def test_fit_with_unused_fit_params(tmp_path: pathlib.Path) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(n_estimators=n_estimators,
n_trials=n_trials,
model_dir=tmp_path)
clf.fit(X, y, eval_set=None)
def test_fit_with_invalid_study(tmp_path: pathlib.Path) -> None:
X, y = load_breast_cancer(return_X_y=True)
study = study_module.create_study(direction="maximize")
clf = OGBMClassifier(study=study, model_dir=tmp_path)
with pytest.raises(ValueError):
clf.fit(X, y)
def test_hasattr(tmp_path: pathlib.Path, refit: bool,
early_stopping_rounds: int) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(
n_estimators=n_estimators,
n_trials=n_trials,
refit=refit,
model_dir=tmp_path,
)
attrs = {
"classes_": np.ndarray,
"best_index_": int,
"best_iteration_": (int, type(None)),
"best_params_": dict,
"best_score_": float,
"booster_": (lgb.Booster, _VotingBooster),
"encoder_": LabelEncoder,
"feature_importances_": np.ndarray,
"n_classes_": int,
"n_features_": int,
"n_splits_": int,
"study_": study_module.Study,
}
for attr in attrs:
with pytest.raises(AttributeError):
getattr(clf, attr)
clf.fit(X, y, early_stopping_rounds=early_stopping_rounds)
for attr, klass in attrs.items():
assert isinstance(getattr(clf, attr), klass)
if refit:
assert hasattr(clf, "refit_time_")
assert clf.booster_.best_iteration == 0
else:
assert not hasattr(clf, "refit_time_")
boosters = clf.booster_.boosters
for b in boosters:
if early_stopping_rounds is None:
assert b.best_iteration == clf.n_estimators
else:
assert b.best_iteration == clf.best_iteration_
if early_stopping_rounds is None:
assert clf.best_iteration_ is None
else:
assert clf.best_iteration_ > 0
def test_fit_with_fit_params(
tmp_path: pathlib.Path,
callbacks: Optional[List[Callable]],
eval_metric: Union[Callable, str],
) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(n_estimators=n_estimators,
n_trials=n_trials,
model_dir=tmp_path)
clf.fit(X, y, callbacks=callbacks, eval_metric=eval_metric)
def test_predict_with_unused_predict_params(tmp_path: pathlib.Path) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(n_estimators=n_estimators,
n_trials=n_trials,
model_dir=tmp_path)
clf.fit(X, y)
y_pred = clf.predict(X, pred_leaf=False)
assert isinstance(y_pred, np.ndarray)
assert y.shape == y_pred.shape
def test_plot_importance(tmp_path: pathlib.Path, n_jobs: int) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(
n_estimators=n_estimators,
n_jobs=n_jobs,
n_trials=n_trials,
refit=False,
model_dir=tmp_path,
)
clf.fit(X, y)
lgb.plot_importance(clf)
def test_refit(tmp_path: pathlib.Path,
early_stopping_rounds: Optional[int]) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(
n_estimators=n_estimators,
n_trials=n_trials,
random_state=random_state,
refit=True,
model_dir=tmp_path,
)
clf.fit(X, y, early_stopping_rounds=early_stopping_rounds)
y_pred = clf.predict(X)
if early_stopping_rounds is None:
_n_estimators = n_estimators
else:
_n_estimators = clf.best_iteration_
clf = lgb.LGBMClassifier(n_estimators=_n_estimators, **clf.best_params_)
clf.fit(X, y)
np.testing.assert_array_equal(y_pred, clf.predict(X))
def test_predict_with_predict_params(tmp_path: pathlib.Path,
num_iteration: Optional[int]) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(n_estimators=n_estimators,
n_trials=n_trials,
model_dir=tmp_path)
clf.fit(X, y)
y_pred = clf.predict(X, num_iteration=num_iteration)
assert isinstance(y_pred, np.ndarray)
assert y.shape == y_pred.shape
