guess_species_from_gtf.scalah

#!/bin/sh
exec scalas "$0" "$@"
!#

/** Work in progress: Guess the species from a gtf file. By name first and then by file content. A more simplistic approach is already implementd in dge_workflow/dge_utils.sh
  */

import java.io.File

import scala.io.Source

// http://alvinalexander.com/scala/scala-shell-script-command-line-arguments-args
val gtfFile = args(1)
//val gtfFile="mm10_igenomes_pc.gtf"

val pattern = "mm10|mm9|h19|zv9".r
val genomeByName = pattern.findFirstIn(gtfFile)

if (genomeByName.isEmpty) {
  System.exit(1)
}

def guessFromContent(gtfFile: File): Option[String] = {
  //  Source.fromString(s"grep ENSMUSG $gtfFile | "!!).getLines().hasNext
  //  Bash.evalCapture(s"grep ENSMUSG $gtfFile | wc -l")

  if (Source.fromFile(gtfFile).getLines().exists(_.contains("ENSMUSG"))) return Some("mouse")
  if (Source.fromFile(gtfFile).getLines().exists(_.contains("ENSCAFG"))) return Some("dog")

  None
}


genomeByName.get match {
  case "mm9" =>
}
println(genomeByName)