outputDir - only applies for cohort.list im DaraProcessingPipeline.scala

Johan_Dahlberg

Looking at the DataProcessingPipeline script I noticed that the "outputDir" is only applied to the cohort list file, but not to the actually processed files, which seems a bit inconsistent with the docs, which say "Output path for the processed BAM files". Here is my fix for this:

diff --git a/public/scala/qscript/org/broadinstitute/sting/queue/qscripts/DataProcessingPipeline.scala b/public/scala/qscript/org/broadinstitute/sting/queue/qscripts/DataProcessingPipeline.scala
index 56f6460..6dd84b5 100755
--- a/public/scala/qscript/org/broadinstitute/sting/queue/qscripts/DataProcessingPipeline.scala
+++ b/public/scala/qscript/org/broadinstitute/sting/queue/qscripts/DataProcessingPipeline.scala
@@ -249,8 +249,12 @@ class DataProcessingPipeline extends QScript {
     // put each sample through the pipeline
     for ((sample, bamList) <- sampleBAMFiles) {

-      // BAM files generated by the pipeline
-      val bam        = new File(qscript.projectName + "." + sample + ".bam")
+      // BAM files generated by the pipeline      
+      val bam        = if(outputDir.isEmpty()) 
+                                      new File(qscript.projectName + "." + sample + ".bam")
+                                  else
+                                      new File(outputDir + qscript.projectName + "." + sample + ".bam")
+      
       val cleanedBam = swapExt(bam, ".bam", ".clean.bam")
       val dedupedBam = swapExt(bam, ".bam", ".clean.dedup.bam")
       val recalBam   = swapExt(bam, ".bam", ".clean.dedup.recal.bam")
@@ -292,6 +296,15 @@ class DataProcessingPipeline extends QScript {
     add(writeList(cohortList, cohortFile))
   }

+  // Override the normal swapExt metod by adding the outputDir to the file path by default if it is defined.
+  override
+  def swapExt(file: File, oldExtension: String, newExtension: String) = {
+      if(outputDir.isEmpty())
+         new File(file.getName.stripSuffix(oldExtension) + newExtension)
+      else
+          swapExt(outputDir, file, oldExtension, newExtension);
+  }      
+      


   /****************************************************************************

This of course, puts all the files in the directory specified by outputDir, but I think that this seems reasonable than putting them in the execution directory of the script.

Carneiro

Thanks, do you want to send me your patch? I can add it to the codebase.

Here are the instructions how to do it: http://gatk.vanillaforums.com/discussion/1267/how-to-submit-patches-to-the-gatk-codebase

Johan_Dahlberg

Sure. I've emailed it to you.