Stripped out some annoying whitespace on empty lines.

R/aggregateReference.R

@@ -140,7 +140,7 @@ aggregateReference <- function(
     first <- labels[vapply(by.label, function(i) i[1], 0L)]
     num <- vapply(output.vals, ncol, 0L)
     output.labels <- rep(first, num)
-    
+
     output <- SummarizedExperiment(list(logcounts=do.call(cbind, output.vals)), colData=DataFrame(label=output.labels))
     colnames(output) <- sprintf("%s.%s", output.labels, sequence(num))
     output

R/combineRecomputedResults.R

@@ -186,7 +186,7 @@ combineRecomputedResults <- function(
     # Organizing the outputs.
     base.scores <- vector("list", length(results))
     for (r in seq_along(base.scores)) {
-        mat <- results[[r]]$scores   
+        mat <- results[[r]]$scores
         mat[] <- NA_real_
         idx <- cbind(seq_len(nrow(mat)), collated[[r]] + 1L)
         mat[idx] <- irun$scores[,r]

R/getClassicMarkers.R

@@ -83,7 +83,7 @@ getClassicMarkers <- function(ref, labels, assay.type="logcounts", check.missing
         colnames(gm) <- levels(flabels)
         ref[[i]] <- gm
     }
-    
+
     ulabels <- .get_levels(unlist(lapply(ref, colnames)))
     labels <- list()
     for (i in seq_along(ref)) {

R/plotScoreDistribution.R

@@ -211,7 +211,7 @@ plotScoreDistribution <- function(
             name = labels.title,
             breaks = c("assigned", "pruned", "other"),
             values = c(this.color, pruned.color, other.color))
-    
+
     jit <- ggplot2::geom_jitter(height = 0, width = 0.3, color = "black",
         shape = 16, size = size, na.rm = TRUE)
 
@@ -223,7 +223,7 @@ plotScoreDistribution <- function(
     p <- p + ggplot2::theme_classic() +
         ggplot2::facet_wrap(facets = ~label, ncol = ncol) +
         ggplot2::ylab(scores.title)
-    
+
     if (nlevels(as.factor(df$label)) == 1) {
         p <- p + ggplot2::scale_x_discrete(name = NULL, labels = NULL)
     } else {

R/pruneScores.R

@@ -90,7 +90,7 @@
 pruneScores <- function(results, nmads=3, min.diff.med=-Inf, min.diff.next=0, get.thresholds=FALSE) {
     delta <- getDeltaFromMedian(results)
     keep <- delta >= min.diff.med
-    
+
     dn <- results$delta.next
     if (!is.null(dn)) {
         keep <- keep & dn >= min.diff.next

src/find_classic_markers.cpp

@@ -33,7 +33,7 @@ Rcpp::List find_classic_markers(int nlabels, int ngenes, Rcpp::List labels, Rcpp
         }
         lab_ptrs.push_back(static_cast<const int*>(lab_vec.back().begin()));
     }
-    
+
     singlepp::ChooseClassicMarkersOptions opts;
     opts.number = de_n;
     opts.num_threads = nthreads;

tests/testthat/setup.R

@@ -4,13 +4,13 @@ set.seed(100)
 ###########################################
 ## Mocking up some example training data ##
 ###########################################
-                                                                    
+
 Ngroups <- 5
 Ngenes <- 1000
 means <- matrix(rnorm(Ngenes*Ngroups), nrow=Ngenes)
 means[1:900,] <- 0
 colnames(means) <- LETTERS[1:5]
-                                                                    
+
 N <- 100
 g <- sample(LETTERS[1:5], N, replace=TRUE)
 training <- SingleCellExperiment(
@@ -24,7 +24,7 @@ training <- scuttle::logNormCounts(training)
 ##################################################
 ## Mocking up some test data for classification ##
 ##################################################
-                                                                    
+
 N <- 100
 g <- sample(LETTERS[1:5], N, replace=TRUE)
 test <- SingleCellExperiment(

