VK music parsing

Main.scala

import java.util.regex.Pattern

import net.logstash.logback.encoder.org.apache.commons.lang.StringEscapeUtils

import scala.io.Source
import scala.util.control.NonFatal

object Main {

  def main(args: Array[String]): Unit = 
    args.toList match {
      case input :: output :: Nil=>
        println(s"Input: [$input]. Output: [$output]")
        process(input, output)
        
      case _ =>
        sys.error("Invalid args format. Please, pass 2 args: 1) input file path 2) output file path")
    }
  
  def process(input: String, output: String): Unit = {
    val regex = """(,\")|(\",)"""

    val dataAudioRegex = Pattern.compile("data-audio=\\\"(?<text>.*)\\\"")

    import java.io._

    val pw = new PrintWriter(new File(output))

    try {
      Source.fromFile(new File(input))
        .getLines()
        .filter { line => dataAudioRegex.matcher(line).find() }
        .map { line =>
          val matcher = dataAudioRegex.matcher(line)
          matcher.find()


          val text = StringEscapeUtils.unescapeHtml(matcher.group("text"))
          val result = text.split(regex)
          try {
            (result(3).drop(1), result(2).drop(1))
          } catch {
            case NonFatal(e) =>
              ("error", line)
          }
        }
        .toList
        .sortBy { case (band, _) => band }
        .foreach { case (band, song) =>
          pw.write(band + " - " + song)
          pw.write("\r\n")
        }
    } finally {
      pw.close()
    }
  }
}