 +74,23 @@ exclude_also = [
 ]
 
 [tool.ruff.lint]
-select = ["E", "F", "UP", "B", "C4", "SIM", "T", "PT003", "PT014", "I", "D", "RUF010", "RUF015"]
+select = [
+    "B",
+    "C4",
+    "T10",
+    "T20",
+    "PYI055", "PYI059",
+    "PT003", "PT014",
+    "SIM",
+    "I",
+    "PERF",
+    "E",
+    "D",
+    "F",
+    "UP",
+    "RUF010", "RUF015",
+]
+ignore = ["PERF203"]
 
 [tool.ruff.lint.pydocstyle]
 convention = "numpy"
diff --git a/skore-local-project/src/skore_local_project/project.py b/skore-local-project/src/skore_local_project/project.py
@@ -208,7 +208,7 @@ def put(self, key: str, report: EstimatorReport | CrossValidationReport):
         if not isinstance(key, str):
             raise TypeError(f"Key must be a string (found '{type(key)}')")
 
-        Metadata: type[EstimatorReportMetadata] | type[CrossValidationReportMetadata]
+        Metadata: type[EstimatorReportMetadata | CrossValidationReportMetadata]
 
         if isinstance(report, EstimatorReport):
             Metadata = EstimatorReportMetadata
diff --git a/skore/pyproject.toml b/skore/pyproject.toml
@@ -135,7 +135,23 @@ exclude_also = [
 ]
 
 [tool.ruff.lint]
-select = ["E", "F", "UP", "B", "C4", "SIM", "T", "PT003", "PT014", "I", "D", "RUF010", "RUF015"]
+select = [
+    "B",
+    "C4",
+    "T10",
+    "T20",
+    "PYI055", "PYI059",
+    "PT003", "PT014",
+    "SIM",
+    "I",
+    "PERF",
+    "E",
+    "D",
+    "F",
+    "UP",
+    "RUF010", "RUF015",
+]
+ignore = ["PERF203"]
 
 [tool.ruff.lint.pydocstyle]
 convention = "numpy"
diff --git a/skore/src/skore/_sklearn/_comparison/feature_importance_accessor.py b/skore/src/skore/_sklearn/_comparison/feature_importance_accessor.py
@@ -59,33 +59,31 @@ def coefficients(self) -> FeatureImportanceDisplay:
                 }
             )
 
-        coef_frames = []
         if self._parent._reports_type == "EstimatorReport":
-            for reports_with_same_features in similar_reports.values():
-                coef_frames.append(
-                    pd.DataFrame(
-                        {
-                            report_data["estimator_name"]: (
-                                report_data["report_obj"]
-                                .feature_importance.coefficients()
-                                .frame()
-                                .iloc[:, 0]
-                            )
-                            for report_data in reports_with_same_features
-                        },
-                        index=reports_with_same_features[-1]["feature_names"],
-                    )
+            coef_frames = [
+                pd.DataFrame(
+                    {
+                        report_data["estimator_name"]: (
+                            report_data["report_obj"]
+                            .feature_importance.coefficients()
+                            .frame()
+                            .iloc[:, 0]
+                        )
+                        for report_data in reports_with_same_features
+                    },
+                    index=reports_with_same_features[-1]["feature_names"],
                 )
-
+                for reports_with_same_features in similar_reports.values()
+            ]
         elif self._parent._reports_type == "CrossValidationReport":
-            for reports_with_same_features in similar_reports.values():
-                for report_data in reports_with_same_features:
-                    coef_frames.append(
-                        report_data["report_obj"]
-                        .feature_importance.coefficients()
-                        .frame()
-                        .add_prefix(f"{report_data['estimator_name']}__")
-                    )
+            coef_frames = [
+                report_data["report_obj"]
+                .feature_importance.coefficients()
+                .frame()
+                .add_prefix(f"{report_data['estimator_name']}__")
+                for reports_with_same_features in similar_reports.values()
+                for report_data in reports_with_same_features
+            ]
         else:
             raise TypeError(f"Unexpected report type: {self._parent._reports_type}")
 
diff --git a/skore/src/skore/_sklearn/_estimator/report.py b/skore/src/skore/_sklearn/_estimator/report.py
@@ -294,11 +294,11 @@ def cache_predictions(
         progress.update(task, total=total_iterations)
 
