def setUp(self):
warnings.filterwarnings("ignore", ".*", ConvergenceWarning)
self.widget = self.create_widget(OWSGD,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
# Loss params for classification
ParameterMapping("loss",
self.widget.cls_loss_function_combo,
list(zip(*self.widget.cls_losses))[1],
problem_type="classification"),
ParameterMapping("epsilon",
self.widget.cls_epsilon_spin,
problem_type="classification"),
# Loss params for regression
ParameterMapping("loss",
self.widget.reg_loss_function_combo,
list(zip(*self.widget.reg_losses))[1],
problem_type="regression"),
ParameterMapping("epsilon",
self.widget.reg_epsilon_spin,
problem_type="regression"),
# Shared params
ParameterMapping("penalty", self.widget.penalty_combo,
list(zip(*self.widget.penalties))[1]),
ParameterMapping.from_attribute(self.widget, "alpha"),
ParameterMapping.from_attribute(self.widget, "l1_ratio"),
ParameterMapping("learning_rate", self.widget.learning_rate_combo,
list(zip(*self.widget.learning_rates))[1]),
ParameterMapping.from_attribute(self.widget, "eta0"),
ParameterMapping.from_attribute(self.widget, "power_t"),
]
def setUp(self):
self.widget = self.create_widget(OWPLS,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
ParameterMapping('max_iter', self.widget.n_iters),
ParameterMapping('n_components', self.widget.ncomps_spin)]
def setUp(self):
warnings.filterwarnings("ignore", ".*", ConvergenceWarning)
self.widget = self.create_widget(
OWSGD, stored_settings={"auto_apply": False})
self.init()
self.parameters = [
# Loss params for classification
ParameterMapping("loss", self.widget.cls_loss_function_combo,
list(zip(*self.widget.cls_losses))[1],
problem_type="classification"),
ParameterMapping("epsilon", self.widget.cls_epsilon_spin,
problem_type="classification"),
# Loss params for regression
ParameterMapping("loss", self.widget.reg_loss_function_combo,
list(zip(*self.widget.reg_losses))[1],
problem_type="regression"),
ParameterMapping("epsilon", self.widget.reg_epsilon_spin,
problem_type="regression"),
# Shared params
ParameterMapping("penalty", self.widget.penalty_combo,
list(zip(*self.widget.penalties))[1]),
ParameterMapping.from_attribute(self.widget, "alpha"),
ParameterMapping.from_attribute(self.widget, "l1_ratio"),
ParameterMapping("learning_rate", self.widget.learning_rate_combo,
list(zip(*self.widget.learning_rates))[1]),
ParameterMapping.from_attribute(self.widget, "eta0"),
ParameterMapping.from_attribute(self.widget, "power_t"),
]
def setUp(self):
self.widget = self.create_widget(OWAdaBoostClassification,
stored_settings={"auto_apply": False})
self.init()
losses = self.widget.losses
self.parameters = [
ParameterMapping('algorithm', self.widget.algorithm_combo, losses),
ParameterMapping('learning_rate', self.widget.learning_rate_spin),
ParameterMapping('n_estimators', self.widget.n_estimators_spin)]
def setUp(self):
self.widget = self.create_widget(OWAdaBoostRegression,
stored_settings={"auto_apply": False})
self.init()
losses = [loss.lower() for loss in self.widget.losses]
self.parameters = [
ParameterMapping('loss', self.widget.loss_combo, losses),
ParameterMapping('learning_rate', self.widget.learning_rate_spin),
ParameterMapping('n_estimators', self.widget.n_estimators_spin)]
def setUp(self):
self.widget = self.create_widget(OWKNNRegression,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
ParameterMapping('metric', self.widget.metrics_combo,
self.widget.metrics),
ParameterMapping('weights', self.widget.weights_combo,
self.widget.weights),
ParameterMapping('n_neighbors', self.widget.n_neighbors_spin)]
def setUp(self):
self.widget = self.create_widget(OWRandomForest,
stored_settings={"auto_apply": False})
self.init()
nest_spin = self.widget.n_estimators_spin
nest_min_max = [nest_spin.minimum() * 10, nest_spin.minimum()]
