change style and use alternative builtin methods

mllib/src/main/scala/org/apache/spark/mllib/evaluation/RankingMetrics.scala

@@ -38,7 +38,7 @@ class RankingMetrics(predictionAndLabels: RDD[(Array[Double], Array[Double])]) {
   lazy val precAtK: RDD[Array[Double]] = predictionAndLabels.map {case (pred, lab)=>
     val labSet : Set[Double] = lab.toSet
     val n = pred.length
-    val topkPrec = Array.fill[Double](n)(.0)
+    val topkPrec = Array.fill[Double](n)(0.0)
     var (i, cnt) = (0, 0)
 
     while (i < n) {
@@ -56,7 +56,7 @@ class RankingMetrics(predictionAndLabels: RDD[(Array[Double], Array[Double])]) {
    */
   lazy val avePrec: RDD[Double] = predictionAndLabels.map {case (pred, lab) =>
     val labSet: Set[Double] = lab.toSet
-    var (i, cnt, precSum) = (0, 0, .0)
+    var (i, cnt, precSum) = (0, 0, 0.0)
     val n = pred.length
 
     while (i < n) {
@@ -72,15 +72,15 @@ class RankingMetrics(predictionAndLabels: RDD[(Array[Double], Array[Double])]) {
   /**
    * Returns the mean average precision (MAP) of all the queries
    */
-  lazy val meanAvePrec: Double = computeMean(avePrec)
+  lazy val meanAvePrec: Double = avePrec.mean
 
   /**
    * Returns the normalized discounted cumulative gain for each query
    */
   lazy val ndcg: RDD[Double] = predictionAndLabels.map {case (pred, lab) =>
     val labSet = lab.toSet
     val n = math.min(pred.length, labSet.size)
-    var (maxDcg, dcg, i) = (.0, .0, 0)
+    var (maxDcg, dcg, i) = (0.0, 0.0, 0)
     while (i < n) {
       /* Calculate 1/log2(i + 2) */
       val gain = 1.0 / (math.log(i + 2) / math.log(2))
@@ -96,13 +96,5 @@ class RankingMetrics(predictionAndLabels: RDD[(Array[Double], Array[Double])]) {
   /**
    * Returns the mean NDCG of all the queries
    */
-  lazy val meanNdcg: Double = computeMean(ndcg)
-
-  private def computeMean(data: RDD[Double]): Double = {
-    val stat = data.aggregate((.0, 0))(
-      seqOp = (c, v) => (c, v) match {case ((sum, cnt), a) => (sum + a, cnt + 1)},
-      combOp = (c1, c2) => (c1, c2) match {case (x, y) => (x._1 + y._1, x._2 + y._2)}
-    )
-    stat._1 / stat._2
-  }
+  lazy val meanNdcg: Double = ndcg.mean
 }

mllib/src/test/scala/org/apache/spark/mllib/evaluation/RankingMetricsSuite.scala

@@ -18,7 +18,7 @@
 package org.apache.spark.mllib.evaluation
 
 import org.scalatest.FunSuite
-
+import org.apache.spark.mllib.util.TestingUtils._
 import org.apache.spark.mllib.util.LocalSparkContext
 
 class RankingMetricsSuite extends FunSuite with LocalSparkContext {
@@ -28,7 +28,7 @@ class RankingMetricsSuite extends FunSuite with LocalSparkContext {
         (Array[Double](1, 6, 2, 7, 8, 3, 9, 10, 4, 5), Array[Double](1, 2, 3, 4, 5)),
         (Array[Double](4, 1, 5, 6, 2, 7, 3, 8, 9, 10), Array[Double](1, 2, 3))
       ), 2)
-    val eps: Double = 1e-5
+    val eps: Double = 1E-5
 
     val metrics = new RankingMetrics(predictionAndLabels)
     val precAtK = metrics.precAtK.collect()
@@ -37,13 +37,13 @@ class RankingMetricsSuite extends FunSuite with LocalSparkContext {
     val ndcg = metrics.ndcg.collect()
     val aveNdcg = metrics.meanNdcg
 
-    assert(math.abs(precAtK(0)(4) - 0.4) < eps)
-    assert(math.abs(precAtK(1)(6) - 3.0/7) < eps)
-    assert(math.abs(avePrec(0) - 0.622222) < eps)
-    assert(math.abs(avePrec(1) - 0.442857) < eps)
-    assert(math.abs(map - 0.532539) < eps)
-    assert(math.abs(ndcg(0) - 0.508740) < eps)
-    assert(math.abs(ndcg(1) - 0.296082) < eps)
-    assert(math.abs(aveNdcg - 0.402411) < eps)
+    assert(precAtK(0)(4) ~== 0.4 absTol eps)
+    assert(precAtK(1)(6) ~== 3.0/7 absTol eps)
+    assert(avePrec(0) ~== 0.622222 absTol eps)
+    assert(avePrec(1) ~== 0.442857 absTol eps)
+    assert(map ~== 0.532539 absTol eps)
+    assert(ndcg(0) ~== 0.508740 absTol eps)
+    assert(ndcg(1) ~== 0.296082 absTol eps)
+    assert(aveNdcg ~== 0.402411 absTol eps)
   }
 }