fix: Fix pipeline_test that partially stopped working due to introduction of shap and confusion matrix.

[arima-tsukasa]

@kimusaku
@AkiraUra

• Scripts generated with generatedcode fail with shap.plots.beeswarm.(Scripts generated with generatedcode fail with shap.plots.beeswarm. sapientml#68)
  • beeswarm cannot plot a graph with multiple shap_values.
  • Deal with multi-class and LightGBM binary classification where multiple shap_values are computed as follows:
    • For multiclass, do not plot
    • For LightGBM, specify masker for Explainer.
• Scripts generated with generatedcode fail with shap.Explainer. (Scripts generated with generatedcode fail with shap.Explainer. sapientml#67)
  • If you enter masker uniformly in the Explainer, an error occurs in the calculation of shap_values depending on the data.
  • Specify masker only if necessary.
• There is a lack of action to undo converted values when using LabelEncoder.
  • Add processing immediately for predict, or on input to confusion matrix for predict_proba.

[arima-tsukasa]

Fix to output predict_result.csv in the format specified by predict_option and metric.

• Added None to predict_option of SapientMLConfig.
  • Specify predict method (default: predict(), probability: predict_proba(), None: Comply with metric requirements.)

[kimusaku]

Thanks!