def test_no_attributes_set_in_init(name, Estimator):
# input validation etc for non-meta estimators
with ignore_warnings(category=(DeprecationWarning, ConvergenceWarning,
UserWarning, FutureWarning)):
estimator = Estimator()
# check this on class
check_no_attributes_set_in_init(name, estimator)
def test_no_attributes_set_init_sublcassed():
"""Tests that subclassed models can be made that
set all parameters in a single __init__
"""
estimator = SubclassedClassifier()
check_no_attributes_set_in_init(estimator.__name__, estimator)
basic_checks(estimator, load_iris)
def custom_check_estimator(Estimator):
# Same as sklearn.check_estimator, skipping tests that can't succeed.
if isinstance(Estimator, type):
# got a class
name = Estimator.__name__
estimator = Estimator()
check_parameters_default_constructible(name, Estimator)
check_no_attributes_set_in_init(name, estimator)
else:
# got an instance
estimator = Estimator
name = type(estimator).__name__
for check in _yield_all_checks(name, estimator):
if (check is estimator_checks.check_estimators_dtypes
or check is estimator_checks.check_fit_score_takes_y
or check is estimator_checks.check_dtype_object
or check is estimator_checks.check_sample_weights_list
or check is estimator_checks.check_estimators_overwrite_params
or check is estimator_checks.check_classifiers_classes
or check is estimator_checks.check_supervised_y_2d
or check is estimator_checks.check_fit2d_predict1d
or check is estimator_checks.check_class_weight_classifiers
or check is estimator_checks.check_methods_subset_invariance
or check is estimator_checks.check_dont_overwrite_parameters
or "check_estimators_fit_returns_self" in check.__repr__()
or "check_classifiers_train" in check.__repr__()):
continue
try:
check(name, estimator)
except SkipTest as exception:
# the only SkipTest thrown currently results from not
# being able to import pandas.
warnings.warn(str(exception), SkipTestWarning)
def test_check_no_attributes_set_in_init():
class NonConformantEstimatorPrivateSet(BaseEstimator):
def __init__(self):
self.you_should_not_set_this_ = None
class NonConformantEstimatorNoParamSet(BaseEstimator):
def __init__(self, you_should_set_this_=None):
pass
msg = (
"Estimator estimator_name should not set any"
" attribute apart from parameters during init."
r" Found attributes \['you_should_not_set_this_'\]."
)
with raises(AssertionError, match=msg):
check_no_attributes_set_in_init(
"estimator_name", NonConformantEstimatorPrivateSet()
)
msg = (
"Estimator estimator_name should store all parameters as an attribute"
" during init"
)
with raises(AttributeError, match=msg):
check_no_attributes_set_in_init(
"estimator_name", NonConformantEstimatorNoParamSet()
)
def test_no_attributes_set_in_init(name, Estimator):
# input validation etc for non-meta estimators
with ignore_warnings(category=(DeprecationWarning, ConvergenceWarning,
UserWarning, FutureWarning)):
estimator = Estimator()
# check this on class
check_no_attributes_set_in_init(name, estimator)
def check_no_attributes_set_in_init(estimator):
from sklearn.utils.estimator_checks import check_no_attributes_set_in_init
check_no_attributes_set_in_init("test", estimator)
return True
def test_no_attributes_set_in_init(name, estimator):
# input validation etc for all estimators
with ignore_warnings(category=(DeprecationWarning, ConvergenceWarning,
UserWarning, FutureWarning)):
tags = _safe_tags(estimator)
if tags['_skip_test']:
warnings.warn(
"Explicit SKIP via _skip_test tag for "
"{}.".format(name), SkipTestWarning)
return
# check this on class
check_no_attributes_set_in_init(name, estimator)
def check_estimator_autofeat(Estimator):
# usually, this would be
# from sklearn.utils.estimator_checks import check_estimator
# but first this issue needs to be resolved:
# https://github.com/pandas-dev/pandas/issues/26247
# check class
name = Estimator.__name__
estimator = Estimator()
check_parameters_default_constructible(name, Estimator)
check_no_attributes_set_in_init(name, estimator)
# check with fewer feateng steps and featsel runs to speed things up
check_estimator(
Estimator(feateng_steps=1, featsel_runs=1, featsel_max_it=20))
def test_no_attributes_set_init_no_args():
"""Tests that models with no build arguments
set all parameters in a single __init__
"""
def build_fn():
model = Sequential()
model.add(layers.Dense(1, input_dim=1, activation="relu"))
model.add(layers.Dense(1))
model.compile(loss="mse")
return model
estimator = KerasRegressor(model=build_fn)
check_no_attributes_set_in_init(estimator.__name__, estimator)
estimator.fit([[1]], [1])
def check_estimator(Estimator):
"""Check if estimator adheres to scikit-learn conventions.
