Merge branch 'fix-BIOPET-483' into 'develop'

Adding bgzipped fasta



See merge request !498

biopet-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/Bgzip.scala

@@ -17,21 +17,28 @@ package nl.lumc.sasc.biopet.extensions
 import java.io.File
 
 import nl.lumc.sasc.biopet.core.BiopetCommandLineFunction
+import nl.lumc.sasc.biopet.utils.Logging
 import nl.lumc.sasc.biopet.utils.config.Configurable
 import org.broadinstitute.gatk.utils.commandline.{ Input, Output }
 
 /** Wrapper for the bgzip command */
 class Bgzip(val root: Configurable) extends BiopetCommandLineFunction {
 
-  @Input(doc = "Input files", required = true)
+  @Input(doc = "Input files", required = false)
   var input: List[File] = Nil
 
-  @Output(doc = "Compressed output file", required = true)
+  @Output(doc = "Compressed output file", required = false)
   var output: File = null
 
   var f: Boolean = config("f", default = false)
   executable = config("exe", default = "bgzip", freeVar = false)
 
+  override def beforeGraph(): Unit = {
+    super.beforeGraph()
+    if (input.isEmpty && !inputAsStdin) Logging.addError("Input is missing for Bgzip")
+    if (output == null && !outputAsStsout) Logging.addError("Output is missing for Bgzip")
+  }
+
   def cmdLine = required(executable) +
     conditional(f, "-f") +
     " -c " + repeat(input) +

biopet-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/samtools/SamtoolsFaidx.scala

@@ -25,13 +25,14 @@ class SamtoolsFaidx(val root: Configurable) extends Samtools {
   var input: File = _
 
   @Output(doc = "output File")
-  private var _output: File = _
+  private var _output: List[File] = Nil
 
   def output = _output
 
   override def beforeGraph: Unit = {
     super.beforeGraph
-    _output = new File(input.getParentFile, input.getName + ".fai")
+    _output = List(new File(input.getAbsolutePath + ".fai"))
+    if (input.getName.endsWith(".gz")) _output :+= new File(input.getAbsolutePath + ".gzi")
   }
 
   /** Returns command to execute */
@@ -42,7 +43,8 @@ object SamtoolsFaidx {
   def apply(root: Configurable, input: File): SamtoolsFaidx = {
     val faidx = new SamtoolsFaidx(root)
     faidx.input = input
-    faidx._output = new File(input.getParentFile, input.getName + ".fai")
+    faidx._output = List(new File(input.getAbsolutePath + ".fai"))
+    if (input.getName.endsWith(".gz")) faidx._output :+= new File(input.getAbsolutePath + ".gzi")
     faidx
   }
 }
\ No newline at end of file

generate-indexes/src/main/scala/nl/lumc/sasc/biopet/pipelines/generateindexes/GenerateIndexes.scala

@@ -2,7 +2,7 @@ package nl.lumc.sasc.biopet.pipelines.generateindexes
 
 import nl.lumc.sasc.biopet.core.{ BiopetQScript, PipelineCommand }
 import nl.lumc.sasc.biopet.extensions.bowtie.{ Bowtie2Build, BowtieBuild }
-import nl.lumc.sasc.biopet.extensions.{ Ln, Star }
+import nl.lumc.sasc.biopet.extensions.{ Bgzip, Ln, Star }
 import nl.lumc.sasc.biopet.extensions.bwa.BwaIndex
 import nl.lumc.sasc.biopet.extensions.gmap.GmapBuild
 import nl.lumc.sasc.biopet.extensions.hisat.Hisat2Build
@@ -44,9 +44,13 @@ class GenerateIndexes(val root: Configurable) extends QScript with BiopetQScript
     var outputConfig: Map[String, Any] = Map("reference_fasta" -> fastaFile)
     val configDeps = new ListBuffer[File]()
 
+    val bgzipFasta = new File(fastaFile.getAbsolutePath + ".gz")
+    add(fastaFile :<: new Bgzip(this) > bgzipFasta)
+    add(SamtoolsFaidx(this, bgzipFasta))
+
     val faidx = SamtoolsFaidx(this, fastaFile)
     add(faidx)
-    configDeps += faidx.output
+    faidx.output.foreach(configDeps += _)
 
     val createDict = new CreateSequenceDictionary(this)
     createDict.reference = fastaFile
@@ -59,10 +63,10 @@ class GenerateIndexes(val root: Configurable) extends QScript with BiopetQScript
 
     def createLinks(dir: File): File = {
       val newFastaFile = new File(dir, fastaFile.getName)
-      val newFai = new File(dir, faidx.output.getName)
+      val newFai = new File(dir, faidx.output.head.getName)
       val newDict = new File(dir, createDict.output.getName)
 
-      add(Ln(this, faidx.output, newFai))
+      add(Ln(this, faidx.output.head, newFai))
       add(Ln(this, createDict.output, newDict))
       val lnFasta = Ln(this, fastaFile, newFastaFile)
       lnFasta.deps ++= List(newFai, newDict)