sv_importance returns unused argument error

[Karol-Gawlowski]

Calling sv_importance results in:

Error in FUN(newX[, i], ...) : unused argument (simplify = FALSE)

Looking through shapviz/tree/main/R)/sv_importance.R; the .min_max_scale helper function indeed does not include the simplify argument,

repro steps as in the shapviz vignette

X_train_test <- data.matrix(iris[, -1])
dtrain_test <- xgboost::xgb.DMatrix(X_train_test, label = iris[, 1])
fit_test <- xgboost::xgb.train(data = dtrain_test, nrounds = 50)
x_test <- shapviz(fit_test, X_pred = X_train_test)
sv_importance(x_test)
sv_importance(x_test, kind = "beeswarm", show_numbers = TRUE)

Note, the following works as intended:
sv_importance(x_test, kind = "no")

Cheers

[mayer79]

Hello and thanks for reporting. The simplify option refers to apply(..., simplify = FALSE), which was introduced in R 4.1. I was not aware of this version dependency, so I'd love to replace the snipped by something that also runs on lower R versions.

[mayer79]

Fixed by #30

Test with

devtools::install_github("mayer79/shapviz")

[mayer79]

Now fixed on CRAN