This estimator will run an extensive test-suite for input validation,
shapes, etc.
Additional tests for classifiers, regressors, clustering or transformers
will be run if the Estimator class inherits from the corresponding mixin
from sklearn.base.
This test can be applied to classes or instances.
Classes currently have some additional tests that related to construction,
while passing instances allows the testing of multiple options.
Parameters
----------
estimator : estimator object or class
Estimator to check. Estimator is a class object or instance.
"""
if isinstance(Estimator, type):
# got a class
name = Estimator.__name__
estimator = Estimator()
check_parameters_default_constructible(name, Estimator)
check_no_attributes_set_in_init(name, estimator)
else:
# got an instance
estimator = Estimator
name = type(estimator).__name__
if hasattr(estimator, 'max_iter'):
if (isinstance(estimator, ShapeletModel)
or isinstance(estimator, SerializableShapeletModel)):
estimator.set_params(max_iter=100)
else:
estimator.set_params(max_iter=10)
if hasattr(estimator, 'total_lengths'):
estimator.set_params(total_lengths=1)
if hasattr(estimator, 'probability'):
estimator.set_params(probability=True)
for check in checks._yield_all_checks(name, estimator):
try:
check(name, estimator)
except SkipTest as exception:
# the only SkipTest thrown currently results from not
# being able to import pandas.
warnings.warn(str(exception), SkipTestWarning)
def rewritten_check_estimator(Estimator):
# Same as sklearn.check_estimator, re-writing tests that can't succeed.
if isinstance(Estimator, type):
# got a class
name = Estimator.__name__
estimator = Estimator()
check_parameters_default_constructible(name, Estimator)
check_no_attributes_set_in_init(name, estimator)
else:
# got an instance
estimator = Estimator
name = type(estimator).__name__
for check in _yield_rewritten_checks(name, estimator):
try:
check(name, estimator)
except SkipTest as exception:
# the only SkipTest thrown currently results from not
# being able to import pandas.
warnings.warn(str(exception), SkipTestWarning)
def check_estimator(Estimator):
"""Check if estimator adheres to scikit-learn conventions.
This estimator will run an extensive test-suite for input validation,
shapes, etc.
Additional tests for classifiers, regressors, clustering or transformers
will be run if the Estimator class inherits from the corresponding mixin
from sklearn.base.
This test can be applied to classes or instances.
Classes currently have some additional tests that related to construction,
while passing instances allows the testing of multiple options.
Parameters
----------
estimator : estimator object or class
Estimator to check. Estimator is a class object or instance.
"""
warnings.filterwarnings(action='ignore', category=DataConversionWarning)
if isinstance(Estimator, type):
# got a class
name = Estimator.__name__
estimator = Estimator()
check_parameters_default_constructible(name, Estimator)
check_no_attributes_set_in_init(name, estimator)
else:
# got an instance
estimator = Estimator
name = type(estimator).__name__
for check in _yield_all_checks(name, estimator):
if hasattr(check, "__name__"):
if check.__name__ == "check_estimators_nan_inf":
# we allow NaNs in transform...
continue
print("##", check.__name__)
try:
check(name, estimator)
except SkipTest as exception:
# the only SkipTest thrown currently results from not
# being able to import pandas.
warnings.warn(str(exception), SkipTestWarning)
def test_check_estimators_voting_estimator(estimator):
# FIXME: to be removed when meta-estimators can be specified themselves
# their testing parameters (for required parameters).
check_estimator(estimator)
check_no_attributes_set_in_init(estimator.__class__.__name__, estimator)
def test_no_attributes_set_in_init(name, Estimator):
# input validation etc for non-meta estimators
estimator = Estimator()
# check this on class
check_no_attributes_set_in_init(name, estimator)
def test_check_estimators_stacking_estimator(estimator):
check_estimator(estimator)
check_no_attributes_set_in_init(estimator.__class__.__name__, estimator)
def test_no_attributes_set_in_init(name, Estimator):
# input validation etc for non-meta estimators
estimator = Estimator()
# check this on class
check_no_attributes_set_in_init(name, estimator)
