Fix test fixtures

tests/test_dogs_vs_cats.py

@@ -16,58 +16,67 @@
 from fuel.schemes import SequentialScheme
 
 
-class TestDogsVsCats(object):
-    @classmethod
-    def setUpClass(cls):
-        config._old_data_path = config.data_path
-        config.data_path = tempfile.mkdtemp()
-        cls._make_dummy_data(config.data_path[0])
-
-    @classmethod
-    def tearDownClass(cls):
-        shutil.rmtree(config.data_path[0])
-        config.data_path = config._old_data_path
-        del config._old_data_path
-
-    @classmethod
-    def _make_dummy_data(cls, output_directory):
-        data = six.BytesIO()
-        Image.new('RGB', (1, 1)).save(data, 'JPEG')
-        image = data.getvalue()
-
-        output_files = [os.path.join(output_directory,
-                                     'dogs_vs_cats.{}.zip'.format(set_))
-                        for set_ in ['train', 'test1']]
-        with zipfile.ZipFile(output_files[0], 'w') as zip_file:
-            zif = zipfile.ZipInfo('train/')
-            zip_file.writestr(zif, "")
-            for i in range(25000):
-                zip_file.writestr('train/cat.{}.jpeg'.format(i), image)
-        with zipfile.ZipFile(output_files[1], 'w') as zip_file:
-            zif = zipfile.ZipInfo('train/')
-            zip_file.writestr(zif, "")
-            for i in range(12500):
-                zip_file.writestr('test/dog.{}.jpeg'.format(i), image)
-
-    def test_conversion(self):
-        convert_dogs_vs_cats(config.data_path[0], config.data_path[0])
-        output_file = "dogs_vs_cats.hdf5"
-        output_file = os.path.join(config.data_path[0], output_file)
-        with h5py.File(output_file, 'r') as h5:
-            assert numpy.all(h5['targets'][:25000] == 0)
-            assert numpy.all(h5['targets'][25000:] == 1)
-            assert numpy.all(numpy.array(
-                [img for img in h5['image_features'][:]]) == 0)
-            assert numpy.all(h5['image_features_shapes'][:, 0] == 3)
-            assert numpy.all(h5['image_features_shapes'][:, 1:] == 1)
-
-    def test_dataset(self):
-        train = DogsVsCats(('train',))
-        assert train.num_examples == 25000
-        assert_raises(ValueError, DogsVsCats, ('valid',))
-
-        test = DogsVsCats(('test',))
-        stream = DataStream.default_stream(
-            test, iteration_scheme=SequentialScheme(10, 10))
-        data = next(stream.get_epoch_iterator())[0][0]
-        assert data.dtype.kind == 'f'
+def setup():
+    config._old_data_path = config.data_path
+    config.data_path = tempfile.mkdtemp()
+    _make_dummy_data(config.data_path[0])
+
+
+def _make_dummy_data(output_directory):
+    data = six.BytesIO()
+    Image.new('RGB', (1, 1)).save(data, 'JPEG')
+    image = data.getvalue()
+
+    output_files = [os.path.join(output_directory,
+                                 'dogs_vs_cats.{}.zip'.format(set_))
+                    for set_ in ['train', 'test1']]
+    with zipfile.ZipFile(output_files[0], 'w') as zip_file:
+        zif = zipfile.ZipInfo('train/')
+        zip_file.writestr(zif, "")
+        for i in range(25000):
+            zip_file.writestr('train/cat.{}.jpeg'.format(i), image)
+    with zipfile.ZipFile(output_files[1], 'w') as zip_file:
+        zif = zipfile.ZipInfo('train/')
+        zip_file.writestr(zif, "")
+        for i in range(12500):
+            zip_file.writestr('test/dog.{}.jpeg'.format(i), image)
+
+
+def teardown():
+    shutil.rmtree(config.data_path[0])
+    config.data_path = config._old_data_path
+    del config._old_data_path
+
+
+def test_dogs_vs_cats():
+    _test_conversion()
+    _test_dataset()
+
+
+def _test_conversion():
+    convert_dogs_vs_cats(config.data_path[0], config.data_path[0])
+    output_file = "dogs_vs_cats.hdf5"
+    output_file = os.path.join(config.data_path[0], output_file)
+    with h5py.File(output_file, 'r') as h5:
+        assert numpy.all(h5['targets'][:25000] == 0)
+        assert numpy.all(h5['targets'][25000:] == 1)
+        assert numpy.all(numpy.array(
+            [img for img in h5['image_features'][:]]) == 0)
+        assert numpy.all(h5['image_features_shapes'][:, 0] == 3)
+        assert numpy.all(h5['image_features_shapes'][:, 1:] == 1)
+
+
+def _test_dataset():
+    train = DogsVsCats(('train',))
+    assert train.num_examples == 25000
+    assert_raises(ValueError, DogsVsCats, ('valid',))
+
+    test = DogsVsCats(('test',))
+    stream = DataStream.default_stream(
+        test, iteration_scheme=SequentialScheme(10, 10))
+    data = next(stream.get_epoch_iterator())[0][0]
+    assert data.dtype.kind == 'f'
+
+
+test_dogs_vs_cats.setup = setup
+test_dogs_vs_cats.teardown = teardown