def testInit(self):
self.assertEqual(self.t.cutoffs["m1"], (-float("inf"), 2.0))
self.assertEqual(self.t.cutoffs["m2"], (-float("inf"), 2.0))
self.assertEqual(self.t1.cutoffs["m1"], (-float("inf"), 1.0))
self.assertEqual(self.t1.cutoffs["m2"], (-float("inf"), 1.0))
self.assertEqual(self.t2.cutoffs["m1"], (0.0, float("inf")))
self.assertEqual(self.t2.cutoffs["m2"], (0.0, float("inf")))
with self.assertRaises(ValueError):
Winsorize(search_space=None,
observation_features=[],
observation_data=[])
def testInit(self):
self.assertEqual(self.t.percentiles["m1"], (0.0, 2.0))
self.assertEqual(self.t.percentiles["m2"], (0.0, 2.0))
self.assertEqual(self.t1.percentiles["m1"], (0.0, 1.0))
self.assertEqual(self.t1.percentiles["m2"], (0.0, 1.0))
self.assertEqual(self.t2.percentiles["m1"], (0.0, 2.0))
self.assertEqual(self.t2.percentiles["m2"], (0.0, 2.0))
with self.assertRaises(ValueError):
Winsorize(search_space=None,
observation_features=[],
observation_data=[])
def setUp(self):
self.obsd1 = ObservationData(
metric_names=["m1", "m2", "m2"],
means=np.array([0.0, 0.0, 1.0]),
covariance=np.array([[1.0, 0.2, 0.4], [0.2, 2.0, 0.8],
[0.4, 0.8, 3.0]]),
)
self.obsd2 = ObservationData(
metric_names=["m1", "m1", "m2", "m2"],
means=np.array([1.0, 2.0, 2.0, 1.0]),
covariance=np.array([
[1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.2, 0.4],
[0.0, 0.2, 2.0, 0.8],
[0.0, 0.4, 0.8, 3.0],
]),
)
self.t = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.2},
)
self.t1 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.8},
)
self.t2 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_lower": 0.2},
)
def testValueError(self):
with self.assertRaises(ValueError):
Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": 0.8,
"percentile_bounds": {
"m1": (0.1, 0.2, 0.3)
}, # Too many inputs..
},
)
def testPrintDeprecationWarning(self):
warnings.simplefilter("always", DeprecationWarning)
with warnings.catch_warnings(record=True) as ws:
Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.2},
)
self.assertTrue(
"Winsorization received an out-of-date `transform_config`, containing "
"the following deprecated keys: {'winsorization_upper'}. Please "
"update the config according to the docs of "
"`ax.modelbridge.transforms.winsorize.Winsorize`." in
[str(w.message) for w in ws])
def setUp(self):
self.obsd1 = ObservationData(
metric_names=["m1", "m2", "m2"],
means=np.array([0.0, 0.0, 1.0]),
covariance=np.array([[1.0, 0.2, 0.4], [0.2, 2.0, 0.8],
[0.4, 0.8, 3.0]]),
)
self.obsd2 = ObservationData(
metric_names=["m1", "m1", "m2", "m2"],
means=np.array([1.0, 2.0, 2.0, 1.0]),
covariance=np.array([
[1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.2, 0.4],
[0.0, 0.2, 2.0, 0.8],
[0.0, 0.4, 0.8, 3.0],
]),
)
self.t = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.2},
)
self.t1 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.8},
)
self.t2 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_lower": 0.2},
)
self.t3 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_upper": 0.6,
"percentile_bounds": {
"m1": (None, None),
"m2": (None, 1.9),
},
},
)
self.t4 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": 0.8,
"percentile_bounds": {
"m1": (None, None),
"m2": (0.3, None),
},
},
)
self.obsd3 = ObservationData(
metric_names=["m3", "m3", "m3", "m3"],
means=np.array([0.0, 1.0, 5.0, 3.0]),
covariance=np.eye(4),
)
self.t5 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[
deepcopy(self.obsd1),
deepcopy(self.obsd2),
deepcopy(self.obsd3),
],
config={
"winsorization_lower": {
"m2": 0.4
},
"winsorization_upper": {
"m1": 0.6
},
},
)
self.t6 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": {
