add the possibility to run some steps in parallel and some not

R/VSURF.R

@@ -82,7 +82,9 @@
 #'   passed to each intermediate function: \code{VSURF_thres},
 #'   \code{VSURF_interp}, \code{VSURF_pred}, in this order.
 #' @param parallel A logical indicating if you want VSURF to run in parallel on
-#'   multiple cores (default to FALSE).
+#'   multiple cores (default to FALSE). If a vector of length 3 is given,
+#'   each coordinate is passed to each intermediate function: \code{VSURF_thres},
+#'   \code{VSURF_interp}, \code{VSURF_pred}, in this order.
 #' @param ncores Number of cores to use. Default is set to the number of cores
 #'   detected by R minus 1.
 #' @param clusterType Type of the multiple cores cluster used to run VSURF in
@@ -97,6 +99,7 @@
 #'   \code{VSURF_interp}, in this order.
 #' @param verbose A logical indicating if information about method's progress
 #' (included progress bars for each step) must be printed (default to TRUE).
+#' Adds a small extra overload.
 #' @param ...  others parameters to be passed on to the \code{randomForest}
 #'   function (see ?randomForest for further information).
 #' 
@@ -215,39 +218,39 @@ VSURF <- function (x, ...) {
 VSURF.default <- function(
   x, y, ntree = 2000, mtry = max(floor(ncol(x)/3), 1),
   nfor.thres = 50, nmin = 1, nfor.interp = 25, nsd = 1, nfor.pred = 25, nmj = 1,
-  RFimplem = "randomForest", parallel = FALSE,
-  ncores = detectCores() - 1, clusterType = "PSOCK", verbose = TRUE, ...) {
+  RFimplem = "randomForest", parallel = FALSE, ncores = detectCores() - 1,
+  clusterType = "PSOCK", verbose = TRUE, ...) {
 
   start <- Sys.time()
 
-  if (!parallel) {
-    clusterType <- NULL
-    ncores <- NULL
-  }
-
   thres <- VSURF_thres(
     x=x, y=y, ntree=ntree, mtry=mtry, nfor.thres=nfor.thres, nmin=nmin,
     RFimplem = ifelse(length(RFimplem) == 3, RFimplem[1], RFimplem),
-    parallel=parallel,
-    clusterType = ifelse(length(clusterType) == 2, clusterType[1], clusterType),
+    parallel = ifelse(length(parallel) == 3, parallel[1], parallel),
+    clusterType = ifelse(length(clusterType) > 1, clusterType[1], clusterType),
     ncores=ncores, verbose = verbose, ...)
 
   interp <- VSURF_interp(
     x=x, y=y, ntree=ntree, vars=thres$varselect.thres, nfor.interp=nfor.interp,
     nsd=nsd, RFimplem = ifelse(length(RFimplem) == 3, RFimplem[2], RFimplem),
-    parallel=parallel,
-    clusterType = ifelse(length(clusterType) == 2, clusterType[2], clusterType),
+    parallel = ifelse(length(parallel) == 3, parallel[2], parallel),
+    clusterType = ifelse(length(clusterType) > 1, clusterType[2], clusterType),
     ncores=ncores, verbose = verbose, ...)
 
   pred <- VSURF_pred(x=x, y=y, ntree=ntree, err.interp=interp$err.interp,
     varselect.interp=interp$varselect.interp, nfor.pred=nfor.pred, nmj=nmj,
     RFimplem = ifelse(length(RFimplem) == 3, RFimplem[3], RFimplem),
-    verbose = verbose,...)
+    parallel = ifelse(length(parallel) == 3, parallel[3], parallel),
+    ncores = ncores, verbose = verbose, ...)
 
   cl <- match.call()
   cl[[1]] <- as.name("VSURF")
 
-  overall.time <- Sys.time()-start
+  if (identical(parallel, FALSE) | identical(parallel, rep(FALSE, 3))) {
+    clusterType <- NULL
+    ncores <- NULL
+  }
+  overall.time <- Sys.time() - start
 
   output <- list('varselect.thres'=thres$varselect.thres,
                  'varselect.interp'=interp$varselect.interp,

R/VSURF_interp.R

@@ -117,7 +117,6 @@ VSURF_interp.default <- function(
   if (verbose == TRUE) cat(paste("\nInterpretation step (on", length(vars), "variables)\n"))
 
   if (!parallel) {
-    clusterType <- NULL
     ncores <- NULL
   }  
 
@@ -247,7 +246,7 @@ VSURF_interp.default <- function(
   }
 
   # initialization of the progress bar
-  if (verbose == TRUE & parallel == FALSE) {
+  if (verbose == TRUE & (parallel == FALSE | clusterType %in% c("ranger", "Rborist"))) {
     pb <- utils::txtProgressBar(style = 3)
     nBar <- 1
   }
@@ -301,6 +300,7 @@ VSURF_interp.default <- function(
     }
   } else {
     if (clusterType == "ranger") {
+      if (RFimplem != "ranger") stop("RFimplem must be set to 'ranger' to use clusterType 'ranger'")
       for (i in 1:nvars){
         res <- rf.interp.ranger(i, nfor.interp, num.threads = ncores, ...)
         err.interp[i] <- res[1]
@@ -312,6 +312,7 @@ VSURF_interp.default <- function(
       }
     } else {
       if (clusterType == "Rborist") {
+        if (RFimplem != "Rborist") stop("RFimplem must be set to 'Rborist' to use clusterType 'Rborist'")
         for (i in 1:nvars){
           res <- rf.interp.Rborist(i, nfor.interp, nThread = ncores, ...)
           err.interp[i] <- res[1]
@@ -396,6 +397,8 @@ VSURF_interp.default <- function(
   cl <- match.call()
   cl[[1]] <- as.name("VSURF_interp")
 
