[SPARK-13761] [ML] Deprecate validateParams

mllib/src/main/scala/org/apache/spark/ml/Pipeline.scala

@@ -110,12 +110,6 @@ class Pipeline @Since("1.4.0") (
   @Since("1.2.0")
   def getStages: Array[PipelineStage] = $(stages).clone()
 
-  @Since("1.4.0")
-  override def validateParams(): Unit = {
-    super.validateParams()
-    $(stages).foreach(_.validateParams())
-  }
-
   /**
    * Fits the pipeline to the input dataset with additional parameters. If a stage is an
    * [[Estimator]], its [[Estimator#fit]] method will be called on the input dataset to fit a model.
@@ -175,7 +169,6 @@ class Pipeline @Since("1.4.0") (
 
   @Since("1.2.0")
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val theStages = $(stages)
     require(theStages.toSet.size == theStages.length,
       "Cannot have duplicate components in a pipeline.")
@@ -297,12 +290,6 @@ class PipelineModel private[ml] (
     this(uid, stages.asScala.toArray)
   }
 
-  @Since("1.4.0")
-  override def validateParams(): Unit = {
-    super.validateParams()
-    stages.foreach(_.validateParams())
-  }
-
   @Since("1.2.0")
   override def transform(dataset: DataFrame): DataFrame = {
     transformSchema(dataset.schema, logging = true)
@@ -311,7 +298,6 @@ class PipelineModel private[ml] (
 
   @Since("1.2.0")
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     stages.foldLeft(schema)((cur, transformer) => transformer.transformSchema(cur))
   }
 

mllib/src/main/scala/org/apache/spark/ml/Predictor.scala

@@ -46,7 +46,6 @@ private[ml] trait PredictorParams extends Params
       schema: StructType,
       fitting: Boolean,
       featuresDataType: DataType): StructType = {
-    validateParams()
     // TODO: Support casting Array[Double] and Array[Float] to Vector when FeaturesType = Vector
     SchemaUtils.checkColumnType(schema, $(featuresCol), featuresDataType)
     if (fitting) {

mllib/src/main/scala/org/apache/spark/ml/Transformer.scala

@@ -103,7 +103,6 @@ abstract class UnaryTransformer[IN, OUT, T <: UnaryTransformer[IN, OUT, T]]
   protected def validateInputType(inputType: DataType): Unit = {}
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     validateInputType(inputType)
     if (schema.fieldNames.contains($(outputCol))) {

mllib/src/main/scala/org/apache/spark/ml/clustering/KMeans.scala

@@ -81,7 +81,6 @@ private[clustering] trait KMeansParams extends Params with HasMaxIter with HasFe
    * @return output schema
    */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(featuresCol), new VectorUDT)
     SchemaUtils.appendColumn(schema, $(predictionCol), IntegerType)
   }

mllib/src/main/scala/org/apache/spark/ml/clustering/LDA.scala

@@ -263,13 +263,6 @@ private[clustering] trait LDAParams extends Params with HasFeaturesCol with HasM
    * @return output schema
    */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
-    SchemaUtils.checkColumnType(schema, $(featuresCol), new VectorUDT)
-    SchemaUtils.appendColumn(schema, $(topicDistributionCol), new VectorUDT)
-  }
-
-  @Since("1.6.0")
-  override def validateParams(): Unit = {
     if (isSet(docConcentration)) {
       if (getDocConcentration.length != 1) {
         require(getDocConcentration.length == getK, s"LDA docConcentration was of length" +
@@ -297,6 +290,8 @@ private[clustering] trait LDAParams extends Params with HasFeaturesCol with HasM
             s" must be >= 1.  Found value: $getTopicConcentration")
       }
     }
+    SchemaUtils.checkColumnType(schema, $(featuresCol), new VectorUDT)
+    SchemaUtils.appendColumn(schema, $(topicDistributionCol), new VectorUDT)
   }
 
   private[clustering] def getOldOptimizer: OldLDAOptimizer = getOptimizer match {

mllib/src/main/scala/org/apache/spark/ml/feature/Bucketizer.scala

@@ -86,7 +86,6 @@ final class Bucketizer(override val uid: String)
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(inputCol), DoubleType)
     SchemaUtils.appendColumn(schema, prepOutputField(schema))
   }

mllib/src/main/scala/org/apache/spark/ml/feature/ChiSqSelector.scala

@@ -88,7 +88,6 @@ final class ChiSqSelector(override val uid: String)
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(featuresCol), new VectorUDT)
     SchemaUtils.checkColumnType(schema, $(labelCol), DoubleType)
     SchemaUtils.appendColumn(schema, $(outputCol), new VectorUDT)
@@ -136,7 +135,6 @@ final class ChiSqSelectorModel private[ml] (
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(featuresCol), new VectorUDT)
     val newField = prepOutputField(schema)
     val outputFields = schema.fields :+ newField

mllib/src/main/scala/org/apache/spark/ml/feature/CountVectorizer.scala

@@ -70,7 +70,6 @@ private[feature] trait CountVectorizerParams extends Params with HasInputCol wit
 
