wehub-resource-sync
122 lines
3.3 KiB
Python
122 lines
3.3 KiB
Python
import matplotlib.pyplot as plt
|
|
import numpy as np
|
|
import pytest
|
|
|
|
import shap
|
|
from shap.plots.colors import (
|
|
blue_rgb,
|
|
gray_rgb,
|
|
light_blue_rgb,
|
|
light_red_rgb,
|
|
red_blue,
|
|
red_blue_circle,
|
|
red_blue_no_bounds,
|
|
red_blue_transparent,
|
|
red_rgb,
|
|
red_transparent_blue,
|
|
red_white_blue,
|
|
transparent_blue,
|
|
transparent_red,
|
|
)
|
|
from shap.utils._exceptions import DimensionError
|
|
|
|
|
|
@pytest.fixture(
|
|
params=[
|
|
blue_rgb,
|
|
gray_rgb,
|
|
light_blue_rgb,
|
|
light_red_rgb,
|
|
red_blue,
|
|
red_blue_circle,
|
|
red_blue_no_bounds,
|
|
red_blue_transparent,
|
|
red_rgb,
|
|
red_transparent_blue,
|
|
red_white_blue,
|
|
transparent_blue,
|
|
transparent_red,
|
|
]
|
|
)
|
|
def color(request):
|
|
return request.param
|
|
|
|
|
|
def test_beeswarm_input_is_explanation():
|
|
"""Checks an error is raised if a non-Explanation object is passed as input."""
|
|
with pytest.raises(
|
|
TypeError,
|
|
match="beeswarm plot requires an `Explanation` object",
|
|
):
|
|
_ = shap.plots.beeswarm(np.random.randn(20, 5), show=False) # type: ignore
|
|
|
|
|
|
def test_beeswarm_wrong_features_shape():
|
|
"""Checks that DimensionError is raised if the features data matrix
|
|
has an incompatible shape with the shap_values matrix.
|
|
"""
|
|
rs = np.random.RandomState(42)
|
|
|
|
emsg = (
|
|
"The shape of the shap_values matrix does not match the shape of "
|
|
"the provided data matrix. Perhaps the extra column"
|
|
)
|
|
with pytest.raises(DimensionError, match=emsg):
|
|
expln = shap.Explanation(
|
|
values=rs.randn(20, 5),
|
|
data=rs.randn(20, 4),
|
|
)
|
|
shap.plots.beeswarm(expln, show=False)
|
|
|
|
emsg = "The shape of the shap_values matrix does not match the shape of the provided data matrix."
|
|
with pytest.raises(DimensionError, match=emsg):
|
|
expln = shap.Explanation(
|
|
values=rs.randn(20, 5),
|
|
data=rs.randn(20, 1),
|
|
)
|
|
shap.plots.beeswarm(expln, show=False)
|
|
|
|
|
|
@pytest.mark.mpl_image_compare
|
|
def test_beeswarm(explainer):
|
|
"""Check a beeswarm chart renders correctly with shap_values as an Explanation
|
|
object (default settings).
|
|
"""
|
|
fig = plt.figure()
|
|
shap_values = explainer(explainer.data)
|
|
shap.plots.beeswarm(shap_values, show=False)
|
|
plt.tight_layout()
|
|
return fig
|
|
|
|
|
|
@pytest.mark.mpl_image_compare
|
|
def test_beeswarm_no_group_remaining(explainer):
|
|
"""Beeswarm with group_remaining_features=False."""
|
|
fig = plt.figure()
|
|
shap_values = explainer(explainer.data)
|
|
shap.plots.beeswarm(shap_values, show=False, group_remaining_features=False)
|
|
plt.tight_layout()
|
|
return fig
|
|
|
|
|
|
def test_beeswarm_basic_explanation_works():
|
|
# GH 3901
|
|
explanation = shap.Explanation([[1.0, 2.0, 3.0]])
|
|
shap.plots.beeswarm(explanation, show=False)
|
|
|
|
|
|
def test_beeswarm_works_with_colors(color):
|
|
# GH 3901
|
|
explanation = shap.Explanation([[1.0, 2.0, 3.0]])
|
|
shap.plots.beeswarm(explanation, show=False, color_bar=True, color=color)
|
|
|
|
|
|
def test_beeswarm_colors_values_with_data(color):
|
|
np.random.seed(42)
|
|
|
|
explanation = shap.Explanation(
|
|
values=np.random.randn(100, 5),
|
|
data=np.array([["cat"] * 5] * 100),
|
|
)
|
|
shap.plots.beeswarm(explanation, show=False, color_bar=True, color=color)
|