self.parameters = [
ParameterMapping("n_estimators", nest_spin, nest_min_max),
ParameterMapping("min_samples_split",
self.widget.min_samples_split_spin[1]),
]
def setUp(self):
self.widget = self.create_widget(OWClassificationTree,
stored_settings={"auto_apply": False})
self.init()
scores = [score[1] for score in self.widget.scores]
self.parameters = [
ParameterMapping('criterion', self.widget.score_combo, scores),
ParameterMapping('max_depth', self.widget.max_depth_spin[1]),
ParameterMapping('min_samples_split',
self.widget.min_internal_spin[1]),
ParameterMapping('min_samples_leaf', self.widget.min_leaf_spin[1])]
def test_parameters_svr_type(self):
"""Check learner and model for various values of all parameters
when NuSVR is chosen
"""
self.assertEqual(self.widget.svrtype, OWSVMRegression.Epsilon_SVR)
# setChecked(True) does not trigger callback event
self.widget.nu_radio.click()
self.assertEqual(self.widget.svrtype, OWSVMRegression.Nu_SVR)
self.parameters[0] = ParameterMapping("C", self.widget.nu_C_spin)
self.parameters[1] = ParameterMapping("nu", self.widget.nu_spin)
self.test_parameters()
def setUp(self):
self.widget = self.create_widget(OWKNNLearner,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
ParameterMapping("metric", self.widget.metrics_combo,
self.widget.metrics),
ParameterMapping("weights", self.widget.weights_combo,
self.widget.weights),
ParameterMapping("n_neighbors", self.widget.n_neighbors_spin),
]
def test_parameters_checked(self):
"""Check learner and model for various values of all parameters
when all properties are checked
"""
self.widget.max_features_spin[0].setCheckState(True)
self.widget.max_depth_spin[0].setCheckState(True)
self.parameters.extend([
ParameterMapping("max_features", self.widget.max_features_spin[1]),
ParameterMapping("max_depth", self.widget.max_depth_spin[1])
])
self.test_parameters()
def test_settings(self):
"""Check neighbors for various distance metrics"""
settings = [ParameterMapping("", self.widget.distance_combo,
self.widget.DISTANCES),
ParameterMapping("", self.widget.nn_spin)]
for setting in settings:
for val in setting.values:
self.send_signal("Data", self.iris)
self.send_signal("Reference", self.iris[:10])
setting.set_value(val)
self.widget.apply_button.button.click()
self.assertIsNotNone(self.get_output("Neighbors"))
def setUp(self):
self.widget = self.create_widget(OWGradientBoosting,
stored_settings={"auto_apply": False})
self.init()
controls = self.widget.editor.controls
self.parameters = [
ParameterMapping("n_estimators", controls.n_estimators, [500, 10]),
ParameterMapping("learning_rate", controls.learning_rate),
ParameterMapping("max_depth", controls.max_depth),
ParameterMapping("min_samples_split", controls.min_samples_split),
]
def test_settings(self):
"""Check neighbors for various distance metrics"""
widget = self.widget
settings = [
ParameterMapping("", widget.controls.distance_index, METRICS),
ParameterMapping("", widget.controls.n_neighbors)]
for setting in settings:
for val in setting.values:
self.send_signal(widget.Inputs.data, self.iris)
self.send_signal(widget.Inputs.reference, self.iris[:10])
setting.set_value(val)
widget.apply_button.button.click()
if METRICS[widget.distance_index][0] != "Jaccard" \
and widget.n_neighbors != 0:
self.assertIsNotNone(self.get_output("Neighbors"))
def setUp(self):
self.widget = self.create_widget(
OWTreeLearner, stored_settings={"auto_apply": False})
self.init()
self.model_class = Model
self.parameters = [
ParameterMapping.from_attribute(self.widget, 'max_depth'),
ParameterMapping.from_attribute(
self.widget, 'min_internal', 'min_samples_split'),
ParameterMapping.from_attribute(
self.widget, 'min_leaf', 'min_samples_leaf')]
# NB. sufficient_majority is divided by 100, so it cannot be tested like
# this
self.checks = [sb.gui_element.cbox for sb in self.parameters]
