Refactored data16S() by defining internal functions.

internals.R

+## High level Helper functions
+
+.import_biom <- function(input) {
+  ## Select appropriate import_function
+  import_function <- switch(input$biomFormat,
+                            "std"   = import_biom,
+                            "frogs" = import_frogs)
+  ## Import
+  return(import_function(input$fileBiom$datapath, ## biom file
+                         input$fileTree$datapath  ## tree file
+  ))
+}
+
+.import_sample_data <- function(input, physeq) {
+  ## Unhappy path
+  if (is.null(input$fileMeta)) {
+    return(data.frame(SampleID = sample_names(physeq) , row.names = sample_names(physeq)))
+  }
+
+  ## Happy path: excel version
+  if (input$CSVsep == "excel") {
+    ## For efficiency and consistency, maybe replace with readr::read_xl
+    sdf <- RcmdrMisc::readXL(input$fileMeta$datapath,
+                             rownames = TRUE,
+                             header = TRUE)
+    sdf$SampleID <- rownames(sdf)
+    return(sdf)
+  }
+
+  ## Happy path: csv version
+  sdf <- read.csv(
+    input$fileMeta$datapath,
+    header = TRUE,
+    sep = input$CSVsep,
+    row.names = 1,
+    na.strings = NA
+  )
+  sdf$SampleID <- rownames(sdf)
+  return(sdf)
+}
+
+.format_tax_table <- function(tdf) {
+  ## explicit rank names
+  colnames(tdf) <- c("Kingdom", "Phylum", "Class", "Order",
+                     "Family", "Genus", "Species", "Strain")[1:ncol(tdf)]
+  ## Replace unknown by NA
+  tdf[grep("unknown ", tdf)] <- NA
+  #tdf[grep("Unclassified", tdf)] <- NA
+  return(tdf)
+}
\ No newline at end of file

server.R

@@ -49,60 +49,28 @@ shinyServer
     "https://raw.githubusercontent.com/mahendra-mariadassou/phyloseq-extended/master/R/load-extra-functions.R"
   })
 
+  source("internals.R")
+
   data16S <- reactive({
+
+    ## BIOM input
     if (input$dataset == "input")
     {
+      ## Unhappy path
       if (is.null(input$fileBiom))
-      {
         return()
-      }
-      if (input$biomFormat == "std")
-      {
-        d <- import_biom(
-          BIOMfilename = input$fileBiom$datapath,
-          treefilename = input$fileTree$datapath# ,
-          # refseqfilename = input$fileSeq$datapath
-        )
-      } else if (input$biomFormat == "frogs") {
-        d <- import_frogs(
-          biom = input$fileBiom$datapath,
-          treefilename = input$fileTree$datapath# ,
-          # refseqfilename = input$fileSeq$datapath
-        )
-      }
 
-      colnames(tax_table(d)) <-
-        c("Kingdom",
-          "Phylum",
-          "Class",
-          "Order",
-          "Family",
-          "Genus",
-          "Species",
-          "Strain")[1:length(rank_names(d))]
-      tax_table(d)[grep("unknown ", tax_table(d))] <- NA
-      #tax_table(d)[grep("Unclassified", tax_table(d))] <- NA
-      if (!is.null(input$fileMeta)) {
-        if (input$CSVsep == "excel") {
-          sample_data(d) <-
-            RcmdrMisc::readXL(input$fileMeta$datapath,
-                              rownames = TRUE,
-                              header = TRUE)
-        } else {
-          sample_data(d) <- read.csv(
-            input$fileMeta$datapath,
-            header = TRUE,
-            sep = input$CSVsep,
-            row.names = 1,
-            na.strings = NA
-          )
-          sample_data(d)$SampleID <- rownames(sample_data(d))
-        }
-      } else {
-        n <- data.frame(sample_names(d) , row.names = sample_names(d))
-        names(n) <- "SampleID"
-        sample_data(d) <- n
-      }
+      ## Happy path
+      ## Import biom
+      d <- .import_biom(input)
+
+      ## Format tax table
+      tax_table(d) <- .format_tax_table(tax_table(d))
+
+      ## import metadata and store it in phyloseq object
+      sample_data(d) <- .import_sample_data(input, d)
+
+      ## Rarefy data
       if (input$rareData) {
         d <- rarefy_even_depth(
           d,
@@ -114,21 +82,19 @@ shinyServer
       return(d)
     }
 
+    ## Rdata input
     if (input$dataset == "rdata")
     {
-      if (is.null(input$fileRData))
-      {
-        return()
-      }
-      load(input$fileRData$datapath)
-      if (exists("data"))
-      {
+      ## Happy path
+      if (!is.null(input$fileRData))
+        load(input$fileRData$datapath)
+      if (exists("data") && class(data) == "phyloseq")
         return(data)
-      } else {
-        return()
-      }
+      ## Unhappy path
+      return()
     }
 
+    ## Default case
     load("demo/demo.RData")
     return(get(input$dataset))
 
@@ -174,7 +140,7 @@ shinyServer
     validate(
       need(
         data16S(),
-        "Firstly, you should select a demo dataset or upload an abundance BIOM file.\nFor example, with Galaxy, a BIOM file can be obtained at the end of FROGS workflow with the 'FROGS BIOM to std BIOM' tool"
+        "Firstly, you should select a demo dataset or upload an abundance BIOM file.\nFor example, with Galaxy, a BIOM file can be obtained at the end of FROGS workflow with the 'FROGS BIOM to std BIOM' tool.\n Make sure that the phyloseq object in the RData file is called 'data'."
       )
     )
     data16S()