Update hush functions to work on windows; remove random phenotype exa…

…mples to reduce computation time

.Rbuildignore

@@ -2,3 +2,4 @@
 ^\.Rproj\.user$
 ^cran-comments\.md$
 ^revdep$
+^CRAN-SUBMISSION$

CRAN-SUBMISSION

@@ -0,0 +1,3 @@
+Version: 1.0.0
+Date: 2023-05-03 18:27:23 UTC
+SHA: 5e8241576dcf70e63d13d9c1004c7643edecef95

DESCRIPTION

@@ -26,6 +26,7 @@ Imports:
     ggplot2,
     partition,
     doParallel,
+    parallel,
     utils,
     stats,
     tidyr

R/ACDC.R

@@ -54,6 +54,10 @@
 #' @import CCP
 #' @import utils
 #' @import stats
+#' @import tidyr
+#' @import foreach
+#' @import doParallel
+#' @import parallel
 ACDC <- function(fullData, ILC = 0.50, externalVar, identifierList = colnames(fullData), numNodes = 1) {
 
   # to remove "no visible binding" note
@@ -62,7 +66,7 @@ ACDC <- function(fullData, ILC = 0.50, externalVar, identifierList = colnames(fu
   ## function to suppress output
   # use for p.asym
   hush = function(code){
-    sink("/dev/null")
+    sink(nullfile())
     tmp = code
     sink()
     return(tmp)
@@ -78,6 +82,9 @@ ACDC <- function(fullData, ILC = 0.50, externalVar, identifierList = colnames(fu
   if(ncol(fullData) != length(identifierList)) stop("identifierList must be the same length as the number of columns in fullData.")
   if(0 > ILC | 1 < ILC) stop("ILC must be between 0 and 1.")
 
+  # iteration counter
+  i = 0
+
   # partition and find modules
   part    <- partition(fullData, threshold = ILC)
   modules <- part$mapping_key[which(grepl("reduced_var_", part$mapping_key$variable)), ]$indices
@@ -86,13 +93,8 @@ ACDC <- function(fullData, ILC = 0.50, externalVar, identifierList = colnames(fu
   if(length(modules) == 0) stop("No modules identified. Try a smaller ILC.")
 
   # parallel set up
-  my.cluster <- parallel::makeCluster(
-    numNodes, 
-    type = "PSOCK")
-  doParallel::registerDoParallel(cl = my.cluster)
-
-  # iteration counter
-  i = 0
+  my.cluster <- parallel::makeCluster(numNodes)
+  doParallel::registerDoParallel(my.cluster)
 
   # for each module...
   results <- foreach (i = 1:length(modules),

R/ACDChighdim.R

@@ -26,16 +26,7 @@
 #' # load dataset
 #' data("nutrimouse")
 #' 
-#' # generate random phenotype
-#' r.pheno <- rnorm(nrow(nutrimouse$lipid))
-#' 
-#' # run function for random phenotype
-#' ACDChighdim(moduleIdentifier = 1,
-#'             moduleCols = list(1:ncol(nutrimouse$lipid)),
-#'             fullData = nutrimouse$lipid,
-#'             externalVar = r.pheno)
-#' 
-#' # run function for diet and genotype (non-random)
+#' # run function for diet and genotype
 #' ACDChighdim(moduleIdentifier = 1,
 #'             moduleCols = list(1:ncol(nutrimouse$lipid)),
 #'             fullData = nutrimouse$lipid,
@@ -73,7 +64,7 @@ ACDChighdim <- function(moduleIdentifier = 1, moduleCols, fullData, externalVar,
   ## function to suppress output
   # use for p.asym
   hush = function(code){
-    sink("/dev/null")
+    sink(nullfile())
     tmp = code
     sink()
     return(tmp)

R/ACDCmod.R

@@ -25,20 +25,12 @@
 #' # load dataset
 #' data("nutrimouse")
 #' 
-#' # generate random phenotype
-#' r.pheno <- rnorm(nrow(nutrimouse$lipid))
-#' 
 #' # partition dataset and save modules
 #' library(partition)
 #' part <- partition(nutrimouse$lipid, threshold = 0.50)
 #' mods <- part$mapping_key[which(grepl("reduced_var_", part$mapping_key$variable)), ]$mapping
 #' 
-#' # run function for random phenotype
-#' ACDCmod(fullData = nutrimouse$lipid,
-#'         modules=mods,
-#'         externalVar = r.pheno)
-#' 
-#' # run function for diet and genotype (non-random)
+#' # run function for diet and genotype
 #' ACDCmod(fullData = nutrimouse$lipid,
 #'         modules = mods,
 #'         externalVar = data.frame(diet=as.numeric(nutrimouse$diet), 
@@ -66,10 +58,10 @@
 #' @import CCP
 #' @import utils
 #' @import stats
-#' @import parallel
+#' @import tidyr
 #' @import foreach
 #' @import doParallel
-#' @import tidyr
+#' @import parallel
 ACDCmod <- function(fullData, modules, externalVar, identifierList=colnames(fullData), numNodes = 1) {
 
