def test_add_constraint(self):
init = self.sk_pca.hyperparam_schema()
expected = {
'allOf': [{
'type': 'object',
'properties': {}
}, {
'anyOf': [{
'type': 'object',
'properties': {
'n_components': {
'not': {
'enum': ['mle']
},
}
},
}, {
'type': 'object',
'properties': {
'svd_solver': {
'enum': ['full', 'auto']
},
}
}]
}]
}
foo = self.sk_pca.customize_schema(constraint=schemas.AnyOf([
schemas.Object(
{'n_components': schemas.Not(schemas.Enum(['mle']))}),
schemas.Object({'svd_solver': schemas.Enum(['full', 'auto'])})
]))
self.assertEqual(foo.hyperparam_schema(), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.sk_pca.hyperparam_schema(), init)
def test_override_null_param(self):
init = self.ll_pca.hyperparam_schema('n_components')
expected = {'enum': [None]}
foo = self.ll_pca.customize_schema(n_components=schemas.Null())
self.assertEqual(foo.hyperparam_schema('n_components'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca.hyperparam_schema('n_components'), init)
def test_override_output2(self):
init_output_schema = self.sk_pca.get_schema('output')
pca_output = schemas.AnyOf([
schemas.Array(schemas.Array(schemas.Float())),
schemas.Array(schemas.Float())
])
expected = {
'anyOf': [{
'type': 'array',
'items': {
'type': 'array',
'items': {
'type': 'number'
}
}
}, {
'type': 'array',
'items': {
'type': 'number'
}
}]
}
foo = self.sk_pca.customize_schema(output=pca_output)
self.assertEqual(foo.get_schema('output'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.sk_pca.get_schema('output'), init_output_schema)
def test_override_bool_param_ll(self):
init = self.ll_pca.hyperparam_schema('whiten')
expected = {'default': True, 'type': 'boolean'}
foo = self.ll_pca.customize_schema(whiten=schemas.Bool(default=True))
self.assertEqual(foo.hyperparam_schema('whiten'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca.hyperparam_schema('whiten'), init)
self.assertRaises(Exception, self.ll_pca.customize_schema, whitenX={})
def test_override_enum_param(self):
init = self.ll_pca.hyperparam_schema('svd_solver')
expected = {'default': 'full', 'enum': ['auto', 'full']}
foo = self.ll_pca.customize_schema(
svd_solver=schemas.Enum(default='full', values=['auto', 'full']))
self.assertEqual(foo.hyperparam_schema('svd_solver'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca.hyperparam_schema('svd_solver'), init)
def test_override_output(self):
init_output_schema = self.sk_pca.get_schema('output')
pca_output = self.ll_pca.get_schema('output')
foo = self.sk_pca.customize_schema(output=schemas.JSON(pca_output))
self.assertEqual(foo.get_schema('output'), pca_output)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.sk_pca.get_schema('output'), init_output_schema)
self.assertRaises(Exception, self.sk_pca.customize_schema, output={})
self.assertRaises(Exception,
self.sk_pca.customize_schema,
output_foo=pca_output)
def test_override_json_param(self):
init = self.ll_pca.hyperparam_schema('tol')
expected = {
'description': 'Tol',
'type': 'number',
'minimum': 0.2,
'default': 1.0
}
foo = self.ll_pca.customize_schema(tol=schemas.JSON(expected))
self.assertEqual(foo.hyperparam_schema('tol'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca.hyperparam_schema('tol'), init)
def test_override_tags(self):
init = self.ll_pca._schemas['tags']
tags = {
'pre': ['~categoricals'],
'op': ['estimator', 'classifier', 'interpretable'],
'post': ['probabilities']
}
foo = self.ll_pca.customize_schema(tags=tags)
self.assertEqual(foo._schemas['tags'], tags)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca._schemas['tags'], init)
self.assertRaises(Exception, self.sk_pca.customize_schema, tags=42)