def setUp(self):
self.widget = self.create_widget(
OWTreeLearner, stored_settings={"auto_apply": False})
self.init()
self.model_class = Model
self.parameters = [
ParameterMapping.from_attribute(self.widget, 'max_depth'),
ParameterMapping.from_attribute(
self.widget, 'min_internal', 'min_samples_split'),
ParameterMapping.from_attribute(
self.widget, 'min_leaf', 'min_samples_leaf')]
# NB. sufficient_majority is divided by 100, so it cannot be tested like
# this
self.checks = [sb.gui_element.cbox for sb in self.parameters]
def setUp(self):
self.widget = self.create_widget(OWAdaBoost,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
# TODO Due to the way params are tested on the learner and the fact
# that learners only receive a subset of these parameters, this
# method of testing these parameters is not viable
# ParameterMapping('algorithm', self.widget.cls_algorithm_combo,
# self.widget.algorithms),
# ParameterMapping('loss', self.widget.reg_algorithm_combo,
# self.widget.losses),
ParameterMapping('learning_rate', self.widget.learning_rate_spin),
ParameterMapping('n_estimators', self.widget.n_estimators_spin),
ParameterMapping.from_attribute(self.widget, 'random_seed',
'random_state')
]
def setUp(self):
self.widget = self.create_widget(OWAdaBoost,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
ParameterMapping('algorithm',
self.widget.cls_algorithm_combo,
self.widget.algorithms,
problem_type="classification"),
ParameterMapping('loss',
self.widget.reg_algorithm_combo,
[x.lower() for x in self.widget.losses],
problem_type="regression"),
ParameterMapping('learning_rate', self.widget.learning_rate_spin),
ParameterMapping('n_estimators', self.widget.n_estimators_spin),
ParameterMapping.from_attribute(self.widget, 'random_seed',
'random_state')
]
def setUp(self):
self.widget = self.create_widget(OWRuleLearner,
stored_settings={"auto_apply": False})
self.init()
self.radio_button_groups = self.widget.findChildren(QButtonGroup)
self.radio_buttons = self.widget.findChildren(QRadioButton)
self.spin_boxes = self.widget.findChildren(QSpinBox)
self.double_spin_boxes = self.widget.findChildren(QDoubleSpinBox)
self.combo_boxes = self.widget.findChildren(QComboBox)
self.parameters = [
ParameterMapping("Evaluation measure", self.combo_boxes[0],
self.widget.storage_measures),
ParameterMapping("Beam width", self.spin_boxes[0]),
ParameterMapping("Minimum rule coverage", self.spin_boxes[1]),
ParameterMapping("Maximum rule length", self.spin_boxes[2]),
]
def test_xgb_params(self):
simulate.combobox_activate_index(self.widget.controls.method_index, 1)
editor = self.widget.editor
controls = editor.controls
reg_slider = controls.lambda_index
self.parameters = [
ParameterMapping("n_estimators", controls.n_estimators, [500, 10]),
ParameterMapping("learning_rate", controls.learning_rate),
ParameterMapping("max_depth", controls.max_depth),
ParameterMapping("reg_lambda",
reg_slider,
values=[editor.LAMBDAS[0], editor.LAMBDAS[-1]],
getter=lambda: editor.LAMBDAS[reg_slider.value()],
setter=lambda val: reg_slider.setValue(
editor.LAMBDAS.index(val))),
]
self.test_parameters()
def setUp(self):
self.widget = self.create_widget(OWLogisticRegression,
stored_settings={"auto_apply": False})
self.init()
c_slider = self.widget.c_slider
def setter(val):
index = self.widget.C_s.index(val)
self.widget.C_s[c_slider.value()]
c_slider.setValue(index)
self.parameters = [
ParameterMapping('penalty', self.widget.penalty_combo,
self.widget.penalty_types_short),
ParameterMapping('C',
c_slider,
values=[self.widget.C_s[0], self.widget.C_s[-1]],
getter=lambda: self.widget.C_s[c_slider.value()],
setter=setter)
]
def setUp(self):
self.widget = self.create_widget(
OWSVM, stored_settings={"auto_apply": False})
self.init()
