Github user mengxr commented on a diff in the pull request:
https://github.com/apache/spark/pull/1520#discussion_r15389694
--- Diff:
mllib/src/test/scala/org/apache/spark/mllib/random/RandomRDDGeneratorsSuite.scala
---
@@ -0,0 +1,171 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.mllib.random
+
+import scala.collection.mutable.ArrayBuffer
+
+import org.scalatest.FunSuite
+
+import org.apache.spark.SparkContext._
+import org.apache.spark.mllib.linalg.Vector
+import org.apache.spark.mllib.rdd.{RandomRDDPartition, RandomRDD}
+import org.apache.spark.mllib.util.LocalSparkContext
+import org.apache.spark.rdd.RDD
+import org.apache.spark.util.StatCounter
+
+/**
+ * Note: avoid including APIs that do not set the seed for the RNG in unit
tests
+ * in order to guarantee deterministic behavior.
+ *
+ * TODO update tests to use TestingUtils for floating point comparison
after PR 1367 is merged
+ */
+class RandomRDDGeneratorsSuite extends FunSuite with LocalSparkContext
with Serializable {
+
+ def testGeneratedRDD(rdd: RDD[Double],
+ expectedSize: Long,
+ expectedNumPartitions: Int,
+ expectedMean: Double,
+ expectedStddev: Double,
+ epsilon: Double = 0.01) {
+ val stats = rdd.stats()
+ assert(expectedSize === stats.count)
+ assert(expectedNumPartitions === rdd.partitions.size)
+ assert(math.abs(stats.mean - expectedMean) < epsilon)
+ assert(math.abs(stats.stdev - expectedStddev) < epsilon)
+ }
+
+ // assume test RDDs are small
+ def testGeneratedVectorRDD(rdd: RDD[Vector],
+ expectedRows: Long,
+ expectedColumns: Int,
+ expectedNumPartitions: Int,
+ expectedMean: Double,
+ expectedStddev: Double,
+ epsilon: Double = 0.01) {
+ assert(expectedNumPartitions === rdd.partitions.size)
+ val values = new ArrayBuffer[Double]()
+ rdd.collect.foreach { vector => {
+ assert(vector.size === expectedColumns)
+ values ++= vector.toArray
+ }}
+ assert(expectedRows === values.size / expectedColumns)
+ val stats = new StatCounter(values)
+ assert(math.abs(stats.mean - expectedMean) < epsilon)
+ assert(math.abs(stats.stdev - expectedStddev) < epsilon)
+ }
+
+ test("RandomRDD sizes") {
+
+ // some cases where size % numParts != 0 to test getPartitions behaves
correctly
+ for ((size, numPartitions) <- List((10000, 6), (12345, 1), (1000,
101))) {
+ val rdd = new RandomRDD(sc, size, numPartitions, new
UniformGenerator, 0L)
+ assert(rdd.count() === size)
+ assert(rdd.partitions.size === numPartitions)
+
+ // check that partition sizes are balanced
+ val partSizes = rdd.partitions.map( p =>
p.asInstanceOf[RandomRDDPartition].size.toDouble)
+ val partStats = new StatCounter(partSizes)
+ assert(partStats.stdev < 1.0)
--- End diff --
Why checking the stdev of partition sizes? It should be `maxPartitionSize -
minPartitionSize <= 1`.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
