[SPARK-2993] [MLLib] colStats (wrapper around MultivariateStatisticalSummary) in Statistics

For both Scala and Python.

The ser/de util functions were moved out of `PythonMLLibAPI` and into their own object to avoid creating the `PythonMLLibAPI` object inside of `MultivariateStatisticalSummarySerialized`, which is then referenced inside of a method in `PythonMLLibAPI`.

`MultivariateStatisticalSummarySerialized` was created to serialize the `Vector` fields in `MultivariateStatisticalSummary`.

Author: Doris Xin <doris.s.xin@gmail.com>

Closes #1911 from dorx/colStats and squashes the following commits:

77b9924 [Doris Xin] developerAPI tag
de9cbbe [Doris Xin] reviewer comments and moved more ser/de
459faba [Doris Xin] colStats in Statistics for both Scala and Python

1 files changed, 8 insertions, 9 deletions

diff --git a/mllib/src/test/scala/org/apache/spark/mllib/api/python/PythonMLLibAPISuite.scala b/mllib/src/test/scala/org/apache/spark/mllib/api/python/PythonMLLibAPISuite.scala
index bd413a80f5..092d67bbc5 100644
--- a/mllib/src/test/scala/org/apache/spark/mllib/api/python/PythonMLLibAPISuite.scala
+++ b/mllib/src/test/scala/org/apache/spark/mllib/api/python/PythonMLLibAPISuite.scala
@@ -23,7 +23,6 @@ import org.apache.spark.mllib.linalg.{Matrices, Vectors}
 import org.apache.spark.mllib.regression.LabeledPoint
 
 class PythonMLLibAPISuite extends FunSuite {
-  val py = new PythonMLLibAPI
 
   test("vector serialization") {
     val vectors = Seq(
@@ -34,8 +33,8 @@ class PythonMLLibAPISuite extends FunSuite {
       Vectors.sparse(1, Array.empty[Int], Array.empty[Double]),
       Vectors.sparse(2, Array(1), Array(-2.0)))
     vectors.foreach { v =>
-      val bytes = py.serializeDoubleVector(v)
-      val u = py.deserializeDoubleVector(bytes)
+      val bytes = SerDe.serializeDoubleVector(v)
+      val u = SerDe.deserializeDoubleVector(bytes)
       assert(u.getClass === v.getClass)
       assert(u === v)
     }
@@ -50,8 +49,8 @@ class PythonMLLibAPISuite extends FunSuite {
       LabeledPoint(1.0, Vectors.sparse(1, Array.empty[Int], Array.empty[Double])),
       LabeledPoint(-0.5, Vectors.sparse(2, Array(1), Array(-2.0))))
     points.foreach { p =>
-      val bytes = py.serializeLabeledPoint(p)
-      val q = py.deserializeLabeledPoint(bytes)
+      val bytes = SerDe.serializeLabeledPoint(p)
+      val q = SerDe.deserializeLabeledPoint(bytes)
       assert(q.label === p.label)
       assert(q.features.getClass === p.features.getClass)
       assert(q.features === p.features)
@@ -60,8 +59,8 @@ class PythonMLLibAPISuite extends FunSuite {
 
   test("double serialization") {
     for (x <- List(123.0, -10.0, 0.0, Double.MaxValue, Double.MinValue, Double.NaN)) {
-      val bytes = py.serializeDouble(x)
-      val deser = py.deserializeDouble(bytes)
+      val bytes = SerDe.serializeDouble(x)
+      val deser = SerDe.deserializeDouble(bytes)
       // We use `equals` here for comparison because we cannot use `==` for NaN
       assert(x.equals(deser))
     }
@@ -70,14 +69,14 @@ class PythonMLLibAPISuite extends FunSuite {
   test("matrix to 2D array") {
     val values = Array[Double](0, 1.2, 3, 4.56, 7, 8)
     val matrix = Matrices.dense(2, 3, values)
-    val arr = py.to2dArray(matrix)
+    val arr = SerDe.to2dArray(matrix)
     val expected = Array(Array[Double](0, 3, 7), Array[Double](1.2, 4.56, 8))
     assert(arr === expected)
 
     // Test conversion for empty matrix
     val empty = Array[Double]()
     val emptyMatrix = Matrices.dense(0, 0, empty)
-    val empty2D = py.to2dArray(emptyMatrix)
+    val empty2D = SerDe.to2dArray(emptyMatrix)
     assert(empty2D === Array[Array[Double]]())
   }
 }