def test_categorical_features(self, lung_X, lung_y,
respect_categorical_features):
# add a categorical feature
categorical_col = np.atleast_2d(
np.array([random.choice([0, 1]) for _ in range(lung_X.shape[0])]))
lung_X_c = np.hstack((lung_X, categorical_col.transpose()))
categorical_features = [lung_X.shape[1]]
rfs = RangerForestSurvival(
respect_categorical_features=respect_categorical_features,
categorical_features=categorical_features,
)
if respect_categorical_features not in [
"partition", "ignore", "order"
]:
with pytest.raises(ValueError):
rfs.fit(lung_X_c, lung_y)
return
rfs.fit(lung_X_c, lung_y)
if respect_categorical_features in ("ignore", "order"):
assert rfs.categorical_features_ == []
else:
assert rfs.categorical_features_ == [
str(c).encode() for c in categorical_features
]
def test_importance(
self, lung_X, lung_y, importance, scale_permutation_importance, local_importance
):
forest = RangerForestSurvival(
importance=importance,
scale_permutation_importance=scale_permutation_importance,
local_importance=local_importance,
)
if importance not in ["none", "impurity", "impurity_corrected", "permutation"]:
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
return
forest.fit(lung_X, lung_y)
if importance == "none":
assert forest.importance_mode_ == 0
elif importance == "impurity":
assert forest.importance_mode_ == 1
elif importance == "impurity_corrected":
assert forest.importance_mode_ == 5
elif importance == "permutation":
if local_importance:
assert forest.importance_mode_ == 6
elif scale_permutation_importance:
assert forest.importance_mode_ == 2
else:
assert forest.importance_mode_ == 3
def test_sample_fraction_replace(self, lung_X, lung_y, replace):
forest = RangerForestSurvival(replace=replace)
forest.fit(lung_X, lung_y)
if replace:
assert forest.sample_fraction_ == [1.0]
else:
assert forest.sample_fraction_ == [0.632]
def test_verbose(self, lung_X, lung_y, verbose, capfd):
forest = RangerForestSurvival(verbose=verbose)
forest.fit(lung_X, lung_y)
captured = capfd.readouterr()
if verbose:
assert len(captured.out) > 0
else:
assert len(captured.out) == 0
def test_sample_fraction(self, lung_X, lung_y):
forest = RangerForestSurvival(sample_fraction=0.69)
forest.fit(lung_X, lung_y)
assert forest.sample_fraction_ == [0.69]
# test with single record
lung_X_record = lung_X.values[0:1, :]
pred = forest.predict(lung_X_record)
assert len(pred) == 1
def test_serialize(self, lung_X, lung_y):
tf = tempfile.TemporaryFile()
forest = RangerForestSurvival(n_estimators=N_ESTIMATORS)
forest.fit(lung_X, lung_y)
pickle.dump(forest, tf)
tf.seek(0)
new_forest = pickle.load(tf)
pred = new_forest.predict(lung_X)
assert len(pred) == lung_X.shape[0]
def test_fit(self, lung_X, lung_y):
forest = RangerForestSurvival(n_estimators=N_ESTIMATORS)
with pytest.raises(NotFittedError):
check_is_fitted(forest)
forest.fit(lung_X, lung_y)
check_is_fitted(forest)
assert hasattr(forest, "event_times_")
assert hasattr(forest, "cumulative_hazard_function_")
assert hasattr(forest, "ranger_forest_")
assert hasattr(forest, "n_features_in_")
def test_sample_weight(self, lung_X, lung_y):
forest_w = RangerForestSurvival()
forest_w.fit(lung_X, lung_y, sample_weight=[1] * len(lung_y))
forest = RangerForestSurvival()
forest.fit(lung_X, lung_y)
pred_w = forest_w.predict(lung_X)
pred = forest.predict(lung_X)
np.testing.assert_array_equal(pred.reshape(-1, 1), pred_w.reshape(-1, 1))
def test_predict(self, lung_X, lung_y):
forest = RangerForestSurvival(n_estimators=N_ESTIMATORS)
forest.fit(lung_X, lung_y)
pred = forest.predict(lung_X)
assert len(pred) == lung_X.shape[0]
# test with single record
lung_X_record = lung_X.values[0:1, :]
pred = forest.predict(lung_X_record)
assert len(pred) == 1
