Merge branch 'main' into issue-1578

Author: fkiraly

openml/_api_calls.py

@@ -22,8 +22,8 @@
 from . import config
 from .__version__ import __version__
 from .exceptions import (
+    OpenMLAuthenticationError,
     OpenMLHashException,
-    OpenMLNotAuthorizedError,
     OpenMLServerError,
     OpenMLServerException,
     OpenMLServerNoResult,
@@ -515,11 +515,7 @@ def __parse_server_exception(
         400,  # run/42 delete
         460,  # task/42 delete
     ]:
-        msg = (
-            f"The API call {url} requires authentication via an API key.\nPlease configure "
-            "OpenML-Python to use your API as described in this example:"
-            "\nhttps://openml.github.io/openml-python/latest/examples/Basics/introduction_tutorial/#authentication"
-        )
-        return OpenMLNotAuthorizedError(message=msg)
+        msg = f"The API call {url} requires authentication via an API key."
+        return OpenMLAuthenticationError(message=msg)
 
     return OpenMLServerException(code=code, message=full_message, url=url)

openml/cli.py

@@ -109,7 +109,7 @@ def check_server(server: str) -> str:
 
     def replace_shorthand(server: str) -> str:
         if server == "test":
-            return "https://test.openml.org/api/v1/xml"
+            return f"{config.TEST_SERVER_URL}/api/v1/xml"
         if server == "production_server":
             return "https://www.openml.org/api/v1/xml"
         return server

openml/config.py

@@ -28,6 +28,8 @@
 OPENML_TEST_SERVER_ADMIN_KEY_ENV_VAR = "OPENML_TEST_SERVER_ADMIN_KEY"
 _TEST_SERVER_NORMAL_USER_KEY = "normaluser"
 
+TEST_SERVER_URL = "https://test.openml.org"
+
 
 class _Config(TypedDict):
     apikey: str
@@ -214,7 +216,7 @@ class ConfigurationForExamples:
     _last_used_server = None
     _last_used_key = None
     _start_last_called = False
-    _test_server = "https://test.openml.org/api/v1/xml"
+    _test_server = f"{TEST_SERVER_URL}/api/v1/xml"
     _test_apikey = _TEST_SERVER_NORMAL_USER_KEY
 
     @classmethod
@@ -470,7 +472,8 @@ def get_cache_directory() -> str:
 
     """
     url_suffix = urlparse(server).netloc
-    reversed_url_suffix = os.sep.join(url_suffix.split(".")[::-1])  # noqa: PTH118
+    url_parts = url_suffix.replace(":", "_").split(".")[::-1]
+    reversed_url_suffix = os.sep.join(url_parts)  # noqa: PTH118
     return os.path.join(_root_cache_directory, reversed_url_suffix)  # noqa: PTH118
 
 

openml/exceptions.py

@@ -63,5 +63,28 @@ class OpenMLNotAuthorizedError(OpenMLServerError):
     """Indicates an authenticated user is not authorized to execute the requested action."""
 
 
+class OpenMLAuthenticationError(OpenMLServerError):
+    """Exception raised when API authentication fails.
+
+    This typically occurs when:
+    - No API key is configured
+    - The API key is invalid or expired
+    - The API key format is incorrect
+
+    This is different from authorization (OpenMLNotAuthorizedError), which occurs
+    when a valid API key lacks permissions for the requested operation.
+    """
+
+    def __init__(self, message: str):
+        help_text = (
+            "\n\nTo fix this:\n"
+            "1. Get your API key from https://www.openml.org/\n"
+            "   (you'll need to register for a free account if you don't have one)\n"
+            "2. Configure your API key by following the authentication guide:\n"
+            "   https://openml.github.io/openml-python/latest/examples/Basics/introduction_tutorial/#authentication"
+        )
+        super().__init__(message + help_text)
+
+
 class ObjectNotPublishedError(PyOpenMLError):
     """Indicates an object has not been published yet."""

openml/tasks/functions.py

@@ -415,9 +415,10 @@ def get_task(
     if not isinstance(task_id, int):
         raise TypeError(f"Task id should be integer, is {type(task_id)}")
 
-    cache_key_dir = openml.utils._create_cache_directory_for_id(TASKS_CACHE_DIR_NAME, task_id)
-    tid_cache_dir = cache_key_dir / str(task_id)
-    tid_cache_dir_existed = tid_cache_dir.exists()
+    task_cache_directory = openml.utils._create_cache_directory_for_id(
+        TASKS_CACHE_DIR_NAME, task_id
+    )
+    task_cache_directory_existed = task_cache_directory.exists()
     try:
         task = _get_task_description(task_id)
         dataset = get_dataset(task.dataset_id, **get_dataset_kwargs)
@@ -434,8 +435,8 @@ def get_task(
         if download_splits and isinstance(task, OpenMLSupervisedTask):
             task.download_split()
     except Exception as e:
-        if not tid_cache_dir_existed:
-            openml.utils._remove_cache_dir_for_id(TASKS_CACHE_DIR_NAME, tid_cache_dir)
+        if not task_cache_directory_existed:
+            openml.utils._remove_cache_dir_for_id(TASKS_CACHE_DIR_NAME, task_cache_directory)
         raise e
 