def test_override_int_param(self):
init = self.ll_pca.hyperparam_schema('iterated_power')
expected = {
'default': 1,
'type': 'integer',
'minimum': -10,
'maximum': 10,
'exclusiveMaximum': True,
'exclusiveMinimum': False
}
foo = self.ll_pca.customize_schema(iterated_power=schemas.Int(
default=1, min=-10, max=10, exclusiveMax=True, exclusiveMin=False))
self.assertEqual(foo.hyperparam_schema('iterated_power'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca.hyperparam_schema('iterated_power'), init)
def test_override_array_param(self):
init = self.sk_pca.hyperparam_schema('copy')
expected = {
'type': 'array',
'minItems': 1,
'maxItems': 20,
'items': {
'type': 'integer'
}
}
foo = self.sk_pca.customize_schema(
copy=schemas.Array(minItems=1, maxItems=20, items=schemas.Int()))
self.assertEqual(foo.hyperparam_schema('copy'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.sk_pca.hyperparam_schema('copy'), init)
def test_override_relevant(self):
init = self.ll_pca.hyperparam_schema(
)['allOf'][0]['relevantToOptimizer']
expected = ['svd_solver']
foo = self.ll_pca.customize_schema(relevantToOptimizer=['svd_solver'])
self.assertEqual(
foo.hyperparam_schema()['allOf'][0]['relevantToOptimizer'],
expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(
self.ll_pca.hyperparam_schema()['allOf'][0]['relevantToOptimizer'],
init)
self.assertRaises(Exception,
self.sk_pca.customize_schema,
relevantToOptimizer={})
def test_override_any_param(self):
init = self.ll_pca.hyperparam_schema('iterated_power')
expected = {
'anyOf': [{
'type': 'integer'
}, {
'enum': ['auto', 'full']
}],
'default': 'auto'
}
foo = self.ll_pca.customize_schema(iterated_power=schemas.AnyOf(
[schemas.Int(), schemas.Enum(['auto', 'full'])], default='auto'))
self.assertEqual(foo.hyperparam_schema('iterated_power'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca.hyperparam_schema('iterated_power'), init)
def test_override_float_param(self):
init = self.ll_pca.hyperparam_schema('tol')
expected = {
'default': 0.1,
'type': 'number',
'minimum': -10,
'maximum': 10,
'exclusiveMaximum': True,
'exclusiveMinimum': False
}
foo = self.ll_pca.customize_schema(
tol=schemas.Float(default=0.1,
min=-10,
max=10,
exclusiveMax=True,
exclusiveMin=False))
self.assertEqual(foo.hyperparam_schema('tol'), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.ll_pca.hyperparam_schema('tol'), init)
def test_override_object_param(self):
init = self.sk_pca.get_schema('input_fit')
expected = {
'type': 'object',
'required': ['X'],
'additionalProperties': False,
'properties': {
'X': {
'type': 'array',
'items': {
'type': 'number'
}
}
}
}
foo = self.sk_pca.customize_schema(
input_fit=schemas.Object(required=['X'],
additionalProperties=False,
X=schemas.Array(schemas.Float())))
self.assertEqual(foo.get_schema('input_fit'), expected)
helpers.validate_is_schema(foo.get_schema('input_fit'))
self.assertEqual(self.sk_pca.get_schema('input_fit'), init)
def test_feature_preprocessor(self):
X_train, y_train = self.X_train, self.y_train
X_test, y_test = self.X_test, self.y_test
import importlib
module_name = ".".join(fproc_name.split('.')[0:-1])
class_name = fproc_name.split('.')[-1]
module = importlib.import_module(module_name)
class_ = getattr(module, class_name)
fproc = class_()
from lale.lib.sklearn.one_hot_encoder import OneHotEncoderImpl
if isinstance(fproc._impl, OneHotEncoderImpl):
#fproc = OneHotEncoder(handle_unknown = 'ignore')
#remove the hack when this is fixed
fproc = PCA()
#test_schemas_are_schemas
from lale.helpers import validate_is_schema
validate_is_schema(fproc.input_schema_fit())