def test_estimators_(self, lung_X, lung_y):
forest = RangerForestSurvival(n_estimators=10)
with pytest.raises(AttributeError):
_ = forest.estimators_
forest.fit(lung_X, lung_y)
with pytest.raises(ValueError):
_ = forest.estimators_
forest = RangerForestSurvival(n_estimators=10, enable_tree_details=True)
forest.fit(lung_X, lung_y)
estimators = forest.estimators_
assert len(estimators) == 10
assert isinstance(estimators[0], RangerTreeSurvival)
check_is_fitted(estimators[0])
def test_regularization(self, lung_X, lung_y):
forest = RangerForestSurvival()
forest.fit(lung_X, lung_y)
assert forest.regularization_factor_ == []
assert not forest.use_regularization_factor_
# vector must be between 0 and 1 and length matching feature num
for r in [[1.1], [-0.1], [1, 1]]:
forest = RangerForestSurvival(regularization_factor=r)
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
# vector of ones isn't applied
forest = RangerForestSurvival(regularization_factor=[1] * lung_X.shape[1])
forest.fit(lung_X, lung_y)
assert forest.regularization_factor_ == []
assert not forest.use_regularization_factor_
# regularization vector is used
reg = [0.5]
forest = RangerForestSurvival(regularization_factor=reg, n_jobs=2)
# warns if n_jobs is not one since parallelization can't be used
with pytest.warns(Warning):
forest.fit(lung_X, lung_y)
assert forest.n_jobs_ == 1
assert forest.regularization_factor_ == reg
assert forest.use_regularization_factor_
def test_get_estimator(self, lung_X, lung_y):
forest = RangerForestSurvival(n_estimators=10)
with pytest.raises(NotFittedError):
_ = forest.get_estimator(idx=0)
forest.fit(lung_X, lung_y)
with pytest.raises(ValueError):
_ = forest.get_estimator(0)
forest = RangerForestSurvival(n_estimators=10, enable_tree_details=True)
forest.fit(lung_X, lung_y)
estimator = forest.get_estimator(0)
estimator.predict(lung_X)
assert isinstance(estimator, RangerTreeSurvival)
with pytest.raises(IndexError):
_ = forest.get_estimator(idx=20)
def test_importance_pvalues(self, lung_X_mod, lung_y, importance, mod):
rfs = RangerForestSurvival(importance=importance)
np.random.seed(42)
if importance not in [
"none", "impurity", "impurity_corrected", "permutation"
]:
with pytest.raises(ValueError):
rfs.fit(lung_X_mod, lung_y)
return
if not importance == "impurity_corrected":
rfs.fit(lung_X_mod, lung_y)
with pytest.raises(ValueError):
rfs.get_importance_pvalues()
return
# Test error for no non-negative importance values
if mod == "none":
rfs.fit(lung_X_mod, lung_y)
with pytest.raises(ValueError):
rfs.get_importance_pvalues()
return
rfs.fit(lung_X_mod, lung_y)
assert len(rfs.get_importance_pvalues()) == lung_X_mod.shape[1]
def test_feature_importances_(self, lung_X, lung_y, importance, local_importance):
forest = RangerForestSurvival(
importance=importance, local_importance=local_importance
)
with pytest.raises(AttributeError):
_ = forest.feature_importances_
if importance == "INVALID":
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
return
forest.fit(lung_X, lung_y)
if importance == "none":
with pytest.raises(ValueError):
_ = forest.feature_importances_
else:
assert len(forest.feature_importances_) == lung_X.shape[1]
def test_categorical_features(self, lung_X, lung_y, respect_categorical_features):
# add a categorical feature
categorical_col = np.atleast_2d(
np.array([random.choice([0, 1]) for _ in range(lung_X.shape[0])])
)
lung_X_c = np.hstack((lung_X, categorical_col.transpose()))
categorical_features = [lung_X.shape[1]]
forest = RangerForestSurvival(
respect_categorical_features=respect_categorical_features,
)
if respect_categorical_features not in ["partition", "ignore", "order"]:
with pytest.raises(ValueError):
forest.fit(lung_X_c, lung_y, categorical_features=categorical_features)
