Merge pull request #123 from biowdl/BIOWDL-284

Use compression level 1 in cutadapt.

CHANGELOG.md

@@ -11,6 +11,8 @@ that users understand how the changes affect the new version.
 
 version 1.0.0-dev
 ---------------------------
++ Cutadapt: If the output is a gzipped file, compress with level 1 (instead of default 6).
++ Cutadapt: Fix issues with read2output when using single-end reads.
 + Add feature type, idattr and additional attributes to htseq-count.
 + Added allow-contain option to bowtie.
 + Added a changelog to keep track of changes.

cutadapt.wdl

@@ -5,7 +5,7 @@ task Cutadapt {
         File read1
         File? read2
         String read1output = "cut_r1.fq.gz"
-        String? read2output = if defined(read2) then "cut_r2.fq.gz" else read2
+        String? read2output
         String? format
         Array[String]+? adapter
         Array[String]+? front
@@ -59,15 +59,19 @@ task Cutadapt {
         Boolean? bwa
         Boolean? zeroCap
         Boolean? noZeroCap
-        String? reportPath
-
+        String reportPath = "cutadapt_report.txt"
+        #Int compressionLevel = 1  # This only affects outputs with the .gz suffix.
+        # --compression-level has a bug in 2.4 https://github.com/marcelm/cutadapt/pull/388
+        #~{"--compression-level=" + compressionLevel} \
+        Boolean Z = true  # equal to compressionLevel=1  # Fixme: replace once upstream is fixed.
         Int cores = 1
         Int memory = 16  # FIXME: Insane memory. Double-check if needed.
-        String dockerImage = "quay.io/biocontainers/cutadapt:2.3--py36h14c3975_0"
+        String dockerImage = "quay.io/biocontainers/cutadapt:2.4--py37h14c3975_0"
     }
 
-    String read2outputArg = if (defined(read2output))
-        then "mkdir -p $(dirname " + read2output + ")"
+    String realRead2output = select_first([read2output, "cut_r2.fq.gz"])
+    String read2outputArg = if (defined(read2))
+        then "mkdir -p $(dirname " + realRead2output + ")"
         else ""
 
     # FIXME: This crappy overengineering can be removed once cromwell can handle subworkflow inputs correctly.
@@ -95,13 +99,14 @@ task Cutadapt {
         ~{read2outputArg}
         cutadapt \
         ~{"--cores=" + cores} \
+        ~{true="-Z" false="" Z} \
         ~{true="-a" false="" defined(adapterForward)} ~{sep=" -a " adapterForward} \
         ~{true="-A" false="" defined(adapterReverse)} ~{sep=" -A " adapterReverse} \
         ~{true="-g" false="" defined(front)} ~{sep=" -g " front} \
         ~{true="-G" false="" defined(frontRead2)} ~{sep=" -G " frontRead2} \
         ~{true="-b" false="" defined(anywhereForward)} ~{sep=" -b " anywhereForward} \
         ~{true="-B" false="" defined(anywhereReverse)} ~{sep=" -B " anywhereReverse} \
-        --output ~{read1output} ~{"--paired-output " + read2output} \
+        --output ~{read1output} ~{if defined(read2) then "-p " + realRead2output else ""} \
         ~{"--to-short-output " + tooShortOutputPath} \
         ~{"--to-short-paired-output " + tooShortPairedOutputPath} \
         ~{"--to-long-output " + tooLongOutputPath} \
@@ -151,9 +156,7 @@ task Cutadapt {
     output{
         File cutRead1 = read1output
         File? cutRead2 = read2output
-        File report = if defined(reportPath)
-            then select_first([reportPath])
-            else stdout()
+        File report = reportPath
         File? tooLongOutput=tooLongOutputPath
         File? tooShortOutput=tooShortOutputPath
         File? untrimmedOutput=untrimmedOutputPath