Small cleanups after original tree API PR

Author: jkbradley

examples/src/main/scala/org/apache/spark/examples/ml/DecisionTreeExample.scala

@@ -70,7 +70,7 @@ object DecisionTreeExample {
     val parser = new OptionParser[Params]("DecisionTreeExample") {
       head("DecisionTreeExample: an example decision tree app.")
       opt[String]("algo")
-        .text(s"algorithm (Classification, Regression), default: ${defaultParams.algo}")
+        .text(s"algorithm (classification, regression), default: ${defaultParams.algo}")
         .action((x, c) => c.copy(algo = x))
       opt[Int]("maxDepth")
         .text(s"max depth of the tree, default: ${defaultParams.maxDepth}")
@@ -221,18 +221,23 @@ object DecisionTreeExample {
     // (1) For classification, re-index classes.
     val labelColName = if (algo == "classification") "indexedLabel" else "label"
     if (algo == "classification") {
-      val labelIndexer = new StringIndexer().setInputCol("labelString").setOutputCol(labelColName)
+      val labelIndexer = new StringIndexer()
+        .setInputCol("labelString")
+        .setOutputCol(labelColName)
       stages += labelIndexer
     }
     // (2) Identify categorical features using VectorIndexer.
     //     Features with more than maxCategories values will be treated as continuous.
-    val featuresIndexer = new VectorIndexer().setInputCol("features")
-      .setOutputCol("indexedFeatures").setMaxCategories(10)
+    val featuresIndexer = new VectorIndexer()
+      .setInputCol("features")
+      .setOutputCol("indexedFeatures")
+      .setMaxCategories(10)
     stages += featuresIndexer
     // (3) Learn DecisionTree
     val dt = algo match {
       case "classification" =>
-        new DecisionTreeClassifier().setFeaturesCol("indexedFeatures")
+        new DecisionTreeClassifier()
+          .setFeaturesCol("indexedFeatures")
           .setLabelCol(labelColName)
           .setMaxDepth(params.maxDepth)
           .setMaxBins(params.maxBins)
@@ -241,7 +246,8 @@ object DecisionTreeExample {
           .setCacheNodeIds(params.cacheNodeIds)
           .setCheckpointInterval(params.checkpointInterval)
       case "regression" =>
-        new DecisionTreeRegressor().setFeaturesCol("indexedFeatures")
+        new DecisionTreeRegressor()
+          .setFeaturesCol("indexedFeatures")
           .setLabelCol(labelColName)
           .setMaxDepth(params.maxDepth)
           .setMaxBins(params.maxBins)

mllib/src/main/scala/org/apache/spark/ml/impl/tree/treeParams.scala

@@ -283,7 +283,7 @@ private[ml] trait TreeRegressorParams extends Params {
   def getImpurity: String = getOrDefault(impurity)
 
   /** Convert new impurity to old impurity. */
-  protected def getOldImpurity: OldImpurity = {
+  private[ml] def getOldImpurity: OldImpurity = {
     getImpurity match {
       case "variance" => OldVariance
       case _ =>

mllib/src/main/scala/org/apache/spark/ml/tree/Split.scala

@@ -38,7 +38,7 @@ sealed trait Split extends Serializable {
   private[tree] def toOld: OldSplit
 }
 
-private[ml] object Split {
+private[tree] object Split {
 
   def fromOld(oldSplit: OldSplit, categoricalFeatures: Map[Int, Int]): Split = {
     oldSplit.featureType match {
@@ -58,7 +58,7 @@ private[ml] object Split {
  *                        left. Otherwise, it goes right.
  * @param numCategories  Number of categories for this feature.
  */
-final class CategoricalSplit(
+final class CategoricalSplit private[ml] (
     override val featureIndex: Int,
     leftCategories: Array[Double],
     private val numCategories: Int)
@@ -130,7 +130,8 @@ final class CategoricalSplit(
  * @param threshold  If the feature value is <= this threshold, then the split goes left.
  *                    Otherwise, it goes right.
  */
-final class ContinuousSplit(override val featureIndex: Int, val threshold: Double) extends Split {
+final class ContinuousSplit private[ml] (override val featureIndex: Int, val threshold: Double)
+  extends Split {
 
   override private[ml] def shouldGoLeft(features: Vector): Boolean = {
     features(featureIndex) <= threshold