"m2": 0.4
},
"winsorization_upper": {
"m1": 0.6
},
"percentile_bounds": {
"m1": (None, None),
"m2": (0.0, None), # This should leave m2 untouched
},
},
)
class WinsorizeTransformTestLegacy(TestCase):
def setUp(self):
self.obsd1 = ObservationData(
metric_names=["m1", "m2", "m2"],
means=np.array([0.0, 0.0, 1.0]),
covariance=np.array([[1.0, 0.2, 0.4], [0.2, 2.0, 0.8],
[0.4, 0.8, 3.0]]),
)
self.obsd2 = ObservationData(
metric_names=["m1", "m1", "m2", "m2"],
means=np.array([1.0, 2.0, 2.0, 1.0]),
covariance=np.array([
[1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.2, 0.4],
[0.0, 0.2, 2.0, 0.8],
[0.0, 0.4, 0.8, 3.0],
]),
)
self.t = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.2},
)
self.t1 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.8},
)
self.t2 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_lower": 0.2},
)
self.t3 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_upper": 0.6,
"percentile_bounds": {
"m1": (None, None),
"m2": (None, 1.9),
},
},
)
self.t4 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": 0.8,
"percentile_bounds": {
"m1": (None, None),
"m2": (0.3, None),
},
},
)
self.obsd3 = ObservationData(
metric_names=["m3", "m3", "m3", "m3"],
means=np.array([0.0, 1.0, 5.0, 3.0]),
covariance=np.eye(4),
)
self.t5 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[
deepcopy(self.obsd1),
deepcopy(self.obsd2),
deepcopy(self.obsd3),
],
config={
"winsorization_lower": {
"m2": 0.4
},
"winsorization_upper": {
"m1": 0.6
},
},
)
self.t6 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": {
"m2": 0.4
},
"winsorization_upper": {
"m1": 0.6
},
"percentile_bounds": {
"m1": (None, None),
"m2": (0.0, None), # This should leave m2 untouched
},
},
)
def testPrintDeprecationWarning(self):
warnings.simplefilter("always", DeprecationWarning)
with warnings.catch_warnings(record=True) as ws:
Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.2},
)
self.assertTrue(
"Winsorization received an out-of-date `transform_config`, containing "
"the following deprecated keys: {'winsorization_upper'}. Please "
"update the config according to the docs of "
"`ax.modelbridge.transforms.winsorize.Winsorize`." in
[str(w.message) for w in ws])
def testInit(self):
self.assertEqual(self.t.cutoffs["m1"], (-float("inf"), 2.0))
self.assertEqual(self.t.cutoffs["m2"], (-float("inf"), 2.0))
self.assertEqual(self.t1.cutoffs["m1"], (-float("inf"), 1.0))
self.assertEqual(self.t1.cutoffs["m2"], (-float("inf"), 1.0))
self.assertEqual(self.t2.cutoffs["m1"], (0.0, float("inf")))
self.assertEqual(self.t2.cutoffs["m2"], (0.0, float("inf")))
with self.assertRaises(ValueError):
Winsorize(search_space=None,
observation_features=[],
observation_data=[])
def testTransformObservations(self):
observation_data = self.t1.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [0.0, 0.0, 1.0])
observation_data = self.t1.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 1.0, 1.0, 1.0])
observation_data = self.t2.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [0.0, 0.0, 1.0])
observation_data = self.t2.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 2.0, 2.0, 1.0])
def testInitPercentileBounds(self):
self.assertEqual(self.t3.cutoffs["m1"], (-float("inf"), 1.0))
self.assertEqual(self.t3.cutoffs["m2"], (-float("inf"), 1.9))
self.assertEqual(self.t4.cutoffs["m1"], (1.0, float("inf")))
self.assertEqual(self.t4.cutoffs["m2"], (0.3, float("inf")))
def testValueError(self):
with self.assertRaises(ValueError):
Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": 0.8,
"percentile_bounds": {
"m1": (0.1, 0.2, 0.3)
}, # Too many inputs..