     return task

openml/testing.py

@@ -47,7 +47,7 @@ class TestBase(unittest.TestCase):
         "user": [],
     }
     flow_name_tracker: ClassVar[list[str]] = []
-    test_server = "https://test.openml.org/api/v1/xml"
+    test_server = f"{openml.config.TEST_SERVER_URL}/api/v1/xml"
     admin_key = os.environ.get(openml.config.OPENML_TEST_SERVER_ADMIN_KEY_ENV_VAR)
     user_key = openml.config._TEST_SERVER_NORMAL_USER_KEY
 

tests/conftest.py

@@ -277,7 +277,7 @@ def with_server(request):
         openml.config.apikey = None
         yield
         return
-    openml.config.server = "https://test.openml.org/api/v1/xml"
+    openml.config.server = f"{openml.config.TEST_SERVER_URL}/api/v1/xml"
     openml.config.apikey = TestBase.user_key
     yield
 

tests/files/localhost_8000

@@ -0,0 +1 @@
+org/openml/test

tests/test_datasets/test_dataset_functions.py

@@ -527,27 +527,20 @@ def test_deletion_of_cache_dir(self):
     def test_deletion_of_cache_dir_faulty_download(self, patch):
         patch.side_effect = Exception("Boom!")
         self.assertRaisesRegex(Exception, "Boom!", openml.datasets.get_dataset, dataset_id=1)
-        datasets_cache_dir = os.path.join(self.workdir, "org", "openml", "test", "datasets")
+        datasets_cache_dir = os.path.join(openml.config.get_cache_directory(), "datasets")
         assert len(os.listdir(datasets_cache_dir)) == 0
 
     @pytest.mark.test_server()
     def test_publish_dataset(self):
-        # lazy loading not possible as we need the arff-file.
-        openml.datasets.get_dataset(3, download_data=True)
-        file_path = os.path.join(
-            openml.config.get_cache_directory(),
-            "datasets",
-            "3",
-            "dataset.arff",
-        )
+        arff_file_path = self.static_cache_dir / "org" / "openml" / "test" / "datasets" / "2" / "dataset.arff"
         dataset = OpenMLDataset(
             "anneal",
             "test",
             data_format="arff",
             version=1,
             licence="public",
             default_target_attribute="class",
-            data_file=file_path,
+            data_file=arff_file_path,
         )
         dataset.publish()
         TestBase._mark_entity_for_removal("data", dataset.dataset_id)
@@ -890,7 +883,7 @@ def test_create_invalid_dataset(self):
 
     @pytest.mark.test_server()
     def test_get_online_dataset_arff(self):
-        dataset_id = 100  # Australian
+        dataset_id = 128  # iris -- one of the few datasets without parquet file
         # lazy loading not used as arff file is checked.
         dataset = openml.datasets.get_dataset(dataset_id, download_data=True)
         decoder = arff.ArffDecoder()
@@ -1468,8 +1461,9 @@ def test_data_edit_critical_field(self):
                     raise e
                 time.sleep(10)
                 # Delete the cache dir to get the newer version of the dataset
+                
                 shutil.rmtree(
-                    os.path.join(self.workdir, "org", "openml", "test", "datasets", str(did)),
+                    os.path.join(openml.config.get_cache_directory(), "datasets", str(did)),
                 )
 
     @pytest.mark.test_server()
@@ -1734,7 +1728,6 @@ def test_delete_dataset(self):
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_dataset_not_owned(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = (
         test_files_directory / "mock_responses" / "datasets" / "data_delete_not_owned.xml"
     )
@@ -1749,14 +1742,13 @@ def test_delete_dataset_not_owned(mock_delete, test_files_directory, test_api_ke
     ):
         openml.datasets.delete_dataset(40_000)
 
-    dataset_url = "https://test.openml.org/api/v1/xml/data/40000"
+    dataset_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/data/40000"
     assert dataset_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_dataset_with_run(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = (
         test_files_directory / "mock_responses" / "datasets" / "data_delete_has_tasks.xml"
     )
@@ -1771,14 +1763,13 @@ def test_delete_dataset_with_run(mock_delete, test_files_directory, test_api_key
     ):
         openml.datasets.delete_dataset(40_000)
 
-    dataset_url = "https://test.openml.org/api/v1/xml/data/40000"
+    dataset_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/data/40000"
     assert dataset_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_dataset_success(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = (
         test_files_directory / "mock_responses" / "datasets" / "data_delete_successful.xml"
     )
@@ -1790,14 +1781,13 @@ def test_delete_dataset_success(mock_delete, test_files_directory, test_api_key)
     success = openml.datasets.delete_dataset(40000)
     assert success
 