   # to remove "no visible binding" note
@@ -78,7 +70,7 @@ ACDCmod <- function(fullData, modules, externalVar, identifierList=colnames(full
   ## function to suppress output
   # use for p.asym
   hush = function(code){
-    sink("/dev/null")
+    sink(nullfile())
     tmp = code
     sink()
     return(tmp)
@@ -94,15 +86,13 @@ ACDCmod <- function(fullData, modules, externalVar, identifierList=colnames(full
   if(ncol(fullData) != length(identifierList)) stop("identifierList must be the same length as the number of columns in fullData.")
   if(length(modules) == 0) stop("No modules input.")
 
-  # parallel set up
-  my.cluster <- parallel::makeCluster(
-    numNodes, 
-    type = "PSOCK")
-  doParallel::registerDoParallel(cl = my.cluster)
-
   # iteration counter
   i = 0
 
+  # parallel set up
+  my.cluster <- parallel::makeCluster(numNodes)
+  doParallel::registerDoParallel(my.cluster)
+
   # for each module...
   results <- foreach (i = 1:length(modules),
                       .combine = rbind,

R/OSCA_par.R

@@ -33,19 +33,6 @@
 #' # load dataset
 #' data("nutrimouse")
 #' 
-#' # generate random phenotype
-#' r.pheno <- rnorm(nrow(nutrimouse$gene))
-#' 
-#' ## random phenotype
-#' # run function; input path to OSCA software
-#' # OSCA_par(df = nutrimouse$gene, 
-#' #          externalVar = r.pheno, 
-#' #          ILCincrement = 0.25, 
-#' #          oscaPath = "pathHere", 
-#' #          numNodes = detectCores()-1, 
-#' #          permute = T)
-#' 
-#' ## observed external variable
 #' # run function; input path to OSCA software
 #' # OSCA_par(df = nutrimouse$gene, 
 #' #          externalVar = as.numeric(nutrimouse$diet),
@@ -70,9 +57,9 @@
 #' @export
 #' @import data.table
 #' @import partition
-#' @import parallel
 #' @import foreach
 #' @import doParallel
+#' @import parallel
 OSCA_par <- function(df, externalVar, ILCincrement = 0.05, oscaPath, numNodes = 1, permute = T) {
 
   # check correct dimensions of input
@@ -94,10 +81,8 @@ OSCA_par <- function(df, externalVar, ILCincrement = 0.05, oscaPath, numNodes =
   externalVar_permute <- sample(externalVar)
 
   # parallel set up
-  my.cluster <- parallel::makeCluster(
-    numNodes, 
-    type = "PSOCK")
-  doParallel::registerDoParallel(cl = my.cluster)
+  my.cluster <- parallel::makeCluster(numNodes)
+  doParallel::registerDoParallel(my.cluster)
 
   # PVE for each value with permutations or PVE for each value without permutations
   if (permute == T) {

R/OSCA_parPlot.R

@@ -18,22 +18,6 @@
 #' # load dataset
 #' data("nutrimouse")
 #' 
-#' # generate random phenotype
-#' r.pheno <- rnorm(nrow(nutrimouse$gene))
-#' 
-#' ## random phenotype
-#' # run OSCA_par and save output; input path to OSCA software
-#' # r.par <- OSCA_par(df = nutrimouse$gene, 
-#' #                 externalVar = r.pheno, 
-#' #                 ILCincrement = 0.25,
-#' #                 oscaPath = "pathHere", 
-#' #                 numNodes = detectCores()-1, 
-#' #                 permute = T)
-#' 
-#' # run function
-#' #OSCA_parPlot(df=r.par, externalVarName = "Random Phenotype", dataName = "Nutritional Issue Genes")
-#' 
-#' ## observed external variables
 #' # run OSCA_par and save output; input path to OSCA software
 #' # par <- OSCA_par(df = nutrimouse$gene, 
 #' #                 externalVar = as.numeric(nutrimouse$diet),

R/OSCA_singleValue.R

@@ -17,17 +17,7 @@
 #' 
 #' # load dataset
 #' data("nutrimouse")
-#'
-#' # generate random phenotype
-#' r.pheno <- rnorm(nrow(nutrimouse$gene))
-#' 
-#' ## random phenotype
-#' # run function; input path to OSCA software
-#' # OSCA_singleValue(df = nutrimouse$gene, 
-#' #                  externalVar= r.pheno, 
-#' #                  oscaPath = "pathHere")
 #' 
-#' ## observed external variable
 #' # run function; input path to OSCA software
 #' # OSCA_singleValue(df = nutrimouse$gene, 
 #' #                  externalVar = as.numeric(nutrimouse$diet),