fix: update xgboost api and add depends R>=4.1.0 to description (#51)

* chore: added depends R>=4.1.0 to description

to fix cran note

* fix: update code to new v3 xgboost api

* fix: updated more old xgboost api calls

* test: fixed treeshap-correctness tests for xgboost

* test: fixed xgboost unit-tests

restored expected behaviour of 'predictions from unified == original predictions'
by now controlling the 'base_score' parameter in order
to prevent automated calculation of the intercept, as it was
introduced to xgboost with v2 and later versions

* test: fix issue with ambigious argument

renamed argument 'model' -> 'model_name' to avoid issues
with also used object named 'model'

* fix: back to new split-condition logic for xgboost

as split value is considered to be 'less than'

* test: quite a hacky workaround to get original unit-tests working

* chore: code formatting and removed library-import of treeshap at beginning of test-files

* test: conditional testing of examples and unit-tests using suggested deps

* chore: re-formatted examples for randomForest and fixed indentation for conditional testing

* chore: replaced deprecated size argument of ggplots element_line

* chore: some house-keeping

renamed globals.R -> zzz.R and moved all import statements there

Author: kapsner

DESCRIPTION

@@ -1,7 +1,7 @@
 Package: treeshap
 Title: Compute SHAP Values for Your Tree-Based Models Using the 'TreeSHAP'
     Algorithm
-Version: 0.3.1.9000
+Version: 0.3.2
 Authors@R: c(
     person("Konrad", "Komisarczyk", email = "komisarczykkonrad@gmail.com", role = "aut"),
     person("Pawel", "Kozminski", email = "pkozminski99@gmail.com", role = "aut"),
@@ -21,7 +21,7 @@ URL: https://modeloriented.github.io/treeshap/,
     https://github.com/ModelOriented/treeshap
 BugReports: https://github.com/ModelOriented/treeshap/issues
 Depends: 
-    R (>= 2.10)
+    R (>= 4.1.0)
 Imports: 
     data.table,
     ggplot2,
@@ -41,4 +41,4 @@ LinkingTo:
 Encoding: UTF-8
 LazyData: true
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.2.3
+RoxygenNote: 7.3.3

R/globals.R

@@ -17,8 +17,6 @@
 #'
 #' @export
 #'
-#' @import ggplot2
-#'
 #' @seealso
 #' \code{\link{treeshap}} for calculation of SHAP values
 #'
@@ -27,17 +25,24 @@
 #'
 #' @examples
 #' \donttest{
-#' library(xgboost)
-#' data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
-#' target <- fifa20$target
-#' param <- list(objective = "reg:squarederror", max_depth = 3)
-#' xgb_model <- xgboost::xgboost(as.matrix(data), params = param, label = target,
-#'                               nrounds = 20, verbose = FALSE)
-#' unified_model <- xgboost.unify(xgb_model, as.matrix(data))
-#' x <- head(data, 1)
-#' shap <- treeshap(unified_model, x)
-#' plot_contribution(shap, 1,  min_max = c(0, 120000000))
-#' }
+#' if (requireNamespace("xgboost", quietly = TRUE) &&
+#'  requireNamespace("scales", quietly = TRUE)) {
+#'   library(xgboost)
+#'   data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
+#'   target <- fifa20$target
+#'   xgb_model <- xgboost::xgboost(
+#'    x = as.matrix(data),
+#'    y = target,
+#'    objective = "reg:squarederror",
+#'    max_depth = 3,
+#'    nrounds = 20
+#'   )
+#'   unified_model <- xgboost.unify(xgb_model, as.matrix(data))
+#'   x <- head(data, 1)
+#'   shap <- treeshap(unified_model, x)
+#'   plot_contribution(shap, 1,  min_max = c(0, 120000000))
+#' }}
+#' 
 plot_contribution <- function(treeshap,
                               obs = 1,
                               max_vars = 5,
@@ -46,6 +51,12 @@ plot_contribution <- function(treeshap,
                               explain_deviation = FALSE,
                               title = "SHAP Break-Down",
                               subtitle = "") {
+  if (!requireNamespace("scales", quietly = TRUE)) {
+    stop(
+      "Package \"scales\" needed for this function to work. Please install it.",
+      call. = FALSE
+    )
+  }
 