validate_is_schema(fproc.input_schema_transform())
validate_is_schema(fproc.output_schema())
validate_is_schema(fproc.hyperparam_schema())
#test_init_fit_transform
trained = fproc.fit(self.X_train, self.y_train)
predictions = trained.transform(self.X_test)
#test_predict_on_trainable
trained = fproc.fit(X_train, y_train)
fproc.transform(X_train)
#test_to_json
fproc.to_json()
#test_in_a_pipeline
#This test assumes that the output of feature processing is compatible with LogisticRegression
from lale.lib.sklearn import LogisticRegression
pipeline = fproc >> LogisticRegression()
trained = pipeline.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
#Tune the pipeline with LR using HyperoptClassifier
from lale.lib.lale import HyperoptClassifier
hyperopt = HyperoptClassifier(model=pipeline, max_evals=1)
trained = hyperopt.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
def test_encoder(self):
import importlib
module_name = ".".join(encoder_name.split('.')[0:-1])
class_name = encoder_name.split('.')[-1]
module = importlib.import_module(module_name)
class_ = getattr(module, class_name)
encoder = class_()
#test_schemas_are_schemas
from lale.helpers import validate_is_schema
validate_is_schema(encoder.input_schema_fit())
validate_is_schema(encoder.input_schema_predict())
validate_is_schema(encoder.output_schema())
validate_is_schema(encoder.hyperparam_schema())
#test_init_fit_transform
trained = encoder.fit(self.X_train, self.y_train)
transformed = trained.transform(self.X_train)
def test_regressor(self):
X_train, y_train = self.X_train, self.y_train
X_test, y_test = self.X_test, self.y_test
import importlib
module_name = ".".join(clf_name.split('.')[0:-1])
class_name = clf_name.split('.')[-1]
module = importlib.import_module(module_name)
class_ = getattr(module, class_name)
regr = class_()
#test_schemas_are_schemas
from lale.helpers import validate_is_schema
validate_is_schema(regr.input_schema_fit())
validate_is_schema(regr.input_schema_predict())
validate_is_schema(regr.output_schema())
validate_is_schema(regr.hyperparam_schema())
#test_init_fit_predict
trained = regr.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
#test_predict_on_trainable
trained = regr.fit(X_train, y_train)
regr.predict(X_train)
#test_to_json
regr.to_json()
#test_in_a_pipeline
pipeline = NoOp() >> regr
trained = pipeline.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
#test_with_hyperopt
from lale.lib.sklearn.ridge import RidgeImpl
if not isinstance(regr._impl, RidgeImpl):
from lale.lib.lale import HyperoptRegressor
hyperopt = HyperoptRegressor(model=pipeline, max_evals=1)
trained = hyperopt.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
'post': []
},
'properties': {
'input_fit': _input_schema_fit,
'input_predict': _input_schema_predict,
'output': _output_schema,
'hyperparams': _hyperparams_schema
}
}
ResNet50 = make_operator(ResNet50Impl, _combined_schemas)
if __name__ == "__main__":
import torchvision.datasets as datasets
from lale.helpers import validate_is_schema
validate_is_schema(_combined_schemas)
transform_train = transforms.Compose([
transforms.RandomCrop(32, padding=4),
transforms.RandomHorizontalFlip(),
transforms.ToTensor(),
transforms.Normalize(mean=[n / 255. for n in [129.3, 124.1, 112.4]],
std=[n / 255. for n in [68.2, 65.4, 70.4]])
]) # meanstd transformation
transform_test = transforms.Compose([
transforms.ToTensor(),
transforms.Normalize(mean=[n / 255. for n in [129.3, 124.1, 112.4]],
std=[n / 255. for n in [68.2, 65.4, 70.4]])
])
def test_add_constraint(self):
init = self.sk_pca.hyperparam_schema()
init_expected = {
'allOf': [{
