def test_result_is_the_same_with_and_without_monitor(self):
monitor = AttributeMonitor(["xcorr.var_"])
res = self.xcorr.transform(self.x, self.y, monitor=monitor)
xcorr_alt = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.seed
)
res_alt = xcorr_alt.transform(self.x, self.y)
np.testing.assert_allclose(res, res_alt)
def setUp(self):
self.n_models = 2
self.n_features = 4
self.n_samples = 34
self.rng = np.random.default_rng(1)
self.x = self.rng.normal(size=(self.n_samples, self.n_features))
self.y = self.rng.normal(size=self.n_samples)
self.xcorr = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.rng
)
def test_results_same_regardless_of_chunk_hint(self):
names = ["xcorr.var_"]
res, history = self.xcorr.transform(self.x, self.y, monitor=names, chunk_hint=5)
xcorr_alt = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.seed
)
res_alt, history_alt = xcorr_alt.transform(
self.x, self.y, monitor=names, chunk_hint=13
)
np.testing.assert_allclose(res, res_alt)
np.testing.assert_allclose(history.xcorr.var_, history_alt.xcorr.var_)
def test_xcorr_constructor_called_with_correct_n_features(self):
n_models = 3
n_features = 4
with mock.patch("bioslds.xcorr_regressor.OnlineCrosscorrelation") as MockXcorr:
CrosscorrelationRegressor(n_models, n_features)
self.assertEqual(MockXcorr.call_args[0][0], n_features)
def test_monitor_as_sequence(self):
names = ["xcorr.var_", "nsm.weights_"]
_, history = self.xcorr.transform(
self.x, self.y, monitor=names, return_history=True
)
monitor = AttributeMonitor(names)
xcorr_alt = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.seed
)
_, history_alt = xcorr_alt.transform(
self.x, self.y, monitor=monitor, return_history=True
)
np.testing.assert_allclose(history.xcorr.var_, history_alt.xcorr.var_)
np.testing.assert_allclose(history.nsm.weights_, history_alt.nsm.weights_)
def test_nsm_constructor_called_with_non_negative_true(self):
n_models = 3
n_features = 4
with mock.patch("bioslds.xcorr_regressor.NonRecurrent") as MockNSM:
CrosscorrelationRegressor(n_models, n_features)
self.assertIn("non_negative", MockNSM.call_args[1])
self.assertTrue(MockNSM.call_args[1]["non_negative"])
def test_nsm_constructor_called_with_correct_number_of_dimensions(self):
n_models = 3
n_features = 4
with mock.patch("bioslds.xcorr_regressor.NonRecurrent") as MockNSM:
CrosscorrelationRegressor(n_models, n_features)
self.assertEqual(MockNSM.call_args[0][0], n_features)
self.assertEqual(MockNSM.call_args[0][1], n_models)
def test_nsm_kws_passed_to_nsm_constructor(self):
n_models = 3
n_features = 4
foo_value = 1.3
with mock.patch("bioslds.xcorr_regressor.NonRecurrent") as MockNSM:
CrosscorrelationRegressor(n_models, n_features, nsm_kws={"foo": foo_value})
self.assertIn("foo", MockNSM.call_args[1])
self.assertEqual(MockNSM.call_args[1]["foo"], foo_value)
def test_rng_passed_to_nsm_constructor(self):
n_models = 3
n_features = 4
rng = mock.Mock()
with mock.patch("bioslds.xcorr_regressor.NonRecurrent") as MockNSM:
CrosscorrelationRegressor(n_models, n_features, rng=rng)
self.assertIn("rng", MockNSM.call_args[1])
self.assertIs(MockNSM.call_args[1]["rng"], rng)
def test_nsm_constructor_called_with_correct_rate(self):
n_models = 3
n_features = 4
nsm_rate = 0.987
with mock.patch("bioslds.xcorr_regressor.NonRecurrent") as MockNSM:
CrosscorrelationRegressor(n_models, n_features, nsm_rate=nsm_rate)
self.assertIn("rate", MockNSM.call_args[1])
self.assertEqual(MockNSM.call_args[1]["rate"], nsm_rate)
def test_xcorr_constructor_called_with_correct_rate(self):
n_models = 3
n_features = 4
xcorr_rate = 0.124
with mock.patch("bioslds.xcorr_regressor.OnlineCrosscorrelation") as MockXcorr:
CrosscorrelationRegressor(n_models, n_features, xcorr_rate=xcorr_rate)
self.assertIn("rate", MockXcorr.call_args[1])
self.assertEqual(MockXcorr.call_args[1]["rate"], xcorr_rate)
def setUp(self):
self.n_models = 4
self.n_features = 3
self.xcorr = CrosscorrelationRegressor(self.n_models, self.n_features, rng=1)