return
forest.fit(lung_X_c, lung_y, categorical_features=categorical_features)
forest.predict(lung_X_c)
def test_split_rule(self, lung_X, lung_y, split_rule):
forest = RangerForestSurvival(split_rule=split_rule)
assert forest.criterion == split_rule
if split_rule not in ["logrank", "extratrees", "C", "C_ignore_ties", "maxstat"]:
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
return
forest.fit(lung_X, lung_y)
if split_rule == "logrank":
assert forest.split_rule_ == 1
elif split_rule == "extratrees":
assert forest.split_rule_ == 5
elif split_rule == "C":
assert forest.split_rule_ == 2
elif split_rule == "C_ignore_ties":
assert forest.split_rule_ == 3
elif split_rule == "maxstat":
assert forest.split_rule_ == 4
if split_rule != "extratrees":
forest = RangerForestSurvival(split_rule=split_rule, num_random_splits=2)
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
def test_inbag(self, lung_X, lung_y):
inbag = [[1, 2, 3], [2, 3, 4]]
forest = RangerForestSurvival(n_estimators=2, inbag=inbag)
forest.fit(lung_X, lung_y)
# inbag list different length from n_estimators
forest = RangerForestSurvival(n_estimators=1, inbag=inbag)
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
# can't use inbag with sample weight
forest = RangerForestSurvival(inbag=inbag)
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y, sample_weight=[1] * len(lung_y))
# can't use class sampling and inbag
forest = RangerForestSurvival(inbag=inbag, sample_fraction=[1, 1])
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
def test_split_select_weights(self, lung_X, lung_y):
n_trees = 10
weights = [0.1] * lung_X.shape[1]
forest = RangerForestSurvival(n_estimators=n_trees,)
forest.fit(lung_X, lung_y, split_select_weights=weights)
weights = [0.1] * (lung_X.shape[1] - 1)
forest = RangerForestSurvival(n_estimators=n_trees)
with pytest.raises(RuntimeError):
forest.fit(lung_X, lung_y, split_select_weights=weights)
weights = [[0.1] * (lung_X.shape[1])] * n_trees
forest = RangerForestSurvival(n_estimators=n_trees)
forest.fit(lung_X, lung_y, split_select_weights=weights)
weights = [[0.1] * (lung_X.shape[1])] * (n_trees + 1)
forest = RangerForestSurvival(n_estimators=n_trees)
with pytest.raises(RuntimeError):
forest.fit(lung_X, lung_y, split_select_weights=weights)
def test_mtry(self, lung_X, lung_y, mtry):
forest = RangerForestSurvival(mtry=mtry)
if callable(mtry) and mtry(5) > 5:
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
return
elif not callable(mtry) and (mtry < 0 or mtry > lung_X.shape[0]):
with pytest.raises(ValueError):
forest.fit(lung_X, lung_y)
return
forest.fit(lung_X, lung_y)
if callable(mtry):
assert forest.mtry_ == mtry(lung_X.shape[1])
else:
assert forest.mtry_ == mtry
def test_predict_cumulative_hazard_function(self, lung_X, lung_y):
rfs = RangerForestSurvival(n_estimators=N_ESTIMATORS)
rfs.fit(lung_X, lung_y)
pred = rfs.predict_cumulative_hazard_function(lung_X)
assert len(pred) == lung_X.shape[0]
def test_clone(self, lung_X, lung_y):
forest = RangerForestSurvival(n_estimators=N_ESTIMATORS)
forest.fit(lung_X, lung_y)
clone(forest)
def test_clone(self, lung_X, lung_y):
rfs = RangerForestSurvival(n_estimators=N_ESTIMATORS)
rfs.fit(lung_X, lung_y)
clone(rfs)
def test_predict_survival_function(self, lung_X, lung_y):
forest = RangerForestSurvival(n_estimators=N_ESTIMATORS)
forest.fit(lung_X, lung_y)
pred = forest.predict_survival_function(lung_X)
assert len(pred) == lung_X.shape[0]
def test_always_split_features(self, lung_X, lung_y):
forest = RangerForestSurvival()
forest.fit(lung_X, lung_y, always_split_features=[0])
# feature 0 is in every tree split
for tree in forest.ranger_forest_["forest"]["split_var_ids"]:
assert 0 in tree