   shap <- treeshap$shaps[obs, ]
   model <- treeshap$unified_model$model

R/plot_contribution.R

@@ -11,8 +11,6 @@
 #'
 #' @export
 #'
-#' @import ggplot2
-#'
 #' @seealso
 #' \code{\link{treeshap}} for calculation of SHAP values
 #'
@@ -21,19 +19,31 @@
 #'
 #' @examples
 #' \donttest{
-#' library(xgboost)
-#' data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
-#' target <- fifa20$target
-#' param <- list(objective = "reg:squarederror", max_depth = 3)
-#' xgb_model <- xgboost::xgboost(as.matrix(data), params = param, label = target,
-#'                               nrounds = 20, verbose = FALSE)
-#' unified_model <- xgboost.unify(xgb_model, as.matrix(data))
-#' x <- head(data, 100)
-#' shaps <- treeshap(unified_model, x)
-#' plot_feature_dependence(shaps, variable = "overall")
-#' }
+#' if (requireNamespace("xgboost", quietly = TRUE) &&
+#'  requireNamespace("scales", quietly = TRUE)) {
+#'   library(xgboost)
+#'   data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
+#'   target <- fifa20$target
+#'   xgb_model <- xgboost::xgboost(
+#'    x = as.matrix(data),
+#'    y = target,
+#'    objective = "reg:squarederror",
+#'    max_depth = 3,
+#'    nrounds = 20
+#'   )
+#'   unified_model <- xgboost.unify(xgb_model, as.matrix(data))
+#'   x <- head(data, 100)
+#'   shaps <- treeshap(unified_model, x)
+#'   plot_feature_dependence(shaps, variable = "overall")
+#' }}
 plot_feature_dependence <- function(treeshap, variable,
                                     title = "Feature Dependence", subtitle = NULL) {
+  if (!requireNamespace("scales", quietly = TRUE)) {
+    stop(
+      "Package \"scales\" needed for this function to work. Please install it.",
+      call. = FALSE
+    )
+  }
 
   shaps <- treeshap$shaps
   x <- treeshap$observations

R/plot_feature_dependence.R

@@ -13,9 +13,6 @@
 #' @return a \code{ggplot2} object
 #'
 #' @export
-#' @import ggplot2
-#' @importFrom stats reorder
-#' @importFrom graphics text
 #'
 #' @seealso
 #' \code{\link{treeshap}} for calculation of SHAP values
@@ -25,21 +22,33 @@
 #'
 #' @examples
 #' \donttest{
-#' library(xgboost)
-#' data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
-#' target <- fifa20$target
-#' param <- list(objective = "reg:squarederror", max_depth = 3)
-#' xgb_model <- xgboost::xgboost(as.matrix(data), params = param, label = target,
-#'                               nrounds = 20, verbose = FALSE)
-#' unified_model <- xgboost.unify(xgb_model, as.matrix(data))
-#' shaps <- treeshap(unified_model, as.matrix(head(data, 3)))
-#' plot_feature_importance(shaps, max_vars = 4)
-#' }
+#' if (requireNamespace("xgboost", quietly = TRUE) &&
+#'  requireNamespace("scales", quietly = TRUE)) {
+#'   library(xgboost)
+#'   data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
+#'   target <- fifa20$target
+#'   xgb_model <- xgboost::xgboost(
+#'    x = as.matrix(data),
+#'    y = target,
+#'    objective = "reg:squarederror",
+#'    max_depth = 3,
+#'    nrounds = 20
+#'   )
+#'   unified_model <- xgboost.unify(xgb_model, as.matrix(data))
+#'   shaps <- treeshap(unified_model, as.matrix(head(data, 3)))
+#'   plot_feature_importance(shaps, max_vars = 4)
+#' }}
 plot_feature_importance <- function(treeshap,
                                     desc_sorting = TRUE,
                                     max_vars = ncol(shaps),
                                     title = "Feature Importance",
                                     subtitle = NULL) {
+  if (!requireNamespace("scales", quietly = TRUE)) {
+    stop(
+      "Package \"scales\" needed for this function to work. Please install it.",
+      call. = FALSE
+    )
+  }
   shaps <- treeshap$shaps
 
