ReadNG27-fold.scala

// Scala 2.7.7, immutable map, uses foldLeft to loop over files/words
import java.io._
import scala.io._

def time(f: => Unit) = {
  val t1 = System.currentTimeMillis
  f
  ((System.currentTimeMillis - t1)/1000.0)
}
 
def processNewsgroups(rootDir: File): Unit = {
  def write(map: Map[String, Int], file: String)(sort: (Tuple2[String, Int], Tuple2[String, Int]) => Boolean) {
    using(new PrintWriter(new FileWriter(file))) { out =>
      map.toList.sort(sort).foreach { pair => out.println(pair._1 + "\t" + pair._2) }
    }
  }

  def using[Closeable <: {def close(): Unit}, B](closeable: Closeable)(getB: Closeable => B): B =
    try {
      getB(closeable)
    } finally {
      closeable.close()
    }

  implicit def file2String(file: File): String = Source.fromFile(file, "ISO-8859-1").getLines.mkString("\n")
 
  var counts = (Map.empty[String, Int].withDefaultValue(0) /:
      rootDir.listFiles.filter(_.isDirectory).flatMap(_.listFiles).flatMap(_.toLowerCase.split("""\W+"""))) (
          (c,word) => c + (word -> (1 + c(word))) )
  write(counts, "counts-descreasing-scala.txt") {_._2 > _._2}
  write(counts, "counts-alphabetical-scala.txt") {_._1 < _._1}
}

time(processNewsgroups(new File("./20_newsgroups")))

ReadNG27.scala

// Scala 2.7.7, mutable map, loops over files/words with foreach
import java.io._
import scala.io._
import scala.collection.mutable._

def time(f: => Unit) = {
  val t1 = System.currentTimeMillis
  f
  ((System.currentTimeMillis - t1)/1000.0)
}
 
def processNewsgroups(rootDir: File): Unit = {
  def write(map: Map[String, Int], file: String)(sort: (Tuple2[String, Int], Tuple2[String, Int]) => Boolean) {
    using(new PrintWriter(new FileWriter(file))) { out =>
      map.toList.sort(sort).foreach { pair => out.println(pair._1 + "\t" + pair._2) }
    }
  }

  def using[Closeable <: {def close(): Unit}, B](closeable: Closeable)(getB: Closeable => B): B =
    try {
      getB(closeable)
    } finally {
      closeable.close()
    }

  implicit def file2String(file: File): String = Source.fromFile(file, "ISO-8859-1").getLines.mkString("\n")
 
  val counts = Map.empty[String, Int]
  rootDir.listFiles.filter(_.isDirectory).flatMap(_.listFiles).foreach { file: File =>
    val s = file2String(file)
    s.toLowerCase.split("""\W+""").foreach { word =>
	  counts(word) = if (counts.contains(word)) counts(word) + 1 else 1
	}
  }
  write(counts, "counts-descreasing-scala.txt") {_._2 > _._2}
  write(counts, "counts-alphabetical-scala.txt") {_._1 < _._1}
}

time(processNewsgroups(new File("./20_newsgroups")))

ReadNG28-fold.scala

// Scala 2.8, immutable map, uses foldLeft to loop over files/words
import java.io._
import scala.io._

def time(f: => Unit) = {
  val t1 = System.currentTimeMillis
  f
  ((System.currentTimeMillis - t1)/1000.0)
}
 
def processNewsgroups(rootDir: File): Unit = {
  def write(map: Map[String, Int], file: String)(sort: (Tuple2[String, Int], Tuple2[String, Int]) => Boolean) {
    using(new PrintWriter(new FileWriter(file))) { out =>
      map.toList.sort(sort).foreach { pair => out.println(pair._1 + "\t" + pair._2) }
    }
  }

  def using[Closeable <: {def close(): Unit}, B](closeable: Closeable)(getB: Closeable => B): B =
    try {
      getB(closeable)
    } finally {
      closeable.close()
    }

  implicit def file2String(file: File): String = Source.fromFile(file, 32768)(Codec.ISO8859).getLines().mkString("\n")
 
  var counts = (Map.empty[String, Int].withDefaultValue(0) /:
      rootDir.listFiles.filter(_.isDirectory).flatMap(_.listFiles).flatMap(_.toLowerCase.split("""\W+"""))) (
          (c,word) => c + (word -> (1 + c(word))) )
  write(counts, "counts-descreasing-scala.txt") {_._2 > _._2}
  write(counts, "counts-alphabetical-scala.txt") {_._1 < _._1}
}

time(processNewsgroups(new File("./20_newsgroups")))

ReadNG28.scala

// Scala 2.8, mutable map, loops over files/words with foreach
import java.io._
import scala.io._
import scala.collection.mutable._

def time(f: => Unit) = {
  val t1 = System.currentTimeMillis
  f
  ((System.currentTimeMillis - t1)/1000.0)
}
 
def processNewsgroups(rootDir: File): Unit = {
  def write(map: Map[String, Int], file: String)(sort: (Tuple2[String, Int], Tuple2[String, Int]) => Boolean) {
    using(new PrintWriter(new FileWriter(file))) { out =>
      map.toList.sort(sort).foreach { pair => out.println(pair._1 + "\t" + pair._2) }
    }
  }

  def using[Closeable <: {def close(): Unit}, B](closeable: Closeable)(getB: Closeable => B): B =
    try {
      getB(closeable)
    } finally {
      closeable.close()
    }

  implicit def file2String(file: File): String = Source.fromFile(file)(Codec.ISO8859).getLines().mkString("\n")
 
  val counts = Map.empty[String, Int]
  rootDir.listFiles.filter(_.isDirectory).flatMap(_.listFiles).foreach { file: File =>
    val s = file2String(file)
    s.toLowerCase.split("""\W+""").foreach { word =>
	  counts(word) = if (counts.contains(word)) counts(word) + 1 else 1
	}
  }
  write(counts, "counts-descreasing-scala.txt") {_._2 > _._2}
  write(counts, "counts-alphabetical-scala.txt") {_._1 < _._1}
}

time(processNewsgroups(new File("./20_newsgroups")))