   /** Validates and transforms the input schema. */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
     val typeCandidates = List(new ArrayType(StringType, true), new ArrayType(StringType, false))
     SchemaUtils.checkColumnTypes(schema, $(inputCol), typeCandidates)
     SchemaUtils.appendColumn(schema, $(outputCol), new VectorUDT)

mllib/src/main/scala/org/apache/spark/ml/feature/HashingTF.scala

@@ -69,7 +69,6 @@ class HashingTF(override val uid: String)
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     require(inputType.isInstanceOf[ArrayType],
       s"The input column must be ArrayType, but got $inputType.")

mllib/src/main/scala/org/apache/spark/ml/feature/IDF.scala

@@ -52,7 +52,6 @@ private[feature] trait IDFBase extends Params with HasInputCol with HasOutputCol
    * Validate and transform the input schema.
    */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(inputCol), new VectorUDT)
     SchemaUtils.appendColumn(schema, $(outputCol), new VectorUDT)
   }

mllib/src/main/scala/org/apache/spark/ml/feature/Interaction.scala

@@ -61,13 +61,15 @@ class Interaction @Since("1.6.0") (override val uid: String) extends Transformer
   // optimistic schema; does not contain any ML attributes
   @Since("1.6.0")
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
+    require(get(inputCols).isDefined, "Input cols must be defined first.")
+    require(get(outputCol).isDefined, "Output col must be defined first.")
+    require($(inputCols).length > 0, "Input cols must have non-zero length.")
+    require($(inputCols).distinct.length == $(inputCols).length, "Input cols must be distinct.")
     StructType(schema.fields :+ StructField($(outputCol), new VectorUDT, false))
   }
 
   @Since("1.6.0")
   override def transform(dataset: DataFrame): DataFrame = {
-    validateParams()
     val inputFeatures = $(inputCols).map(c => dataset.schema(c))
     val featureEncoders = getFeatureEncoders(inputFeatures)
     val featureAttrs = getFeatureAttrs(inputFeatures)
@@ -217,13 +219,6 @@ class Interaction @Since("1.6.0") (override val uid: String) extends Transformer
   @Since("1.6.0")
   override def copy(extra: ParamMap): Interaction = defaultCopy(extra)
 
-  @Since("1.6.0")
-  override def validateParams(): Unit = {
-    require(get(inputCols).isDefined, "Input cols must be defined first.")
-    require(get(outputCol).isDefined, "Output col must be defined first.")
-    require($(inputCols).length > 0, "Input cols must have non-zero length.")
-    require($(inputCols).distinct.length == $(inputCols).length, "Input cols must be distinct.")
-  }
 }
 
 @Since("1.6.0")

mllib/src/main/scala/org/apache/spark/ml/feature/MaxAbsScaler.scala

@@ -37,7 +37,6 @@ private[feature] trait MaxAbsScalerParams extends Params with HasInputCol with H
 
    /** Validates and transforms the input schema. */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     require(inputType.isInstanceOf[VectorUDT],
       s"Input column ${$(inputCol)} must be a vector column")

mllib/src/main/scala/org/apache/spark/ml/feature/MinMaxScaler.scala

@@ -59,7 +59,7 @@ private[feature] trait MinMaxScalerParams extends Params with HasInputCol with H
 
   /** Validates and transforms the input schema. */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
+    require($(min) < $(max), s"The specified min(${$(min)}) is larger or equal to max(${$(max)})")
     val inputType = schema($(inputCol)).dataType
     require(inputType.isInstanceOf[VectorUDT],
       s"Input column ${$(inputCol)} must be a vector column")
@@ -69,9 +69,6 @@ private[feature] trait MinMaxScalerParams extends Params with HasInputCol with H
     StructType(outputFields)
   }
 