   # argument check

R/plot_feature_importance.R

@@ -13,8 +13,6 @@
 #'
 #' @export
 #'
-#' @import ggplot2
-#'
 #' @seealso
 #' \code{\link{treeshap}} for calculation of SHAP Interaction values
 #'
@@ -23,17 +21,30 @@
 #'
 #' @examples
 #' \donttest{
-#' data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
-#' target <- fifa20$target
-#' param2 <- list(objective = "reg:squarederror", max_depth = 5)
-#' xgb_model2 <- xgboost::xgboost(as.matrix(data), params = param2, label = target, nrounds = 10)
-#' unified_model2 <- xgboost.unify(xgb_model2, data)
-#' inters <- treeshap(unified_model2, as.matrix(data[1:50, ]), interactions = TRUE)
-#' plot_interaction(inters, "dribbling", "defending")
-#' }
+#' if (requireNamespace("xgboost", quietly = TRUE) &&
+#'  requireNamespace("scales", quietly = TRUE)) {
+#'   data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
+#'   target <- fifa20$target
+#'   xgb_model2 <- xgboost::xgboost(
+#'    x = as.matrix(data),
+#'    y = target,
+#'    objective = "reg:squarederror",
+#'    max_depth = 5,
+#'    nrounds = 10
+#'   )
+#'   unified_model2 <- xgboost.unify(xgb_model2, data)
+#'   inters <- treeshap(unified_model2, as.matrix(data[1:50, ]), interactions = TRUE)
+#'   plot_interaction(inters, "dribbling", "defending")
+#' }}
 plot_interaction <- function(treeshap, var1, var2,
                              title = "SHAP Interaction Value Plot",
                              subtitle = "") {
+  if (!requireNamespace("scales", quietly = TRUE)) {
+    stop(
+      "Package \"scales\" needed for this function to work. Please install it.",
+      call. = FALSE
+    )
+  }
 
   interactions <- treeshap$interactions
   x <- treeshap$observations

R/plot_interaction.R

@@ -12,19 +12,21 @@
 #'
 #' @examples
 #' \donttest{
-#' library(gbm)
-#' data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
-#' data['value_eur'] <- fifa20$target
-#' gbm_model <- gbm::gbm(
-#'   formula = value_eur ~ .,
-#'   data = data,
-#'   distribution = "laplace",
-#'   n.trees = 20,
-#'   interaction.depth = 4,
-#'   n.cores = 1)
+#' if (requireNamespace("gbm", quietly = TRUE)) {
+#'   library(gbm)
+#'   data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
+#'   data['value_eur'] <- fifa20$target
+#'   gbm_model <- gbm::gbm(
+#'     formula = value_eur ~ .,
+#'     data = data,
+#'     distribution = "laplace",
+#'     n.trees = 20,
+#'     interaction.depth = 4,
+#'     n.cores = 1
+#'   )
 #'   unified <- gbm.unify(gbm_model, data)
 #'   predict(unified, data[2001:2005, ])
-#'   }
+#' }}
 predict.model_unified <- function(object, x, ...) {
   unified_model <- object
   model <- unified_model$model

R/predict.R

@@ -27,19 +27,20 @@
 #'
 #' @examples
 #' \donttest{
-#' library(gbm)
-#' data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
-#' data['value_eur'] <- fifa20$target
-#' gbm_model <- gbm::gbm(
-#' formula = value_eur ~ .,
-#'   data = data,
-#'   distribution = "laplace",
-#'   n.trees = 20,
-#'   interaction.depth = 4,
-#'   n.cores = 1)
-#' unified <- gbm.unify(gbm_model, data)
-#' set_reference_dataset(unified, data[200:700, ])
-#' }
+#' if (requireNamespace("gbm", quietly = TRUE)) {
+#'   library(gbm)
+#'   data <- fifa20$data[colnames(fifa20$data) != 'work_rate']
+#'   data['value_eur'] <- fifa20$target
+#'   gbm_model <- gbm::gbm(
+#'   formula = value_eur ~ .,
+#'     data = data,
+#'     distribution = "laplace",
+#'     n.trees = 20,
+#'     interaction.depth = 4,
+#'     n.cores = 1)
+#'   unified <- gbm.unify(gbm_model, data)
+#'   set_reference_dataset(unified, data[200:700, ])
+#' }}
 set_reference_dataset <- function(unified_model, x) {
   model <- unified_model$model
   data <- x