copilot fixes

Author: anwurl

octopus/predict/feature_importance.py

@@ -442,8 +442,8 @@ def calculate_fi_permutation(
         test_data: Dict mapping outersplit_id to test DataFrame.
         train_data: Dict mapping outersplit_id to train DataFrame.
         target_assignments: Dict mapping semantic target roles to column
-            names.  For single-target tasks: ``{"default": "target_col"}``.
-            For time-to-event: ``{"duration": "...", "event": "..."}``.
+            names.  For single-target tasks: ``{"default": "y"}``.
+            For time-to-event: ``{"duration": "time_col", "event": "event_col"}``.
         target_metric: Metric name for scoring.
         positive_class: Positive class label for classification.
         n_repeats: Number of permutation repeats per feature.

tests/modules/octo/test_classes_attribute.py

@@ -96,7 +96,7 @@ def test_classification_model_has_classes_attribute(sample_data, model_name):
         training = Training(
             training_id=f"test_{model_name}",
             ml_type=MLType.BINARY,
-            target_assignments={"target": "target"},
+            target_assignments={"default": "target"},
             feature_cols=["x1", "x2"],
             row_id_col="row_id",
             data_train=train,

tests/modules/octo/test_column_ordering.py

@@ -139,10 +139,10 @@ def _create_training(
 ) -> Training:
     """Create a Training instance."""
     if ml_type == MLType.REGRESSION:
-        target_assignments = {"target": "target_reg"}
+        target_assignments = {"default": "target_reg"}
         target_metric = "R2"
     else:
-        target_assignments = {"target": "target_class"}
+        target_assignments = {"default": "target_class"}
         target_metric = "AUCROC"
 
     ml_model_params = _get_default_model_params(model_name)

tests/modules/octo/test_model_fitted_validation.py

@@ -89,12 +89,12 @@ def get_model_configs():
     return {
         MLType.BINARY: {
             "models": available_models[MLType.BINARY],
-            "target_assignments": {"target": "target_class"},
+            "target_assignments": {"default": "target_class"},
             "target_metric": "accuracy",
         },
         MLType.REGRESSION: {
             "models": available_models[MLType.REGRESSION],
-            "target_assignments": {"target": "target_reg"},
+            "target_assignments": {"default": "target_reg"},
             "target_metric": "mse",
         },
         MLType.TIMETOEVENT: {

tests/modules/test_training_feature_importances.py

@@ -72,39 +72,27 @@
 
 ML_TYPE_CONFIGS = {
     MLType.BINARY: {
-        "target_assignments": {"target": "target_class"},
+        "target_assignments": {"default": "target_class"},
         "target_metric": "AUCROC",
     },
     MLType.REGRESSION: {
-        "target_assignments": {"target": "target_reg"},
+        "target_assignments": {"default": "target_reg"},
         "target_metric": "R2",
     },
     MLType.TIMETOEVENT: {
         "target_assignments": {"duration": "duration", "event": "event"},
         "target_metric": "CI",
     },
     MLType.MULTICLASS: {
-        "target_assignments": {"target": "target_multiclass"},
+        "target_assignments": {"default": "target_multiclass"},
         "target_metric": "ACCBAL_MC",
     },
 }
 
 
 def _get_available_models_by_type():
     """Get all available models dynamically from the registry, grouped by ML type."""
-    all_models = Models._config_factories.keys()
-    models_by_type = {ml_type: [] for ml_type in MLType}
-
-    for model_name in all_models:
-        try:
-            model_config = Models.get_config(model_name)
-            for ml_type in MLType:
-                if model_config.supports_ml_type(ml_type):
-                    models_by_type[ml_type].append(model_name)
-        except Exception:
-            continue
-
-    return models_by_type
+    return {ml_type: list(Models.get_models_for_type(ml_type)) for ml_type in MLType}
 
 
 def _generate_model_fi_params():