Fix check_datasets_active and corresponding unit test (#642)

* Now use different did for active, as d/1 is deactivated. Test against production server as test server does not have deactivated datasets.

* Fix that reflects dataset_list has integer keys (and can not be indexed). Fix retrieving all datasets instead of only active ones. Add documentation.

* Refactored to have a single use of 'active' and forgo many excessive checks on datasets that were not asked for.

* Remove spaces from empty like (flake error).

* Removed unused import.

* PEP8

Author: PGijsbers

openml/datasets/functions.py

@@ -279,32 +279,28 @@ def _load_features_from_file(features_file: str) -> Dict:
         return xml_dict["oml:data_features"]
 
 
-def check_datasets_active(dataset_ids):
-    """Check if the dataset ids provided are active.
+def check_datasets_active(dataset_ids: List[int]) -> Dict[int, bool]:
+    """ Check if the dataset ids provided are active.
 
     Parameters
     ----------
-    dataset_ids : iterable
-        Integers representing dataset ids.
+    dataset_ids : List[int]
+        A list of integers representing dataset ids.
 
     Returns
     -------
     dict
         A dictionary with items {did: bool}
     """
-    dataset_list = list_datasets()
-    dataset_ids = sorted(dataset_ids)
+    dataset_list = list_datasets(status='all')
     active = {}
 
-    for dataset in dataset_list:
-        active[dataset['did']] = dataset['status'] == 'active'
-
     for did in dataset_ids:
-        if did not in active:
-            raise ValueError('Could not find dataset {} in '
-                             'OpenML dataset list.'.format(did))
-
-    active = {did: active[did] for did in dataset_ids}
+        dataset = dataset_list.get(did, None)
+        if dataset is None:
+            raise ValueError('Could not find dataset {} in OpenML dataset list.'.format(did))
+        else:
+            active[did] = (dataset['status'] == 'active')
 
     return active
 

tests/test_datasets/test_dataset_functions.py

@@ -1,4 +1,3 @@
-import unittest
 import os
 import random
 from itertools import product
@@ -206,17 +205,19 @@ def test_list_datasets_empty(self):
 
         self.assertIsInstance(datasets, dict)
 
-    @unittest.skip('See https://github.com/openml/openml-python/issues/149')
     def test_check_datasets_active(self):
-        active = openml.datasets.check_datasets_active([1, 17])
-        self.assertTrue(active[1])
+        # Have to test on live because there is no deactivated dataset on the test server.
+        openml.config.server = self.production_server
+        active = openml.datasets.check_datasets_active([2, 17])
+        self.assertTrue(active[2])
         self.assertFalse(active[17])
         self.assertRaisesRegex(
             ValueError,
             'Could not find dataset 79 in OpenML dataset list.',
             openml.datasets.check_datasets_active,
             [79],
         )
+        openml.config.server = self.test_server
 
     def test_get_datasets(self):
         dids = [1, 2]