# All this nonsense is necessary because we add an `auto` option to the
# gamma spin box
gamma_spin = self.widget._kernel_params[0]
values = [self.widget._default_gamma, gamma_spin.maximum()]
def getter():
value = gamma_spin.value()
return gamma_spin.specialValueText() \
if value == gamma_spin.minimum() else value
def setter(value):
if value == gamma_spin.specialValueText():
gamma_spin.setValue(gamma_spin.minimum())
else:
gamma_spin.setValue(value)
self.parameters = [
ParameterMapping("C", self.widget.c_spin),
ParameterMapping("gamma", self.widget._kernel_params[0],
values=values, setter=setter, getter=getter),
ParameterMapping("coef0", self.widget._kernel_params[1]),
ParameterMapping("degree", self.widget._kernel_params[2]),
ParameterMapping("tol", self.widget.tol_spin),
ParameterMapping("max_iter", self.widget.max_iter_spin[1])]
def setUp(self):
self.widget = self.create_widget(OWSVMRegression,
stored_settings={"auto_apply": False})
self.init()
gamma_spin = self.widget._kernel_params[0]
values = [self.widget._default_gamma, gamma_spin.maximum()]
def getter():
value = gamma_spin.value()
return gamma_spin.specialValueText() \
if value == gamma_spin.minimum() else value
def setter(value):
if value == gamma_spin.specialValueText():
gamma_spin.setValue(gamma_spin.minimum())
else:
gamma_spin.setValue(value)
self.parameters = [
ParameterMapping("C", self.widget.epsilon_C_spin),
ParameterMapping("epsilon", self.widget.epsilon_spin),
ParameterMapping("gamma",
self.widget._kernel_params[0],
values=values,
setter=setter,
getter=getter),
ParameterMapping("coef0", self.widget._kernel_params[1]),
ParameterMapping("degree", self.widget._kernel_params[2]),
ParameterMapping("tol", self.widget.tol_spin)
]
def setUp(self):
self.widget = self.create_widget(
OWAdaBoost, stored_settings={"auto_apply": False})
self.init()
self.parameters = [
ParameterMapping('algorithm', self.widget.cls_algorithm_combo,
self.widget.algorithms,
problem_type="classification"),
ParameterMapping('loss', self.widget.reg_algorithm_combo,
[x.lower() for x in self.widget.losses],
problem_type="regression"),
ParameterMapping('learning_rate', self.widget.learning_rate_spin),
ParameterMapping('n_estimators', self.widget.n_estimators_spin),
ParameterMapping.from_attribute(
self.widget, 'random_seed', 'random_state')]
def setUp(self):
self.widget = self.create_widget(OWSGD,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
ParameterMapping('loss', self.widget.cls_loss_function_combo,
list(zip(*self.widget.cls_losses))[1]),
ParameterMapping.from_attribute(self.widget, 'cls_epsilon',
'epsilon'),
ParameterMapping('penalty', self.widget.penalty_combo,
list(zip(*self.widget.penalties))[1]),
ParameterMapping.from_attribute(self.widget, 'alpha'),
ParameterMapping.from_attribute(self.widget, 'l1_ratio'),
ParameterMapping('learning_rate', self.widget.learning_rate_combo,
list(zip(*self.widget.learning_rates))[1]),
ParameterMapping.from_attribute(self.widget, 'eta0'),
ParameterMapping.from_attribute(self.widget, 'power_t'),
]
def setUp(self):
self.widget = self.create_widget(OWSGD,
stored_settings={"auto_apply": False})
self.init()
self.parameters = [
# TODO Fix the base tests to support parameter delegation with
# different problem types
# ParameterMapping('loss', self.widget.loss_function_combo,
# list(zip(*self.widget.losses))[1]),
# ParameterMapping.from_attribute(self.widget, 'epsilon'),
ParameterMapping('penalty', self.widget.penalty_combo,
list(zip(*self.widget.penalties))[1]),
ParameterMapping.from_attribute(self.widget, 'alpha'),
ParameterMapping.from_attribute(self.widget, 'l1_ratio'),
ParameterMapping('learning_rate', self.widget.learning_rate_combo,
list(zip(*self.widget.learning_rates))[1]),
ParameterMapping.from_attribute(self.widget, 'eta0'),
ParameterMapping.from_attribute(self.widget, 'power_t'),
]