Added a skip option in merge otu maps tool

biopet-tools-extensions/src/main/scala/nl/lumc/sasc/biopet/extensions/tools/MergeOtuMaps.scala

@@ -31,9 +31,14 @@ class MergeOtuMaps(val root: Configurable) extends ToolCommandFunction {
   @Output(doc = "Output", shortName = "output", required = true)
   var output: File = _
 
+  var skipPrefix: List[String] = config("skip_prefix", default = Nil)
+
   override def defaultCoreMemory = 6.0
 
-  override def cmdLine = super.cmdLine + repeat("-I", input) + required("-o", output)
+  override def cmdLine = super.cmdLine +
+    repeat("-I", input) +
+    required("-o", output) +
+    repeat("-p", skipPrefix)
 
 }
 

biopet-tools/src/main/scala/nl/lumc/sasc/biopet/tools/MergeOtuMaps.scala

@@ -24,7 +24,9 @@ import scala.io.Source
  * Created by pjvan_thof on 12/18/15.
  */
 object MergeOtuMaps extends ToolCommand {
-  case class Args(inputFiles: List[File] = Nil, outputFile: File = null) extends AbstractArgs
+  case class Args(inputFiles: List[File] = Nil,
+                  outputFile: File = null,
+                  skipPrefix: List[String] = Nil) extends AbstractArgs
 
   class OptParser extends AbstractOptParser {
     opt[File]('I', "input") minOccurs 2 required () unbounded () valueName "<file>" action { (x, c) =>
@@ -33,6 +35,9 @@ object MergeOtuMaps extends ToolCommand {
     opt[File]('o', "output") required () unbounded () maxOccurs 1 valueName "<file>" action { (x, c) =>
       c.copy(outputFile = x)
     }
+    opt[String]('p', "skipPrefix") unbounded () valueName "<text>" action { (x, c) =>
+      c.copy(skipPrefix = x :: c.skipPrefix)
+    }
   }
 
   /**
@@ -40,23 +45,24 @@ object MergeOtuMaps extends ToolCommand {
    */
   def main(args: Array[String]): Unit = {
     val argsParser = new OptParser
-    val commandArgs: Args = argsParser.parse(args, Args()) getOrElse (throw new IllegalArgumentException)
+    val cmdArgs: Args = argsParser.parse(args, Args()) getOrElse (throw new IllegalArgumentException)
 
-    var map: Map[Long, String] = Map()
+    var map: Map[String, String] = Map()
 
-    for (inputFile <- commandArgs.inputFiles) {
+    for (inputFile <- cmdArgs.inputFiles) {
       logger.info(s"Start reading $inputFile")
       val reader = Source.fromFile(inputFile)
       reader.getLines().foreach { line =>
         val values = line.split("\t", 2)
-        val key = values.head.toLong
-        map += key -> (line.stripPrefix(s"$key") + map.getOrElse(key, ""))
+        val key = values.head
+        if (!cmdArgs.skipPrefix.exists(key.startsWith))
+          map += key -> (line.stripPrefix(s"$key") + map.getOrElse(key, ""))
       }
       reader.close()
     }
 
-    logger.info(s"Start writing to ${commandArgs.outputFile}")
-    val writer = new PrintWriter(commandArgs.outputFile)
+    logger.info(s"Start writing to ${cmdArgs.outputFile}")
+    val writer = new PrintWriter(cmdArgs.outputFile)
     map.foreach { case (key, list) => writer.println(key + list) }
     writer.close()
   }

gears/src/main/scala/nl/lumc/sasc/biopet/pipelines/gears/Gears.scala

@@ -37,6 +37,8 @@ class Gears(val root: Configurable) extends QScript with MultiSampleQScript { qs
     Some(gearsReport)
   }
 
+  override def defaults = Map("mergeotumaps" -> Map("skip_prefix" -> "New."))
+
   override def fixedValues = Map("gearssingle" -> Map("skip_flexiprep" -> true))
 
   /** Init for pipeline */

gears/src/main/scala/nl/lumc/sasc/biopet/pipelines/gears/GearsReport.scala

@@ -43,7 +43,6 @@ object GearsReport extends MultisampleReportBuilder {
     val qiimeOpenOtuTable = summary.getValue("gears", "files", "pipeline", "qiime_open_otu_table", "path")
       .map(x => new File(x.toString))
 
-
     ReportPage(
       (if (krakenExecuted) List("Kraken analysis" -> ReportPage(List(), List(
         "Krona plot" -> ReportSection("/nl/lumc/sasc/biopet/pipelines/gears/krakenKrona.ssp"