-  override def validateParams(): Unit = {
-    require($(min) < $(max), s"The specified min(${$(min)}) is larger or equal to max(${$(max)})")
-  }
 }
 
 /**

mllib/src/main/scala/org/apache/spark/ml/feature/OneHotEncoder.scala

@@ -66,7 +66,6 @@ class OneHotEncoder(override val uid: String) extends Transformer
   def setOutputCol(value: String): this.type = set(outputCol, value)
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputColName = $(inputCol)
     val outputColName = $(outputCol)
 

mllib/src/main/scala/org/apache/spark/ml/feature/PCA.scala

@@ -77,7 +77,6 @@ class PCA (override val uid: String) extends Estimator[PCAModel] with PCAParams
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     require(inputType.isInstanceOf[VectorUDT],
       s"Input column ${$(inputCol)} must be a vector column")
@@ -133,7 +132,6 @@ class PCAModel private[ml] (
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     require(inputType.isInstanceOf[VectorUDT],
       s"Input column ${$(inputCol)} must be a vector column")

mllib/src/main/scala/org/apache/spark/ml/feature/QuantileDiscretizer.scala

@@ -78,7 +78,6 @@ final class QuantileDiscretizer(override val uid: String)
   def setSeed(value: Long): this.type = set(seed, value)
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(inputCol), DoubleType)
     val inputFields = schema.fields
     require(inputFields.forall(_.name != $(outputCol)),

mllib/src/main/scala/org/apache/spark/ml/feature/RFormula.scala

@@ -167,7 +167,6 @@ class RFormula(override val uid: String) extends Estimator[RFormulaModel] with R
 
   // optimistic schema; does not contain any ML attributes
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     if (hasLabelCol(schema)) {
       StructType(schema.fields :+ StructField($(featuresCol), new VectorUDT, true))
     } else {
@@ -200,7 +199,6 @@ class RFormulaModel private[feature](
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     checkCanTransform(schema)
     val withFeatures = pipelineModel.transformSchema(schema)
     if (hasLabelCol(withFeatures)) {
@@ -263,7 +261,6 @@ private class ColumnPruner(columnsToPrune: Set[String]) extends Transformer {
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     StructType(schema.fields.filter(col => !columnsToPrune.contains(col.name)))
   }
 
@@ -312,7 +309,6 @@ private class VectorAttributeRewriter(
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     StructType(
       schema.fields.filter(_.name != vectorCol) ++
       schema.fields.filter(_.name == vectorCol))

mllib/src/main/scala/org/apache/spark/ml/feature/SQLTransformer.scala

@@ -74,7 +74,6 @@ class SQLTransformer @Since("1.6.0") (override val uid: String) extends Transfor
 
   @Since("1.6.0")
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val sc = SparkContext.getOrCreate()
     val sqlContext = SQLContext.getOrCreate(sc)
     val dummyRDD = sc.parallelize(Seq(Row.empty))

mllib/src/main/scala/org/apache/spark/ml/feature/StandardScaler.scala

@@ -94,7 +94,6 @@ class StandardScaler(override val uid: String) extends Estimator[StandardScalerM
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     require(inputType.isInstanceOf[VectorUDT],
       s"Input column ${$(inputCol)} must be a vector column")
@@ -144,7 +143,6 @@ class StandardScalerModel private[ml] (
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     require(inputType.isInstanceOf[VectorUDT],
       s"Input column ${$(inputCol)} must be a vector column")

mllib/src/main/scala/org/apache/spark/ml/feature/StopWordsRemover.scala

@@ -145,7 +145,6 @@ class StopWordsRemover(override val uid: String)
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputType = schema($(inputCol)).dataType
     require(inputType.sameType(ArrayType(StringType)),
       s"Input type must be ArrayType(StringType) but got $inputType.")

mllib/src/main/scala/org/apache/spark/ml/feature/StringIndexer.scala

@@ -39,7 +39,6 @@ private[feature] trait StringIndexerBase extends Params with HasInputCol with Ha
 
