fix: Do not compute Brier score when predict_proba is not available

skore/src/skore/sklearn/cross_validation/cross_validation_helpers.py

@@ -33,28 +33,27 @@ def _get_scorers_to_add(estimator, y) -> dict[str, Any]:
             ),
         }
     if ml_task == "binary-classification":
-        return {
-            "roc_auc": "roc_auc",
-            "brier_score_loss": metrics.make_scorer(
-                metrics.brier_score_loss, response_method="predict_proba"
-            ),
+        scorers_to_add = {
             "recall": "recall",
             "precision": "precision",
+            "roc_auc": "roc_auc",
         }
-    if ml_task == "multiclass-classification":
         if hasattr(estimator, "predict_proba"):
-            return {
-                "recall_weighted": "recall_weighted",
-                "precision_weighted": "precision_weighted",
-                "roc_auc_ovr_weighted": "roc_auc_ovr_weighted",
-                "log_loss": metrics.make_scorer(
-                    metrics.log_loss, response_method="predict_proba"
-                ),
-            }
-        return {
+            scorers_to_add["brier_score_loss"] = metrics.make_scorer(
+                metrics.brier_score_loss, response_method="predict_proba"
+            )
+        return scorers_to_add
+    if ml_task == "multiclass-classification":
+        scorers_to_add = {
             "recall_weighted": "recall_weighted",
             "precision_weighted": "precision_weighted",
         }
+        if hasattr(estimator, "predict_proba"):
+            scorers_to_add["roc_auc_ovr_weighted"] = "roc_auc_ovr_weighted"
+            scorers_to_add["log_loss"] = metrics.make_scorer(
+                metrics.log_loss, response_method="predict_proba"
+            )
+        return scorers_to_add
     return {}
 
 
@@ -104,9 +103,11 @@ def _add_scorers(scorers, scorers_to_add):
 
         internal_scorer = _MultimetricScorer(
             scorers={
-                name: check_scoring(estimator=None, scoring=scoring)
-                if isinstance(scoring, str)
-                else scoring
+                name: (
+                    check_scoring(estimator=None, scoring=scoring)
+                    if isinstance(scoring, str)
+                    else scoring
+                )
                 for name, scoring in scorers_to_add.items()
             }
         )

skore/tests/unit/sklearn/test_cross_validate.py

@@ -1,4 +1,9 @@
+import pytest
+from sklearn.datasets import make_classification, make_regression
+from sklearn.linear_model import LinearRegression, LogisticRegression
+from sklearn.svm import SVC
 from skore.sklearn.cross_validation import CrossValidationReporter
+from skore.sklearn.cross_validation.cross_validation_helpers import _get_scorers_to_add
 
 
 def prepare_cv():
@@ -35,3 +40,66 @@ def test_cross_validate_return_estimator():
     assert "indices" in reporter.cv_results
     assert "estimator" in reporter._cv_results
     assert "indices" in reporter._cv_results
+
+
+@pytest.mark.parametrize(
+    "estimator,dataset_func,dataset_kwargs,expected_keys",
+    [
+        # Regression case
+        (
+            LinearRegression(),
+            make_regression,
+            {"n_targets": 1},
+            {"r2", "root_mean_squared_error"},
+        ),
+        # Binary classification with predict_proba
+        (
+            LogisticRegression(),
+            make_classification,
+            {"n_classes": 2},
+            {"recall", "precision", "roc_auc", "brier_score_loss"},
+        ),
+        # Binary classification without predict_proba
+        (
+            SVC(probability=False),
+            make_classification,
+            {"n_classes": 2},
+            {"recall", "precision", "roc_auc"},
+        ),
+        # Multiclass classification with predict_proba
+        (
+            LogisticRegression(),
+            make_classification,
+            {"n_classes": 3, "n_clusters_per_class": 1},
+            {
+                "recall_weighted",
+                "precision_weighted",
+                "roc_auc_ovr_weighted",
+                "log_loss",
+            },
+        ),
+        # Multiclass classification without predict_proba
+        (
+            SVC(probability=False),
+            make_classification,
+            {"n_classes": 3, "n_clusters_per_class": 1},
+            {"recall_weighted", "precision_weighted"},
+        ),
+    ],
+    ids=[
+        "regression",
+        "binary_classification_with_proba",
+        "binary_classification_without_proba",
+        "multiclass_with_proba",
+        "multiclass_without_proba",
+    ],
+)
+def test_get_scorers_to_add(estimator, dataset_func, dataset_kwargs, expected_keys):
+    """Check that the scorers to add are correct.
+
+    Non-regression test for:
+    https://github.com/probabl-ai/skore/issues/1050
+    """
+    X, y = dataset_func(**dataset_kwargs)
+    scorers = _get_scorers_to_add(estimator, y)
+    assert set(scorers.keys()) == expected_keys