def test_fit_twice_without_study(tmp_path: pathlib.Path, n_jobs: int) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(
n_estimators=n_estimators,
n_jobs=n_jobs,
n_trials=n_trials,
random_state=random_state,
model_dir=tmp_path,
)
clf.fit(X, y)
df = clf.study_.trials_dataframe()
values = df["value"]
clf = OGBMClassifier(
bagging_fraction=1.0,
bagging_freq=0,
feature_fraction=1.0,
lambda_l1=0.0,
lambda_l2=0.0,
min_data_in_leaf=20,
n_estimators=n_estimators,
n_jobs=n_jobs,
n_trials=n_trials,
random_state=random_state,
model_dir=tmp_path,
)
clf.fit(X, y)
df = clf.study_.trials_dataframe()
np.testing.assert_array_equal(values, df["value"])
def test_fit_with_group_k_fold(tmp_path: pathlib.Path) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(
cv=GroupKFold(5),
n_estimators=n_estimators,
n_trials=n_trials,
model_dir=tmp_path,
)
n_samples, _ = X.shape
groups = np.random.choice(10, size=n_samples)
clf.fit(X, y, groups=groups)
def test_fit_with_pruning(tmp_path: pathlib.Path) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(enable_pruning=True, model_dir=tmp_path)
clf.fit(X, y)
if hasattr(clf.study_, "get_trials"):
trials = clf.study_.get_trials()
else:
trials = clf.study_.trials
pruned_trials = [t for t in trials if t.state == structs.TrialState.PRUNED]
assert len(pruned_trials) > 0
def test_score(tmp_path: pathlib.Path) -> None:
X, y = load_breast_cancer(return_X_y=True)
X_train, X_test, y_train, y_test = train_test_split(
X, y, random_state=random_state)
clf = lgb.LGBMClassifier(random_state=random_state)
clf.fit(X_train, y_train)
score = clf.score(X_test, y_test)
clf = OGBMClassifier(random_state=random_state, model_dir=tmp_path)
clf.fit(X_train, y_train)
assert score <= clf.score(X_test, y_test)
def test_fit_with_empty_param_distributions(tmp_path: pathlib.Path) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(
colsample_bytree=0.1,
n_estimators=n_estimators,
n_trials=n_trials,
param_distributions={},
model_dir=tmp_path,
)
clf.fit(X, y)
df = clf.study_.trials_dataframe()
values = df["value"]
assert values.nunique() == 1
def test_ogbm_classifier(tmp_path: pathlib.Path) -> None:
pytest.importorskip("sklearn", minversion="0.20.0")
# from sklearn.utils.estimator_checks import check_estimator
from sklearn.utils.estimator_checks import check_estimators_pickle
from sklearn.utils.estimator_checks import check_set_params
clf = OGBMClassifier(model_dir=tmp_path)
name = clf.__class__.__name__
# check_estimator(clf)
check_estimators_pickle(name, clf)
check_set_params(name, clf)
def test_fit_twice_with_study(tmp_path: pathlib.Path,
storage: Optional[str]) -> None:
X, y = load_breast_cancer(return_X_y=True)
study = study_module.create_study(storage=storage)
clf = OGBMClassifier(
n_estimators=n_estimators,
n_trials=n_trials,
study=study,
model_dir=tmp_path,
)
clf.fit(X, y)
assert len(study.trials) == n_trials
clf.fit(X, y)
assert len(study.trials) == 2 * n_trials
def test_fit_with_params(
tmp_path: pathlib.Path,
boosting_type: str,
objective: Optional[Union[Callable, str]],
) -> None:
X, y = load_breast_cancer(return_X_y=True)
clf = OGBMClassifier(
boosting_type=boosting_type,
n_estimators=n_estimators,
n_trials=n_trials,
objective=objective,
model_dir=tmp_path,
)
# See https://github.com/microsoft/LightGBM/issues/2328
if boosting_type == "rf" and callable(objective):
with pytest.raises(lgb.basic.LightGBMError):
clf.fit(X, y)
else:
clf.fit(X, y)