refactor: `addLayer` to use `mclapply`

DESCRIPTION

 Type: Package
 Package: RFSurrogates
 Title: Surrogate Minimal Depth Variable Importance
-Version: 0.3.3.9000
+Version: 0.3.3.9001
 Authors@R: c(
     person("Stephan", "Seifert", , "stephan.seifert@uni-hamburg.de", role = c("aut", "cre"),
            comment = c(ORCID = "0000-0003-2567-5728")),

NEWS.md

 # RFSurrogates (development version)
 
+* [`addLayer`]: Refactor for-loop to lapply.
+  * Add `num.threads` param to enable parallelization using [`parallel::mclapply`]. It defaults to 1 for backward compatability.
+
 # RFSurrogates 0.3.3
 
 * Fixed `meanAdjAgree()` bug which caused mean adjusted agreement pairings to be set to NA incorrectly when `variables` was a subset or differently ordered than `candidates`.

R/addLayer.R

 #' Add layer information to a forest that was created by getTreeranger
 #'
 #' This functions adds the layer information to each node in a list with trees that was obtained by getTreeranger.
+#' You should use [`getTreeranger()`] with `add_layer = TRUE` instead.
+#'
+#' @param trees The output of [`getTreeranger()`].
+#' @param num.threads (Default: 1) Number of threads to spawn for parallelization.
+#'
+#' @returns A list of tree data frames of length `RF$num.trees`.
+#' Each row of the tree data frames corresponds to a node of the respective tree and the columns correspond to:
+#' * `nodeID`: ID of the respective node (important for left and right daughters in the next columns)
+#' * `leftdaughter`: ID of the left daughter of this node
+#' * `rightdaughter`: ID of the right daughter of this node
+#' * `splitvariable`: ID of the split variable
+#' * `splitpoint`: Split point of the split variable.
+#'    For categorical variables this is a comma separated lists of values, representing the factor levels (in the original order) going to the right.
+#' * `status`: `0` for terminal (`splitpoint` is `NA`) and `1` for non-terminal.
+#' * `layer`: Tree layer depth information, starting at 0 (root node) and incremented for each layer.
 #'
-#' @param trees list of trees created by getTreeranger
-#' @return a list with trees. Each row of the list elements corresponds to a node of the respective tree and the columns correspond to:
-#' \itemize{
-#' \item nodeID: ID of the respective node (important for left and right daughters in the next columns)
-#' \item leftdaughter: ID of the left daughter of this node
-#' \item rightdaughter: ID of the right daughter of this node
-#' \item splitvariable: ID of the split variable
-#' \item splitpoint: splitpoint of the split variable
-#' \item status: "0" for terminal and "1" for non-terminal
-#' \item layer: layer information (0 means root node, 1 means 1 layer below root, etc)
-#' }
 #' @export
-addLayer <- function(trees) {
-  # This function adds the respective layer to the different nodes in a tree. The tree has to be prepared by getTree function
-  tree.layer <- list()
-  num.trees <- length(trees)
-  for (i in 1:num.trees) {
-    tree <- trees[[i]]
+addLayer <- function(trees, num.threads = 1) {
+  parallel::mclapply(trees, add_layer_to_tree, mc.cores = num.threads)
+}
+
+#' Internal function
+#'
+#' This function adds the respective layer to the different nodes in a tree.
+#' The tree has to be prepared by getTree function.
+#'
+#' @param tree A tree data frame from [getTreeranger()].
+#'
+#' @returns A tree data frame with `layer` added.
+#'
+#' @seealso [addLayer()]
+#'
+#' @keywords internal
+add_layer_to_tree <- function(tree) {
   layer <- rep(NA, nrow(tree))
   layer[1] <- 0
   t <- 1
@@ -30,7 +45,6 @@ addLayer <- function(trees) {
   }
   tree <- cbind(tree, layer)
   tree <- tree[order(as.numeric(tree[, "layer"])), ]
-    tree.layer[[i]] <- tree
-  }
-  return(tree.layer)
+
+  return(tree)
 }

man/addLayer.Rd

@@ -4,23 +4,28 @@
 \alias{addLayer}
 \title{Add layer information to a forest that was created by getTreeranger}
 \usage{
-addLayer(trees)
+addLayer(trees, num.threads = 1)
 }
 \arguments{
-\item{trees}{list of trees created by getTreeranger}
+\item{trees}{The output of \code{\link[=getTreeranger]{getTreeranger()}}.}
+
+\item{num.threads}{(Default: 1) Number of threads to spawn for parallelization.}
 }
 \value{
-a list with trees. Each row of the list elements corresponds to a node of the respective tree and the columns correspond to:
+A list of tree data frames of length \code{RF$num.trees}.
+Each row of the tree data frames corresponds to a node of the respective tree and the columns correspond to:
 \itemize{
-\item nodeID: ID of the respective node (important for left and right daughters in the next columns)
-\item leftdaughter: ID of the left daughter of this node
-\item rightdaughter: ID of the right daughter of this node
-\item splitvariable: ID of the split variable
-\item splitpoint: splitpoint of the split variable
-\item status: "0" for terminal and "1" for non-terminal
-\item layer: layer information (0 means root node, 1 means 1 layer below root, etc)
+\item \code{nodeID}: ID of the respective node (important for left and right daughters in the next columns)
+\item \code{leftdaughter}: ID of the left daughter of this node
+\item \code{rightdaughter}: ID of the right daughter of this node
+\item \code{splitvariable}: ID of the split variable
+\item \code{splitpoint}: Split point of the split variable.
+For categorical variables this is a comma separated lists of values, representing the factor levels (in the original order) going to the right.
+\item \code{status}: \code{0} for terminal (\code{splitpoint} is \code{NA}) and \code{1} for non-terminal.
+\item \code{layer}: Tree layer depth information, starting at 0 (root node) and incremented for each layer.
 }
 }
 \description{
 This functions adds the layer information to each node in a list with trees that was obtained by getTreeranger.
+You should use \code{\link[=getTreeranger]{getTreeranger()}} with \code{add_layer = TRUE} instead.
 }