Module: tfds.testing | TensorFlow Datasets (original) (raw)
Testing utilities.
Classes
class DatasetBuilderTestCase: Inherit this class to test your DatasetBuilder class.
class DummyBeamDataset: Minimal beam DatasetBuilder.
class DummyDataset: Minimal DatasetBuilder.
class DummyDatasetCollection: Minimal Dataset Collection builder.
class DummyDatasetSharedGenerator: Test DatasetBuilder.
class DummyMnist: Test DatasetBuilder.
class DummyParser: To mock example_parser.ExampleParser.
class DummySerializer: To mock example_serializer.ExampleSerializer.
class FeatureExpectationItem: Test item of a FeatureExpectation.
class FeatureExpectationsTestCase: Tests FeatureExpectations with full encode-decode.
class MockFs: This util wraps mock for the tf.io.gfile / epath.Path API.
class MockPolicy: Strategy to use with tfds.testing.mock_data to mock the dataset.
class PickableDataSourceMock: Makes MagicMock pickable in order to work with multiprocessing in Grain.
class RaggedConstant: Container of tf.ragged.constant values.
class SubTestCase: Adds subTest() context manager to the TestCase if supported.
class TestCase: Base TestCase to be used for all tests.
Functions
assert_features_equal(...): Asserts that the 2 nested FeatureConnector structure match.
make_tmp_dir(...): Make a temporary directory.
mock_data(...): Mock tfds to generate random data.
mock_kaggle_api(...): Mock out the kaggle CLI.
rm_tmp_dir(...): Rm temporary directory.
run_in_graph_and_eager_modes(...): Execute the decorated test in both graph mode and eager mode.
test_main(...): Entrypoint for tests.
tmp_dir(...): Context manager for a temporary directory.