Merge pull request #104 from amueller/xml_refactoring

make "close to the metal" xml functions private

Author: mfeurer

doc/api.rst

@@ -29,9 +29,6 @@ Top-level Classes
    :template: function.rst
 
     check_datasets_active
-    get_dataset_description
-    get_dataset_features
-    get_dataset_qualities
     get_dataset
     get_datasets
     list_datasets

openml/datasets/__init__.py

@@ -1,11 +1,11 @@
 from .functions import (list_datasets, list_datasets_by_tag,
                         check_datasets_active, get_datasets, get_dataset,
-                        get_dataset_description,
-                        get_dataset_features, get_dataset_qualities)
+                        _get_dataset_description,
+                        _get_dataset_features, _get_dataset_qualities)
 from .dataset import OpenMLDataset
 
 __all__ = ['check_datasets_active', 'get_dataset', 'get_datasets',
-           'get_datasets_arf', 'get_dataset_features',
-           'get_dataset_qualities', 'OpenMLDataset', 'list_datasets',
+           'get_datasets_arf', '_get_dataset_features',
+           '_get_dataset_qualities', 'OpenMLDataset', 'list_datasets',
            'list_datasets_by_tag',
-           'get_dataset_description', 'list_datasets']
+           '_get_dataset_description', 'list_datasets']

openml/datasets/functions.py

@@ -236,14 +236,14 @@ def get_dataset(did):
         raise ValueError("Dataset ID is neither an Integer nor can be "
                          "cast to an Integer.")
 
-    description = get_dataset_description(did)
+    description = _get_dataset_description(did)
     arff_file = _get_dataset_arff(did, description=description)
 
     dataset = _create_dataset_from_description(description, arff_file)
     return dataset
 
 
-def get_dataset_description(did):
+def _get_dataset_description(did):
     # TODO implement a cache for this that invalidates itself after some
     # time
     # This can be saved on disk, but cannot be cached properly, because
@@ -295,7 +295,7 @@ def _get_dataset_arff(did, description=None):
         pass
 
     if description is None:
-        description = get_dataset_description(did)
+        description = _get_dataset_description(did)
     url = description['oml:url']
     return_code, arff_string = _read_url(url)
     # TODO: it is inefficient to load the dataset in memory prior to
@@ -307,7 +307,7 @@ def _get_dataset_arff(did, description=None):
     return output_file
 
 
-def get_dataset_features(did):
+def _get_dataset_features(did):
     did_cache_dir = _create_dataset_cache_directory(did)
     features_file = os.path.join(did_cache_dir, "features.xml")
 
@@ -337,7 +337,7 @@ def get_dataset_features(did):
     return features
 
 
-def get_dataset_qualities(did):
+def _get_dataset_qualities(did):
     # Dataset qualities are subject to change and must be fetched every time
     did_cache_dir = _create_dataset_cache_directory(did)
     qualities_file = os.path.join(did_cache_dir, "qualities.xml")

tests/test_datasets.py

@@ -109,20 +109,20 @@ def test_download_rowid(self):
         dataset = openml.datasets.get_dataset(did)
         self.assertEqual(dataset.row_id_attribute, 'instance')
 
-    def test_get_dataset_description(self):
+    def test__get_dataset_description(self):
         # Only a smoke test, I don't know exactly how to test the URL
         # retrieval and "caching"
-        description = openml.datasets.get_dataset_description(2)
+        description = openml.datasets._get_dataset_description(2)
         self.assertIsInstance(description, dict)
 
-    def test_get_dataset_features(self):
+    def test__get_dataset_features(self):
         # Only a smoke check
-        features = openml.datasets.get_dataset_features(2)
+        features = openml.datasets._get_dataset_features(2)
         self.assertIsInstance(features, dict)
 
-    def test_get_dataset_qualities(self):
+    def test__get_dataset_qualities(self):
         # Only a smoke check
-        qualities = openml.datasets.get_dataset_qualities(2)
+        qualities = openml.datasets._get_dataset_qualities(2)
         self.assertIsInstance(qualities, dict)
 
     def test_publish_dataset(self):