   /** Validates and transforms the input schema. */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputColName = $(inputCol)
     val inputDataType = schema(inputColName).dataType
     require(inputDataType == StringType || inputDataType.isInstanceOf[NumericType],
@@ -275,7 +274,6 @@ class IndexToString private[ml] (override val uid: String)
   final def getLabels: Array[String] = $(labels)
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputColName = $(inputCol)
     val inputDataType = schema(inputColName).dataType
     require(inputDataType.isInstanceOf[NumericType],

mllib/src/main/scala/org/apache/spark/ml/feature/VectorAssembler.scala

@@ -106,7 +106,6 @@ class VectorAssembler(override val uid: String)
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val inputColNames = $(inputCols)
     val outputColName = $(outputCol)
     val inputDataTypes = inputColNames.map(name => schema(name).dataType)

mllib/src/main/scala/org/apache/spark/ml/feature/VectorIndexer.scala

@@ -126,7 +126,6 @@ class VectorIndexer(override val uid: String) extends Estimator[VectorIndexerMod
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     // We do not transfer feature metadata since we do not know what types of features we will
     // produce in transform().
     val dataType = new VectorUDT
@@ -355,7 +354,6 @@ class VectorIndexerModel private[ml] (
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     val dataType = new VectorUDT
     require(isDefined(inputCol),
       s"VectorIndexerModel requires input column parameter: $inputCol")

mllib/src/main/scala/org/apache/spark/ml/feature/VectorSlicer.scala

@@ -89,11 +89,6 @@ final class VectorSlicer(override val uid: String)
   /** @group setParam */
   def setOutputCol(value: String): this.type = set(outputCol, value)
 
-  override def validateParams(): Unit = {
-    require($(indices).length > 0 || $(names).length > 0,
-      s"VectorSlicer requires that at least one feature be selected.")
-  }
-
   override def transform(dataset: DataFrame): DataFrame = {
     // Validity checks
     transformSchema(dataset.schema)
@@ -139,7 +134,8 @@ final class VectorSlicer(override val uid: String)
   }
 
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
+    require($(indices).length > 0 || $(names).length > 0,
+      s"VectorSlicer requires that at least one feature be selected.")
     SchemaUtils.checkColumnType(schema, $(inputCol), new VectorUDT)
 
     if (schema.fieldNames.contains($(outputCol))) {

mllib/src/main/scala/org/apache/spark/ml/feature/Word2Vec.scala

@@ -92,7 +92,6 @@ private[feature] trait Word2VecBase extends Params
    * Validate and transform the input schema.
    */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(inputCol), new ArrayType(StringType, true))
     SchemaUtils.appendColumn(schema, $(outputCol), new VectorUDT)
   }

mllib/src/main/scala/org/apache/spark/ml/param/params.scala

@@ -58,9 +58,8 @@ class Param[T](val parent: String, val name: String, val doc: String, val isVali
   /**
    * Assert that the given value is valid for this parameter.
    *
-   * Note: Parameter checks involving interactions between multiple parameters should be
-   *       implemented in [[Params.validateParams()]].  Checks for input/output columns should be
-   *       implemented in [[org.apache.spark.ml.PipelineStage.transformSchema()]].
+   * Note: Parameter checks involving interactions between multiple parameters and input/output
+   * columns should be implemented in [[org.apache.spark.ml.PipelineStage.transformSchema()]].
    *
    * DEVELOPERS: This method is only called by [[ParamPair]], which means that all parameters
    *             should be specified via [[ParamPair]].
@@ -549,7 +548,9 @@ trait Params extends Identifiable with Serializable {
    * Parameter value checks which do not depend on other parameters are handled by
    * [[Param.validate()]].  This method does not handle input/output column parameters;
    * those are checked during schema validation.
+   * @deprecated Will be removed in 2.1.0. All the checks should be merged into transformSchema
    */
+  @deprecated("Will be removed in 2.1.0. Checks should be merged into transformSchema.", "2.0.0")
   def validateParams(): Unit = {
     // Do nothing by default.  Override to handle Param interactions.
   }

mllib/src/main/scala/org/apache/spark/ml/recommendation/ALS.scala

@@ -162,7 +162,6 @@ private[recommendation] trait ALSParams extends ALSModelParams with HasMaxIter w
    * @return output schema
    */
   protected def validateAndTransformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(userCol), IntegerType)
     SchemaUtils.checkColumnType(schema, $(itemCol), IntegerType)
     val ratingType = schema($(ratingCol)).dataType
@@ -220,7 +219,6 @@ class ALSModel private[ml] (
 
   @Since("1.3.0")
   override def transformSchema(schema: StructType): StructType = {
-    validateParams()
     SchemaUtils.checkColumnType(schema, $(userCol), IntegerType)
     SchemaUtils.checkColumnType(schema, $(itemCol), IntegerType)
     SchemaUtils.appendColumn(schema, $(predictionCol), FloatType)