tests/testthat/test-SingleR.R

@@ -18,7 +18,7 @@ test_that("SingleR works with custom gene selection", {
     out <- SingleR(test=test, ref=training, labels=training$label, genes=more.collected)
     tab <- table(out$labels, test$label)
     expect_true(sum(diag(tab))/sum(tab) > 0.95)
-    
+
     # We should get, in this case, the same result with a list of vectors.
     out2 <- SingleR(test=test, ref=training, labels=training$label, genes=collected)
     expect_identical(collected, lapply(metadata(out2)$de.genes, unlist, use.names=FALSE))
@@ -92,7 +92,7 @@ test_that("SingleR handles data.frame inputs", {
     ref1 <- SingleR(test=test, ref=training, labels=training$label)
     set.seed(10)
     ref2 <- SingleR(test=data.frame(logcounts(test)), ref=data.frame(logcounts(training)), labels=training$label)
-    
+
     rownames(ref2) <- NULL # as the data.frame coercion changes the cell's names.
     expect_identical(ref1, ref2)
 

tests/testthat/test-markers.R

@@ -38,7 +38,7 @@ REF <- function(ref, labels, de.n=NULL) {
     collected <- list()
     for (i in ulabels) {
         subcollected <- list()
-   
+
         for (j in ulabels) {
             s <- sort(mat[,i] - mat[,j], decreasing=TRUE)
             s <- s[s>0]
@@ -81,7 +81,7 @@ test_that("getClassicMarkers works with blocking", {
     out2 <- getClassicMarkers(list(logcounts(training), logcounts(training)), 
         list(training$label, training$label))
     expect_identical(out, out2)
-    
+
     # Blocking is robust to training sets that don't have the labels. 
     out3 <- getClassicMarkers(list(logcounts(training), logcounts(training)[,0]), 
         list(training$label, training$label[0]))

tests/testthat/test-prune.R

@@ -10,7 +10,7 @@ test_that("validating per-cell check without finetuning", {
         c(1,1,1,1,1)
     )
     colnames(scores) <- LETTERS[1:5]
-    
+
     results <- DataFrame(scores=I(scores), labels=colnames(scores)[max.col(scores)])
 
     expect_identical(pruneScores(results, min.diff.med=0.05), c(FALSE, FALSE, TRUE, TRUE, TRUE))

tests/testthat/test-recomputed.R

@@ -57,7 +57,7 @@ test_that("combineRecomputedResults matrix fragmentation works as expected", {
         test=test,
         trained=list(train1, train2))
 
-    # Testing that it works upon parallelization.   
+    # Testing that it works upon parallelization.
     combined1x <- combineRecomputedResults(
         results=list(pred1, pred2), 
         test=test,

tests/testthat/test-train.R

@@ -176,7 +176,7 @@ test_that("trainSingleR behaves with multiple references, plus recomputation", {
     training1 <- training2 <- training
     training1 <- training1[sample(nrow(training1)),]
     rownames(training1) <- rownames(training)
-    
+
     ref1 <- trainSingleR(training1, training1$label)
     ref2 <- trainSingleR(training2, training2$label)
     out <- trainSingleR(list(training1, training2), list(training1$label, training2$label))
@@ -213,7 +213,7 @@ test_that("trainSingleR works when 'genes' contains markers outside of the refer
     train.sub <- head(training, 90)
     collected <- SingleR:::.get_genes_by_de(logcounts(training), training$label)
     genes <- unique(unlist(collected))
-    
+
     # Make sure more genes than ref
     expect_false(all(genes %in% row.names(train.sub)))
     expect_error(out <- SingleR::trainSingleR(train.sub, training$label, genes = collected), NA)

vignettes/SingleR.Rmd

@@ -18,7 +18,7 @@ bibliography: ref.bib
 vignette: >
   %\VignetteIndexEntry{Annotating scRNA-seq data}
   %\VignetteEngine{knitr::rmarkdown}
-  %\VignetteEncoding{UTF-8}    
+  %\VignetteEncoding{UTF-8}
 ---
 
 ```{r, echo=FALSE, results="hide", message=FALSE}