def test_call_xcorr_constructor(self):
with mock.patch("bioslds.xcorr_regressor.OnlineCrosscorrelation") as MockXcorr:
CrosscorrelationRegressor(2, 3)
MockXcorr.assert_called()
def test_call_nsm_constructor(self):
with mock.patch("bioslds.xcorr_regressor.NonRecurrent") as MockNSM:
CrosscorrelationRegressor(2, 3)
MockNSM.assert_called()
class TestCrosscorrelationRegressorTransformMonitor(unittest.TestCase):
def setUp(self):
self.n_models = 3
self.n_features = 3
self.n_samples = 71
self.rng = np.random.default_rng(2)
self.x = self.rng.normal(size=(self.n_samples, self.n_features))
self.y = self.rng.normal(size=self.n_samples)
self.seed = 4
self.xcorr = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.seed
)
def test_with_monitor_object_history_not_returned_by_default(self):
monitor = AttributeMonitor(["xcorr.coef_"])
res = self.xcorr.transform(self.x, self.y, monitor=monitor)
self.assertEqual(np.shape(res), (self.n_samples, self.n_models))
def test_history_returned_when_return_history_is_true(self):
monitor = AttributeMonitor(["xcorr.coef_"])
_, history = self.xcorr.transform(
self.x, self.y, monitor=monitor, return_history=True
)
self.assertIs(history, monitor.history_)
def test_result_is_the_same_with_and_without_monitor(self):
monitor = AttributeMonitor(["xcorr.var_"])
res = self.xcorr.transform(self.x, self.y, monitor=monitor)
xcorr_alt = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.seed
)
res_alt = xcorr_alt.transform(self.x, self.y)
np.testing.assert_allclose(res, res_alt)
def test_monitor_as_sequence(self):
names = ["xcorr.var_", "nsm.weights_"]
_, history = self.xcorr.transform(
self.x, self.y, monitor=names, return_history=True
)
monitor = AttributeMonitor(names)
xcorr_alt = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.seed
)
_, history_alt = xcorr_alt.transform(
self.x, self.y, monitor=monitor, return_history=True
)
np.testing.assert_allclose(history.xcorr.var_, history_alt.xcorr.var_)
np.testing.assert_allclose(history.nsm.weights_, history_alt.nsm.weights_)
def test_when_monitor_sequence_return_history_is_forced_true(self):
names = ["xcorr.var_", "nsm.weights_"]
res = self.xcorr.transform(self.x, self.y, monitor=names)
self.assertEqual(len(res), 2)
def test_out_matches_monitor_nsm_output(self):
res, history = self.xcorr.transform(self.x, self.y, monitor=["nsm.output_"])
np.testing.assert_allclose(res, history.nsm.output_)
def test_results_same_regardless_of_chunk_hint(self):
names = ["xcorr.var_"]
res, history = self.xcorr.transform(self.x, self.y, monitor=names, chunk_hint=5)
xcorr_alt = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.seed
)
res_alt, history_alt = xcorr_alt.transform(
self.x, self.y, monitor=names, chunk_hint=13
)
np.testing.assert_allclose(res, res_alt)
np.testing.assert_allclose(history.xcorr.var_, history_alt.xcorr.var_)
def test_passes_chunk_hint_to_nsm_transform(self):
chunk = 13
with mock.patch.object(
bioslds.xcorr_regressor.NonRecurrent, "transform"
) as mock_transform:
mock_transform.side_effect = lambda X, **kwargs: (
self.rng.normal(size=(len(X), self.n_models))
)
self.xcorr.transform(
self.x, self.y, monitor=["xcorr.var_"], chunk_hint=chunk
)
self.assertIn("chunk_hint", mock_transform.call_args[1])
self.assertEqual(mock_transform.call_args[1]["chunk_hint"], chunk)
def test_passes_chunk_hint_to_xcorr_transform(self):
chunk = 7
with mock.patch.object(
bioslds.xcorr_regressor.OnlineCrosscorrelation, "transform"
) as mock_transform:
mock_transform.side_effect = lambda X, y, **kwargs: (
self.rng.normal(size=(len(X), self.n_features))
)
self.xcorr.transform(
self.x, self.y, monitor=["nsm.weights_"], chunk_hint=chunk
)
self.assertIn("chunk_hint", mock_transform.call_args[1])
self.assertEqual(mock_transform.call_args[1]["chunk_hint"], chunk)
def test_return_history_ignored_when_monitor_is_none(self):
res = self.xcorr.transform(self.x, self.y, return_history=True)