+  if (!parallel) clusterType <- NULL
+
   comput.time <- Sys.time()-start
 
   output <- list('varselect.interp'=varselect,

R/VSURF_pred.R

@@ -96,7 +96,8 @@ VSURF_pred <- function (x, ...) {
 #' @rdname VSURF_pred
 #' @export
 VSURF_pred.default <-function(x, y, ntree = 2000, err.interp, varselect.interp,
-  nfor.pred = 25, nmj = 1, RFimplem = "randomForest", verbose = TRUE, ...) {
+  nfor.pred = 25, nmj = 1, RFimplem = "randomForest", parallel = FALSE,
+  ncores = detectCores()-1, verbose = TRUE, ...) {
 
   # err.interp: interpretation models errors
   # varselect.interp: interpretation variables indices
@@ -197,15 +198,16 @@ did not eliminate variables")
     if (RFimplem == "ranger") {
       dat <- cbind(w, "y" = y)
       for (j in 1:nfor.pred) {
-      rf[j] <- ranger::ranger(dependent.variable.name="y", data=dat,
-                              num.trees=ntree,
-                              ...)$prediction.error
+        rf[j] <- ranger::ranger(dependent.variable.name="y", data=dat,
+                                num.threads = ifelse(parallel, ncores, 1),
+                                num.trees=ntree, ...)$prediction.error
       }
       err.pred <- mean(rf)
     }
     if (RFimplem == "Rborist") {
       for (j in 1:nfor.pred) {
         rf[j] <- Rborist::Rborist(x = w, y = y, nTree = ntree, minInfo = 0,
+                                  nThread = ifelse(parallel, ncores, 1),
                                   ...)$validation$oobError
       }
       err.pred <- mean(rf)
@@ -251,28 +253,30 @@ did not eliminate variables")
           if (i <= n) {
             for (j in 1:nfor.pred) {
               rf[j] <- ranger::ranger(dependent.variable.name="y", data=dat,
+                                      num.threads = ifelse(parallel, ncores, 1),
                                       num.trees=ntree, ...)$prediction.error
             }
           } else {
             for (j in 1:nfor.pred) {
               rf[j] <- ranger::ranger(dependent.variable.name="y", data=dat,
-                                      mtry=i/3, num.trees=ntree,
-                                      ...)$prediction.error
+                                      num.threads = ifelse(parallel, ncores, 1),
+                                      num.trees=ntree, mtry=i/3, ...)$prediction.error
             }
           }
           z <- mean(rf)
         }
         if (RFimplem == "Rborist") {
           if (i <= n) {
             for (j in 1:nfor.pred) {
-              rf[j] <- Rborist::Rborist(x = w, y = y, nTree = ntree,
-                                        minInfo = 0, ...)$validation$oobError
+              rf[j] <- Rborist::Rborist(x = w, y = y, nTree = ntree, minInfo = 0,
+                                        nThread = ifelse(parallel, ncores, 1),
+                                        ...)$validation$oobError
             }
           } else {
             for (j in 1:nfor.pred) {
-              rf[j] <- Rborist::Rborist(x = w, y = y, nTree = ntree,
-                                        minInfo = 0, predFixed = i/3,
-                                        ...)$validation$oobError
+              rf[j] <- Rborist::Rborist(x = w, y = y, nTree = ntree, minInfo = 0,
+                                        nThread = ifelse(parallel, ncores, 1),
+                                        predFixed = i/3, ...)$validation$oobError
             }
           }
           z <- mean(rf)

R/VSURF_thres.R

@@ -123,7 +123,6 @@ VSURF_thres.default <- function(
   if (verbose == TRUE) cat(paste("Thresholding step\n"))
 
   if (!parallel) {
-    clusterType <- NULL
     ncores <- NULL
   } else {
     ncores <- min(nfor.thres, ncores)
@@ -162,6 +161,7 @@ VSURF_thres.default <- function(
   # filling of perf with the nfor.thres forests OOB errors
 
   if (RFimplem == "Rborist") RFimplem <- "randomForest"
+  if (clusterType == "Rborist") clusterType <- "PSOCK"
 
   if (RFimplem == "randomForest") {
     rf.classif <- function(i, ...) {
@@ -208,7 +208,7 @@ VSURF_thres.default <- function(
   }
 
   # initialization of the progress bar
-  if (verbose == TRUE & parallel == FALSE) {
+  if (verbose == TRUE & (parallel == FALSE | clusterType == "ranger")) {
     pb <- utils::txtProgressBar(style = 3)
     nBar <- 1
   }
@@ -251,6 +251,7 @@ VSURF_thres.default <- function(
     }
   } else {
     if (clusterType == "ranger") {
+      if (RFimplem != "ranger") stop("RFimplem must be set to 'ranger' to use clusterType 'ranger'")
       for (i in 1:nfor.thres) {
         rf <- rf.ranger(i, num.threads = ncores, ...)
         m[i,] <- rf$m
@@ -371,6 +372,8 @@ VSURF_thres.default <- function(
   cl <- match.call()
   cl[[1]] <- as.name("VSURF_thres")
 
+  if (!parallel) clusterType <- NULL
+
   comput.time <- Sys.time()-start
 
   output <- list('varselect.thres'=varselect.thres,