         # do not mutate directly `self._cache` during the execution of Parallel
-        results_to_cache = {}
+        results_to_cache: dict[tuple[Any, ...], Any] = {}
         for results in generator:
-            for key, value, is_cached in results:
-                if not is_cached:
-                    results_to_cache[key] = value
+            results_to_cache.update(
+                (key, value) for key, value, is_cached in results if not is_cached
+            )
             progress.update(task, advance=1, refresh=True)
 
         if results_to_cache:
diff --git a/skore/src/skore/_sklearn/_plot/base.py b/skore/src/skore/_sklearn/_plot/base.py
@@ -194,11 +194,11 @@ class HelpDisplayMixin:
 
     def _get_attributes_for_help(self) -> list[str]:
         """Get the attributes ending with '_' to display in help."""
-        attributes = []
-        for name in dir(self):
-            if name.endswith("_") and not name.startswith("_"):
-                attributes.append(f".{name}")
-        return sorted(attributes)
+        return sorted(
+            f".{name}"
+            for name in dir(self)
+            if name.endswith("_") and not name.startswith("_")
+        )
 
     def _get_methods_for_help(self) -> list[tuple[str, Any]]:
         """Get the public methods to display in help."""
diff --git a/skore/src/skore/project/widget.py b/skore/src/skore/project/widget.py
@@ -576,8 +576,7 @@ def _update_plot(self, change: dict[str, Any] | None = None) -> None:
                 self._color_metric_dropdown[ml_task].value
             ]
 
-            dimensions = []
-            dimensions.append(
+            dimensions = [
                 {
                     "label": "Learner",
                     "values": self._add_jitter_to_categorical(
@@ -586,17 +585,17 @@ def _update_plot(self, change: dict[str, Any] | None = None) -> None:
                     "ticktext": df_dataset["learner"].cat.categories,
                     "tickvals": np.arange(len(df_dataset["learner"].cat.categories)),
                 }
-            )
+            ]
 
-            for col in selected_metrics:  # use the order defined in the constructor
-                dimensions.append(
-                    {
-                        "label": self._metrics[col]["name"],
-                        # convert to float in case that the column has None values and
-                        # thus is object type
-                        "values": df_dataset[col].astype(float).fillna(0),
-                    }
-                )
+            dimensions.extend(
+                {
+                    "label": self._metrics[col]["name"],
+                    # convert to float in case that the column has None values and
+                    # thus is object type
+                    "values": df_dataset[col].astype(float).fillna(0),
+                }
+                for col in selected_metrics  # use the order defined in the constructor
+            )
 
             colorscale = (
                 "Viridis"
diff --git a/skore/tests/unit/reports/test_base.py b/skore/tests/unit/reports/test_base.py
@@ -103,8 +103,7 @@ def test_get_cached_response_values(
     )
 
     # cache the results
-    for key, value, _ in results:
-        cache[key] = value
+    cache.update((key, value) for key, value, _ in results)
 
     # Reload from the cache
     results = _get_cached_response_values(cache=cache, **params)
@@ -131,8 +130,8 @@ def test_get_cached_response_values(
         assert not is_cached
         assert response_values.shape == y.shape
 
-        for key, value, _ in results:
-            cache[key] = value
+        cache.update((key, value) for key, value, _ in results)
+
     else:
         assert len(results) == 1
         assert current_calls == initial_calls
@@ -183,8 +182,7 @@ def test_get_cached_response_values_different_data_source_hash(
     initial_calls = getattr(estimator, f"n_call_{response_method}")
 
     # cache the results
-    for key, value, _ in results:
-        cache[key] = value
+    cache.update((key, value) for key, value, _ in results)
 
     # Second call by passing the hash of the data should not trigger new computation
     # because we consider it trustworthy
diff --git a/skore/tests/unit/utils/test_progress_bar.py b/skore/tests/unit/utils/test_progress_bar.py
@@ -149,10 +149,7 @@ def __init__(self):
 
         @progress_decorator("Parent Process")
         def run(self):
-            results = []
-            for rpt in self.reports_.values():
-                results.append(rpt.process())
-            return results
+            return [rpt.process() for rpt in self.reports_.values()]
 
     parent = Parent()
     results = parent.run()
