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_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_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