},
)
def testTransformObservationsPercentileBounds(self):
observation_data = self.t3.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [0.0, 0.0, 1.0])
observation_data = self.t3.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 1.0, 1.9, 1.0])
observation_data = self.t4.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [1.0, 0.3, 1.0])
observation_data = self.t4.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 2.0, 2.0, 1.0])
def testTransformObservationsDifferentLowerUpper(self):
observation_data = self.t5.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertEqual(self.t5.cutoffs["m1"], (-float("inf"), 1.0))
self.assertEqual(self.t5.cutoffs["m2"], (1.0, float("inf")))
self.assertEqual(self.t5.cutoffs["m3"], (-float("inf"), float("inf")))
self.assertListEqual(list(observation_data.means),
[1.0, 1.0, 2.0, 1.0])
# Nothing should happen to m3
observation_data = self.t5.transform_observation_data(
[deepcopy(self.obsd3)], [])[0]
self.assertListEqual(list(observation_data.means),
[0.0, 1.0, 5.0, 3.0])
# With percentile_bounds
observation_data = self.t6.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertEqual(self.t6.cutoffs["m1"], (-float("inf"), 1.0))
self.assertEqual(self.t6.cutoffs["m2"], (0.0, float("inf")))
self.assertListEqual(list(observation_data.means),
[1.0, 1.0, 2.0, 1.0])
class WinsorizeTransformTest(TestCase):
def setUp(self):
self.obsd1 = ObservationData(
metric_names=["m1", "m2", "m2"],
means=np.array([0.0, 0.0, 1.0]),
covariance=np.array([[1.0, 0.2, 0.4], [0.2, 2.0, 0.8],
[0.4, 0.8, 3.0]]),
)
self.obsd2 = ObservationData(
metric_names=["m1", "m1", "m2", "m2"],
means=np.array([1.0, 2.0, 2.0, 1.0]),
covariance=np.array([
[1.0, 0.0, 0.0, 0.0],
[0.0, 1.0, 0.2, 0.4],
[0.0, 0.2, 2.0, 0.8],
[0.0, 0.4, 0.8, 3.0],
]),
)
self.t = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.2},
)
self.t1 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_upper": 0.8},
)
self.t2 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={"winsorization_lower": 0.2},
)
self.t3 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_upper": 0.6,
"percentile_bounds": {
"m1": (None, None),
"m2": (None, 1.9),
},
},
)
self.t4 = Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": 0.8,
"percentile_bounds": {
"m1": (None, None),
"m2": (0.3, None),
},
},
)
def testInit(self):
self.assertEqual(self.t.percentiles["m1"], (0.0, 2.0))
self.assertEqual(self.t.percentiles["m2"], (0.0, 2.0))
self.assertEqual(self.t1.percentiles["m1"], (0.0, 1.0))
self.assertEqual(self.t1.percentiles["m2"], (0.0, 1.0))
self.assertEqual(self.t2.percentiles["m1"], (0.0, 2.0))
self.assertEqual(self.t2.percentiles["m2"], (0.0, 2.0))
with self.assertRaises(ValueError):
Winsorize(search_space=None,
observation_features=[],
observation_data=[])
def testTransformObservations(self):
observation_data = self.t1.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [0.0, 0.0, 1.0])
observation_data = self.t1.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 1.0, 1.0, 1.0])
observation_data = self.t2.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [0.0, 0.0, 1.0])
observation_data = self.t2.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 2.0, 2.0, 1.0])
def testInitPercentileBounds(self):
self.assertEqual(self.t3.percentiles["m1"], (0.0, 1.0))
self.assertEqual(self.t3.percentiles["m2"], (0.0, 1.9))
self.assertEqual(self.t4.percentiles["m1"], (1.0, 2.0))
self.assertEqual(self.t4.percentiles["m2"], (0.3, 2.0))
def testValueError(self):
with self.assertRaises(ValueError):
Winsorize(
search_space=None,
observation_features=None,
observation_data=[deepcopy(self.obsd1),
deepcopy(self.obsd2)],
config={
"winsorization_lower": 0.8,
"percentile_bounds": {
"m1": (0.1, 0.2, 0.3)
}, # Too many inputs..
},
)
def testTransformObservationsPercentileBounds(self):
observation_data = self.t3.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [0.0, 0.0, 1.0])
observation_data = self.t3.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 1.0, 1.9, 1.0])
observation_data = self.t4.transform_observation_data(
[deepcopy(self.obsd1)], [])[0]
self.assertListEqual(list(observation_data.means), [1.0, 0.3, 1.0])
observation_data = self.t4.transform_observation_data(
[deepcopy(self.obsd2)], [])[0]
self.assertListEqual(list(observation_data.means),
[1.0, 2.0, 2.0, 1.0])