-    dataset_url = "https://test.openml.org/api/v1/xml/data/40000"
+    dataset_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/data/40000"
     assert dataset_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_unknown_dataset(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = (
         test_files_directory / "mock_responses" / "datasets" / "data_delete_not_exist.xml"
     )
@@ -1812,7 +1802,7 @@ def test_delete_unknown_dataset(mock_delete, test_files_directory, test_api_key)
     ):
         openml.datasets.delete_dataset(9_999_999)
 
-    dataset_url = "https://test.openml.org/api/v1/xml/data/9999999"
+    dataset_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/data/9999999"
     assert dataset_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
@@ -1907,9 +1897,8 @@ def _dataset_features_is_downloaded(did: int):
 
 
 def _dataset_data_file_is_downloaded(did: int):
-    parquet_present = _dataset_file_is_downloaded(did, "dataset.pq")
-    arff_present = _dataset_file_is_downloaded(did, "dataset.arff")
-    return parquet_present or arff_present
+    cache_directory = Path(openml.config.get_cache_directory()) / "datasets" / str(did)
+    return any(f.suffix in (".pq", ".arff") for f in cache_directory.iterdir())
 
 
 def _assert_datasets_retrieved_successfully(
@@ -2014,7 +2003,7 @@ def test_get_dataset_parquet(requests_mock, test_files_directory):
             test_files_directory / "mock_responses" / "datasets" / "data_description_61.xml"
     )
     # While the mocked example is from production, unit tests by default connect to the test server.
-    requests_mock.get("https://test.openml.org/api/v1/xml/data/61", text=content_file.read_text())
+    requests_mock.get(f"{openml.config.TEST_SERVER_URL}/api/v1/xml/data/61", text=content_file.read_text())
     dataset = openml.datasets.get_dataset(61, download_data=True)
     assert dataset._parquet_url is not None
     assert dataset.parquet_file is not None

tests/test_flows/test_flow_functions.py

@@ -453,7 +453,6 @@ def test_delete_flow(self):
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_flow_not_owned(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = test_files_directory / "mock_responses" / "flows" / "flow_delete_not_owned.xml"
     mock_delete.return_value = create_request_response(
         status_code=412,
@@ -466,14 +465,13 @@ def test_delete_flow_not_owned(mock_delete, test_files_directory, test_api_key):
     ):
         openml.flows.delete_flow(40_000)
 
-    flow_url = "https://test.openml.org/api/v1/xml/flow/40000"
+    flow_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/flow/40000"
     assert flow_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_flow_with_run(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = test_files_directory / "mock_responses" / "flows" / "flow_delete_has_runs.xml"
     mock_delete.return_value = create_request_response(
         status_code=412,
@@ -486,14 +484,13 @@ def test_delete_flow_with_run(mock_delete, test_files_directory, test_api_key):
     ):
         openml.flows.delete_flow(40_000)
 
-    flow_url = "https://test.openml.org/api/v1/xml/flow/40000"
+    flow_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/flow/40000"
     assert flow_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_subflow(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = test_files_directory / "mock_responses" / "flows" / "flow_delete_is_subflow.xml"
     mock_delete.return_value = create_request_response(
         status_code=412,
@@ -506,14 +503,13 @@ def test_delete_subflow(mock_delete, test_files_directory, test_api_key):
     ):
         openml.flows.delete_flow(40_000)
 
-    flow_url = "https://test.openml.org/api/v1/xml/flow/40000"
+    flow_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/flow/40000"
     assert flow_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
 
 @mock.patch.object(requests.Session, "delete")
 def test_delete_flow_success(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = test_files_directory / "mock_responses" / "flows" / "flow_delete_successful.xml"
     mock_delete.return_value = create_request_response(
         status_code=200,
@@ -523,15 +519,14 @@ def test_delete_flow_success(mock_delete, test_files_directory, test_api_key):
     success = openml.flows.delete_flow(33364)
     assert success
 
-    flow_url = "https://test.openml.org/api/v1/xml/flow/33364"
+    flow_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/flow/33364"
     assert flow_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")
 
 
 @mock.patch.object(requests.Session, "delete")
 @pytest.mark.xfail(reason="failures_issue_1544", strict=False)
 def test_delete_unknown_flow(mock_delete, test_files_directory, test_api_key):
-    openml.config.start_using_configuration_for_example()
     content_file = test_files_directory / "mock_responses" / "flows" / "flow_delete_not_exist.xml"
     mock_delete.return_value = create_request_response(
         status_code=412,
@@ -544,6 +539,6 @@ def test_delete_unknown_flow(mock_delete, test_files_directory, test_api_key):
     ):
         openml.flows.delete_flow(9_999_999)
 
-    flow_url = "https://test.openml.org/api/v1/xml/flow/9999999"
+    flow_url = f"{openml.config.TEST_SERVER_URL}/api/v1/xml/flow/9999999"
     assert flow_url == mock_delete.call_args.args[0]
     assert test_api_key == mock_delete.call_args.kwargs.get("params", {}).get("api_key")