'type': 'object',
'properties': {
'n_components': {
'default': None
},
'copy': {
'default': True
},
'whiten': {
'default': False
},
'svd_solver': {
'default': 'auto'
},
'tol': {
'default': 0.0
},
'iterated_power': {
'default': 'auto'
},
'random_state': {
'default': None
}
}
}]
}
self.assertEqual(init, init_expected)
expected = {
'allOf': [
init_expected['allOf'][0], {
'anyOf': [{
'type': 'object',
'properties': {
'n_components': {
'not': {
'enum': ['mle']
},
}
},
}, {
'type': 'object',
'properties': {
'svd_solver': {
'enum': ['full', 'auto']
},
}
}]
}
]
}
foo = self.sk_pca.customize_schema(constraint=schemas.AnyOf([
schemas.Object(n_components=schemas.Not(schemas.Enum(['mle']))),
schemas.Object(svd_solver=schemas.Enum(['full', 'auto']))
]))
self.assertEqual(foo.hyperparam_schema(), expected)
helpers.validate_is_schema(foo._schemas)
self.assertEqual(self.sk_pca.hyperparam_schema(), init)
def test_classifier(self):
X_train, y_train = self.X_train, self.y_train
X_test, y_test = self.X_test, self.y_test
import importlib
module_name = ".".join(clf_name.split('.')[0:-1])
class_name = clf_name.split('.')[-1]
module = importlib.import_module(module_name)
class_ = getattr(module, class_name)
clf = class_()
#test_schemas_are_schemas
from lale.helpers import validate_is_schema
validate_is_schema(clf.input_schema_fit())
validate_is_schema(clf.input_schema_predict())
validate_is_schema(clf.output_schema())
validate_is_schema(clf.hyperparam_schema())
#test_init_fit_predict
trained = clf.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
#test_with_hyperopt
from lale.lib.lale import HyperoptClassifier
hyperopt = HyperoptClassifier(model=clf, max_evals=1)
trained = hyperopt.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
#test_cross_validation
from lale.helpers import cross_val_score
cv_results = cross_val_score(clf, X_train, y_train, cv=2)
self.assertEqual(len(cv_results), 2)
#test_with_gridsearchcv_auto_wrapped
from sklearn.metrics import accuracy_score, make_scorer
with warnings.catch_warnings():
warnings.simplefilter("ignore")
from lale.lib.sklearn.gradient_boosting_classifier import GradientBoostingClassifierImpl
from lale.lib.sklearn.mlp_classifier import MLPClassifierImpl
if isinstance(clf._impl, GradientBoostingClassifierImpl):
#because exponential loss does not work with iris dataset as it is not binary classification
import lale.schemas as schemas
clf = clf.customize_schema(
loss=schemas.Enum(default='deviance', values=['deviance']))
if not isinstance(clf._impl, MLPClassifierImpl):
#mlp fails due to issue #164.
grid_search = LaleGridSearchCV(
clf,
lale_num_samples=1,
lale_num_grids=1,
cv=2,
scoring=make_scorer(accuracy_score))
grid_search.fit(X_train, y_train)
#test_predict_on_trainable
trained = clf.fit(X_train, y_train)
clf.predict(X_train)
#test_to_json
clf.to_json()
#test_in_a_pipeline
pipeline = NoOp() >> clf
trained = pipeline.fit(self.X_train, self.y_train)
predictions = trained.predict(self.X_test)
}
}]
}
_combined_schemas = {
'$schema':
'http://json-schema.org/draft-04/schema#',
'description':
'Combined schema for expected data and hyperparameters for a transformer for'
' a text data transformer based on pre-trained BERT model '
'(https://github.com/huggingface/pytorch-pretrained-BERT).',
'type':
'object',
'tags': {
'pre': [],
'op': [],
'post': []
},
'properties': {
'input_fit': _input_schema_fit,
'input_predict': _input_schema_predict,
'output': _output_schema,
'hyperparams': _hyperparams_schema
}
}
if __name__ == "__main__":
helpers.validate_is_schema(_combined_schemas)
Batching = make_operator(BatchingImpl, _combined_schemas)