[ENH] Improve NotImplementedError Messages

openml/datasets/dataset.py

@@ -420,7 +420,11 @@ def _get_arff(self, format: str) -> dict:  # noqa: A002
             file_size = filepath.stat().st_size
             if file_size > MB_120:
                 raise NotImplementedError(
-                    f"File {filename} too big for {file_size}-bit system ({bits} bytes).",
+                    f"File '{filename}' ({file_size / 1e6:.1f} MB)"
+                    f"exceeds the maximum supported size of 120 MB. "
+                    f"This limitation applies to {bits}-bit systems. "
+                    f"Large dataset handling is currently not fully supported. "
+                    f"Please consider using a smaller dataset"
                 )
 
         if format.lower() == "arff":
@@ -780,7 +784,12 @@ def get_data(  # noqa: C901
         # All the assumptions below for the target are dependant on the number of targets being 1
         n_targets = len(target_names)
         if n_targets > 1:
-            raise NotImplementedError(f"Number of targets {n_targets} not implemented.")
+            raise NotImplementedError(
+                f"Multi-target prediction is not yet supported."
+                f"Found {n_targets} target columns: {target_names}. "
+                f"Currently, only single-target datasets are supported. "
+                f"Please select a single target column."
+            )
 
         target_name = target_names[0]
         x = data.drop(columns=[target_name])

openml/runs/functions.py

@@ -755,7 +755,12 @@ def _run_task_get_arffcontent_parallel_helper(  # noqa: PLR0913
         test_x = None
         test_y = None
     else:
-        raise NotImplementedError(task.task_type)
+        raise NotImplementedError(
+            f"Task type '{task.task_type}' is not supported. "
+            f"Only OpenMLSupervisedTask and OpenMLClusteringTask are currently implemented. "
+            f"Task details: task_id={getattr(task, 'task_id', 'unknown')}, "
+            f"task_class={task.__class__.__name__}"
+        )
 
     config.logger.info(
         f"Going to run model {model!s} on "
@@ -982,7 +987,13 @@ def obtain_field(xml_obj, fieldname, from_server, cast=None):  # type: ignore
     if "predictions" not in files and from_server is True:
         task = openml.tasks.get_task(task_id)
         if task.task_type_id == TaskType.SUBGROUP_DISCOVERY:
-            raise NotImplementedError("Subgroup discovery tasks are not yet supported.")
+            raise NotImplementedError(
+                f"Subgroup discovery tasks are not yet supported. "
+                f"Task ID: {task_id}. Please check the OpenML documentation"
+                f"for supported task types. "
+                f"Currently supported task types: Classification, Regression,"
+                f"Clustering, and Learning Curve."
+            )
 
         # JvR: actually, I am not sure whether this error should be raised.
         # a run can consist without predictions. But for now let's keep it
@@ -1282,7 +1293,12 @@ def format_prediction(  # noqa: PLR0913
     if isinstance(task, OpenMLRegressionTask):
         return [repeat, fold, index, prediction, truth]
 
-    raise NotImplementedError(f"Formatting for {type(task)} is not supported.")
+    raise NotImplementedError(
+        f"Formatting predictions for task type '{type(task).__name__}' is not supported. "
+        f"Supported task types: OpenMLClassificationTask, OpenMLRegressionTask,"
+        f"and OpenMLLearningCurveTask. "
+        f"Please ensure your task is one of these types."
+    )
 
 
 def delete_run(run_id: int) -> bool:

openml/runs/run.py

@@ -480,7 +480,12 @@ def _generate_arff_dict(self) -> OrderedDict[str, Any]:
             ]
 
         else:
-            raise NotImplementedError(f"Task type {task.task_type!s} is not yet supported.")
+            raise NotImplementedError(
+                f"Task type '{task.task_type}' is not yet supported. "
+                f"Supported task types: Classification, Regression, Clustering, Learning Curve. "
+                f"Task ID: {task.task_id}. "
+                f"Please check the OpenML documentation for supported task types."
+            )
 
         return arff_dict
 

openml/tasks/functions.py

@@ -528,7 +528,12 @@ def _create_task_from_xml(xml: str) -> OpenMLTask:
         TaskType.LEARNING_CURVE: OpenMLLearningCurveTask,
     }.get(task_type)
     if cls is None:
-        raise NotImplementedError(f"Task type {common_kwargs['task_type']} not supported.")
+        raise NotImplementedError(
+            f"Task type '{common_kwargs['task_type']}' is not supported. "
+            f"Supported task types: SUPERVISED_CLASSIFICATION,"
+            f"SUPERVISED_REGRESSION, CLUSTERING, LEARNING_CURVE."
+            f"Please check the OpenML documentation for available task types."
+        )
     return cls(**common_kwargs)  # type: ignore
 
 
@@ -584,7 +589,13 @@ def create_task(
     elif task_type == TaskType.SUPERVISED_REGRESSION:
         task_cls = OpenMLRegressionTask  # type: ignore
     else:
-        raise NotImplementedError(f"Task type {task_type:d} not supported.")
+        raise NotImplementedError(
+            f"Task type ID {task_type:d} is not supported. "
+            f"Supported task type IDs: {TaskType.SUPERVISED_CLASSIFICATION.value},"
+            f"{TaskType.SUPERVISED_REGRESSION.value}, "
+            f"{TaskType.CLUSTERING.value}, {TaskType.LEARNING_CURVE.value}. "
+            f"Please refer to the TaskType enum for valid task type identifiers."
+        )
 
     return task_cls(
         task_type_id=task_type,

openml/tasks/task.py

@@ -290,7 +290,12 @@ def get_X_and_y(self) -> tuple[pd.DataFrame, pd.Series | pd.DataFrame | None]:
             TaskType.SUPERVISED_REGRESSION,
             TaskType.LEARNING_CURVE,
         ):
-            raise NotImplementedError(self.task_type)
+            raise NotImplementedError(
+                f"Task type '{self.task_type}' is not implemented for get_X_and_y(). "
+                f"Supported types: SUPERVISED_CLASSIFICATION, SUPERVISED_REGRESSION,"
+                f"LEARNING_CURVE."
+                f"Task ID: {getattr(self, 'task_id', 'unknown')}. "
+            )
 
         X, y, _, _ = dataset.get_data(target=self.target_name)
         return X, y
@@ -382,7 +387,7 @@ def __init__(  # noqa: PLR0913
         self.cost_matrix = cost_matrix
 
         if cost_matrix is not None:
-            raise NotImplementedError("Costmatrix")
+            raise NotImplementedError("Costmatrix functionality is not yet implemented.")
 
 
 class OpenMLRegressionTask(OpenMLSupervisedTask):