self.assertEqual(np.shape(res), (self.n_samples, self.n_models))
def test_return_history_ignored_when_monitor_is_none_but_progress_is_not(self):
res = self.xcorr.transform(
self.x, self.y, progress=mock.MagicMock(), return_history=True
)
self.assertEqual(np.shape(res), (self.n_samples, self.n_models))
def test_progress_called_with_monitor(self):
mock_progress = mock.MagicMock()
self.xcorr.transform(
self.x, self.y, progress=mock_progress, monitor=["xcorr.var_"]
)
mock_progress.assert_called()
def test_progress_called_without_monitor(self):
mock_progress = mock.MagicMock()
self.xcorr.transform(self.x, self.y, progress=mock_progress)
mock_progress.assert_called()
class TestCrosscorrelationRegressorTransform(unittest.TestCase):
def setUp(self):
self.n_models = 2
self.n_features = 4
self.n_samples = 34
self.rng = np.random.default_rng(1)
self.x = self.rng.normal(size=(self.n_samples, self.n_features))
self.y = self.rng.normal(size=self.n_samples)
self.xcorr = CrosscorrelationRegressor(
self.n_models, self.n_features, rng=self.rng
)
def test_calls_xcorr_transform(self):
with mock.patch.object(
bioslds.xcorr_regressor.OnlineCrosscorrelation, "transform"
) as mock_transform:
mock_transform.return_value = self.x
self.xcorr.transform(self.x, self.y)
mock_transform.assert_called()
def test_calls_xcorr_transform_with_proper_args(self):
with mock.patch.object(
bioslds.xcorr_regressor.OnlineCrosscorrelation, "transform"
) as mock_transform:
mock_transform.return_value = self.x
self.xcorr.transform(self.x, self.y)
self.assertIn("X", mock_transform.call_args[1])
self.assertIn("y", mock_transform.call_args[1])
np.testing.assert_allclose(mock_transform.call_args[1]["X"], self.x)
np.testing.assert_allclose(mock_transform.call_args[1]["y"], self.y)
def test_calls_nsm_transform(self):
with mock.patch.object(
bioslds.xcorr_regressor.NonRecurrent, "transform"
) as mock_transform:
mock_transform.return_value = self.rng.normal(
size=(self.n_samples, self.n_models)
)
self.xcorr.transform(self.x, self.y)
mock_transform.assert_called()
def test_calls_nsm_transform_with_proper_out_from_xcorr_transform(self):
with mock.patch.object(
bioslds.xcorr_regressor.OnlineCrosscorrelation, "transform"
) as mock_xcorr_transform, mock.patch.object(
bioslds.xcorr_regressor.NonRecurrent, "transform"
) as mock_nsm_transform:
trafo_ret = self.rng.normal(size=(self.n_samples, self.n_features))
mock_xcorr_transform.return_value = trafo_ret
mock_nsm_transform.return_value = self.rng.normal(
size=(self.n_samples, self.n_models)
)
self.xcorr.transform(self.x, self.y)
self.assertIn("X", mock_nsm_transform.call_args[1])
np.testing.assert_allclose(mock_nsm_transform.call_args[1]["X"], trafo_ret)
def test_returns_out_from_nsm_transform(self):
with mock.patch.object(
bioslds.xcorr_regressor.NonRecurrent, "transform"
) as mock_transform:
nsm_out = self.rng.normal(size=(self.n_samples, self.n_models))
mock_transform.return_value = nsm_out
res = self.xcorr.transform(self.x, self.y)
np.testing.assert_allclose(res, nsm_out)
def test_passes_chunk_hint_to_nsm_transform(self):
chunk = 13
with mock.patch.object(
bioslds.xcorr_regressor.NonRecurrent, "transform"
) as mock_transform:
mock_transform.side_effect = lambda X, **kwargs: self.rng.normal(
size=(len(X), self.n_models)
)
self.xcorr.transform(self.x, self.y, chunk_hint=chunk)
self.assertIn("chunk_hint", mock_transform.call_args[1])
self.assertEqual(mock_transform.call_args[1]["chunk_hint"], chunk)
def test_passes_chunk_hint_to_xcorr_transform(self):
chunk = 7
with mock.patch.object(
bioslds.xcorr_regressor.OnlineCrosscorrelation, "transform"
) as mock_transform:
trafo_ret = lambda X, y, **kwargs: self.rng.normal(
size=(len(X), self.n_features)
)
mock_transform.side_effect = trafo_ret
self.xcorr.transform(self.x, self.y, chunk_hint=chunk)
self.assertIn("chunk_hint", mock_transform.call_args[1])
self.assertEqual(mock_transform.call_args[1]["chunk_hint"], chunk)
