First completed and passed successfully with colon dataset.

Author: sramirez

src/main/scala/org/apache/spark/mllib/feature/InfoThSelector.scala

@@ -316,8 +316,7 @@ class InfoThSelector @Since("1.6.0") (
         case dv: DenseVector =>
           dv.values
       }
-      val condition = (value: Double) => value <= Byte.MaxValue &&
-        value >= Byte.MinValue && value % 1 == 0.0
+      val condition = (value: Double) => value <= 255 && value >= 0.0 && value % 1 == 0.0
       if (!values.forall(condition(_))) {
         val str = values.mkString(",")
         throw new SparkException(
@@ -412,7 +411,7 @@ class InfoThSelector @Since("1.6.0") (
       case F(feat, rel) =>
         (feat + 1) + "\t" + "%.4f".format(rel)
     }.mkString("\n")
-    logInfo("\n*** Selected features ***\nFeature\tScore\n" + out)
+    println("\n*** Selected features ***\nFeature\tScore\n" + out)
     // Features must be sorted
     new InfoThSelectorModel(selected.map { case F(feat, rel) => feat }.sorted.toArray)
   }

src/test/scala/org/apache/spark/ml/feature/ITSelectorSuite.scala

@@ -13,7 +13,7 @@ import TestHelper._
   * @author Sergio Ramirez
   */
 @RunWith(classOf[JUnitRunner])
-class MDLPDiscretizerSuite extends FunSuite with BeforeAndAfterAll {
+class ITSelectorSuite extends FunSuite with BeforeAndAfterAll {
 
   var sqlContext: SQLContext = null
 
@@ -22,13 +22,15 @@ class MDLPDiscretizerSuite extends FunSuite with BeforeAndAfterAll {
   }
 
   /** Do entropy based binning of cars data from UC Irvine repository. */
-  test("Run ITFS on colon data (nPart = 20, nfeat = 20)") {
+  test("Run ITFS on colon data (nPart = 10, nfeat = 10)") {
 
     val df = readColonData(sqlContext)
     val cols = df.columns
-    val model = getSelectorModel(sqlContext, df, df.columns.drop(1), df.columns.head, 10, 20)
+    val pad = 2
+    val allVectorsDense = true
+    val model = getSelectorModel(sqlContext, df, cols.drop(1), cols.head, 10, 10, allVectorsDense, pad)
 
-    assertResult("764, 1581, 1671, 512, 1670, 1324, 1381, 1971, 1422, 1411") {
+    assertResult("512, 764, 1324, 1380, 1411, 1422, 1581, 1670, 1671, 1971") {
       model.selectedFeatures.mkString(", ")
     }
   }

src/test/scala/org/apache/spark/ml/feature/TestHelper.scala

@@ -36,7 +36,8 @@ object TestHelper {
                              labelColumn: String,
                              nPartitions: Int = 100,
                              numTopFeatures: Int = 20, 
-                             allVectorsDense: Boolean = true): InfoThSelectorModel = {
+                             allVectorsDense: Boolean = true,
+                             padded: Int = 0): InfoThSelectorModel = {
     val featureAssembler = new VectorAssembler()
       .setInputCols(inputCols)
       .setOutputCol("features")
@@ -45,7 +46,14 @@ object TestHelper {
     /** InfoSelector requires all vectors from the same type (either be sparse or dense) **/
     val rddData = processedDf.rdd.map {
         case Row(label: Double, features: Vector) =>
-          val standardv = if(allVectorsDense) features.toDense else features.toSparse
+          val standardv = if(allVectorsDense){
+            Vectors.dense(features.toArray.map(_ + padded))
+          } else {
+              val sparseVec = features.toSparse
+              val newValues: Array[Double] = sparseVec.values.map(_ + padded)
+              Vectors.sparse(sparseVec.size, sparseVec.indices, newValues)
+          }
+          
           Row.fromSeq(Seq(label, standardv))
       }
 
@@ -70,9 +78,12 @@ object TestHelper {
   def getSelectorModel(sqlContext: SQLContext, dataframe: DataFrame, inputCols: Array[String],
                           labelColumn: String,
                              nPartitions: Int = 100,
-                             numTopFeatures: Int = 20): InfoThSelectorModel = {
+                             numTopFeatures: Int = 20,
+                             allVectorsDense: Boolean = true,
+                             padded: Int = 0): InfoThSelectorModel = {
     val processedDf = cleanLabelCol(dataframe, labelColumn)
-    createSelectorModel(sqlContext, processedDf, inputCols, labelColumn, nPartitions, numTopFeatures)
+    createSelectorModel(sqlContext, processedDf, inputCols, labelColumn, 
+        nPartitions, numTopFeatures, allVectorsDense, padded)
   }