BUG: summary_plot ignores plot_type for TreeExplainer

[bhvieira]

Issue Description

In the MWE below, I would expect summary_plot to return a bar plot per class (this was the old behavior). Instead, no matter the argument in plot_type, I always get the interaction dot-plot.

Minimal Reproducible Example

# train a multiclass lgm on random sklearn data
import sklearn
import sklearn.datasets
import sklearn.model_selection
import lightgbm as lgb
import shap

X, y = sklearn.datasets.make_classification(n_samples=100, n_features=20, n_informative=5, n_classes=3, random_state=0)
X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X, y, random_state=1)
gbm = lgb.LGBMClassifier()
gbm.fit(X_train, y_train)

# explain the model's predictions using SHAP
explainer = shap.Explainer(gbm)
shap_values = explainer(X_test)

# summary_plot

shap.summary_plot(shap_values, X_test, plot_type="bar") # wrong plot

shap.summary_plot(shap_values, X_test, plot_type="asdasd asd") # even gibberish works

Traceback

No response

Expected Behavior

When setting plot_type="bar" I would expect shap to return a bar plot, not a compact_dot. When using gibberish arguments, I would expect shap to error out.

Bug report checklist

• I have checked that this issue has not already been reported.
• I have confirmed this bug exists on the latest release of shap.
• I have confirmed this bug exists on the master branch of shap.
• I'd be interested in making a PR to fix this bug

Installed Versions

[[package]]
name = "shap"
version = "0.45.0"
description = "A unified approach to explain the output of any machine learning model."
category = "main"
optional = false
python-versions = ">=3.9"

[[package]]
name = "lightgbm"
version = "4.3.0"
description = "LightGBM Python Package"
category = "main"
optional = false
python-versions = ">=3.6"

[metadata]
lock-version = "1.1"
python-versions = ">=3.10,<3.12"

[CloseChoice]

Thanks for reporting and the MWE, will have a look.