def test_multi_label_deserialize_without_error():
encoder = encoders.MultiLabelEncoder()
dataset = tf.data.Dataset.from_tensor_slices([1, 2]).batch(32)
encoder = preprocessors.deserialize(preprocessors.serialize(encoder))
assert encoder.transform(dataset) is dataset
def test_multi_label_transform_dataset_doesnt_change():
encoder = encoders.MultiLabelEncoder()
dataset = tf.data.Dataset.from_tensor_slices([1, 2]).batch(32)
assert encoder.transform(dataset) is dataset
def test_multi_label_postprocess_to_one_hot_labels():
encoder = encoders.MultiLabelEncoder()
y = encoder.postprocess(np.random.rand(10, 3))
assert set(y.flatten().tolist()) == set([1, 0])