def test_plot_input(X_y, clf):
plotter = TreeDependencePlotter(clf).fit(X_y[0], X_y[1])
with pytest.raises(ValueError):
plotter.plot(feature="not a feature")
with pytest.raises(ValueError):
plotter.plot(feature=0, type_binning=5)
with pytest.raises(ValueError):
plotter.plot(feature=0, min_q=1, max_q=0)
def test_fit_complex(complex_data_split, complex_fitted_lightgbm):
X_train, X_test, y_train, y_test = complex_data_split
plotter = TreeDependencePlotter(complex_fitted_lightgbm)
plotter.fit(X_test, y_test)
pd.testing.assert_frame_equal(plotter.X, X_test)
pd.testing.assert_series_equal(plotter.y, pd.Series(y_test, index=X_test.index))
assert plotter.fitted is True
# Check if plotting doesnt cause errors
with patch('matplotlib.pyplot.figure') as mock_plt:
for binning in ["simple", "agglomerative", "quantile"]:
_ = plotter.plot(feature='f2_missing', type_binning=binning)
def test_plot_class_names(X_y, clf):
plotter = TreeDependencePlotter(clf).fit(X_y[0], X_y[1], class_names=["a", "b"])
fig = plotter.plot(feature=0)
assert plotter.class_names == ["a", "b"]
def test_plot_normal(X_y, clf):
plotter = TreeDependencePlotter(clf).fit(X_y[0], X_y[1])
for binning in ["simple", "agglomerative", "quantile"]:
fig = plotter.plot(feature=0, type_binning=binning)