[SPARK-30144][ML][PYSPARK] Make MultilayerPerceptronClassificationModel extend MultilayerPerceptronParams

### What changes were proposed in this pull request?
Make ```MultilayerPerceptronClassificationModel``` extend ```MultilayerPerceptronParams```

### Why are the changes needed?
Make ```MultilayerPerceptronClassificationModel``` extend ```MultilayerPerceptronParams``` to expose the training params, so user can see these params when calling ```extractParamMap```

### Does this PR introduce any user-facing change?
Yes. The ```MultilayerPerceptronParams``` such as ```seed```, ```maxIter``` ... are available in ```MultilayerPerceptronClassificationModel``` now

### How was this patch tested?
Manually tested ```MultilayerPerceptronClassificationModel.extractParamMap()``` to verify all the new params are there.

Closes #26838 from huaxingao/spark-30144.

Authored-by: Huaxin Gao <huaxing@us.ibm.com>
Signed-off-by: Sean Owen <srowen@gmail.com>

Author: huaxingao

mllib/src/main/scala/org/apache/spark/ml/classification/MultilayerPerceptronClassifier.scala

@@ -30,6 +30,7 @@ import org.apache.spark.ml.param.shared._
 import org.apache.spark.ml.util._
 import org.apache.spark.ml.util.Instrumentation.instrumented
 import org.apache.spark.sql.{Dataset, Row}
+import org.apache.spark.util.VersionUtils.majorMinorVersion
 
 /** Params for Multilayer Perceptron. */
 private[classification] trait MultilayerPerceptronParams extends ProbabilisticClassifierParams
@@ -247,7 +248,7 @@ class MultilayerPerceptronClassifier @Since("1.5.0") (
     }
     trainer.setStackSize($(blockSize))
     val mlpModel = trainer.train(data)
-    new MultilayerPerceptronClassificationModel(uid, myLayers, mlpModel.weights)
+    new MultilayerPerceptronClassificationModel(uid, mlpModel.weights)
   }
 }
 
@@ -273,31 +274,22 @@ object MultilayerPerceptronClassifier
  * Each layer has sigmoid activation function, output layer has softmax.
  *
  * @param uid uid
- * @param layers array of layer sizes including input and output layers
  * @param weights the weights of layers
  */
 @Since("1.5.0")
 class MultilayerPerceptronClassificationModel private[ml] (
     @Since("1.5.0") override val uid: String,
-    @Since("1.5.0") val layers: Array[Int],
     @Since("2.0.0") val weights: Vector)
   extends ProbabilisticClassificationModel[Vector, MultilayerPerceptronClassificationModel]
-  with Serializable with MLWritable {
+  with MultilayerPerceptronParams with Serializable with MLWritable {
 
   @Since("1.6.0")
-  override val numFeatures: Int = layers.head
+  override lazy val numFeatures: Int = $(layers).head
 
-  private[ml] val mlpModel = FeedForwardTopology
-    .multiLayerPerceptron(layers, softmaxOnTop = true)
+  @transient private[ml] lazy val mlpModel = FeedForwardTopology
+    .multiLayerPerceptron($(layers), softmaxOnTop = true)
     .model(weights)
 
-  /**
-   * Returns layers in a Java List.
-   */
-  private[ml] def javaLayers: java.util.List[Int] = {
-    layers.toList.asJava
-  }
-
   /**
    * Predict label for the given features.
    * This internal method is used to implement `transform()` and output [[predictionCol]].
@@ -308,7 +300,8 @@ class MultilayerPerceptronClassificationModel private[ml] (
 
   @Since("1.5.0")
   override def copy(extra: ParamMap): MultilayerPerceptronClassificationModel = {
-    val copied = new MultilayerPerceptronClassificationModel(uid, layers, weights).setParent(parent)
+    val copied = new MultilayerPerceptronClassificationModel(uid, weights)
+      .setParent(parent)
     copyValues(copied, extra)
   }
 
@@ -323,11 +316,11 @@ class MultilayerPerceptronClassificationModel private[ml] (
   @Since("3.0.0")
   override def predictRaw(features: Vector): Vector = mlpModel.predictRaw(features)
 
-  override def numClasses: Int = layers.last
+  override def numClasses: Int = $(layers).last
 
   @Since("3.0.0")
   override def toString: String = {
-    s"MultilayerPerceptronClassificationModel: uid=$uid, numLayers=${layers.length}, " +
+    s"MultilayerPerceptronClassificationModel: uid=$uid, numLayers=${$(layers).length}, " +
       s"numClasses=$numClasses, numFeatures=$numFeatures"
   }
 }
@@ -348,13 +341,13 @@ object MultilayerPerceptronClassificationModel
   class MultilayerPerceptronClassificationModelWriter(
       instance: MultilayerPerceptronClassificationModel) extends MLWriter {
 
-    private case class Data(layers: Array[Int], weights: Vector)
+    private case class Data(weights: Vector)
 
     override protected def saveImpl(path: String): Unit = {
       // Save metadata and Params
       DefaultParamsWriter.saveMetadata(instance, path, sc)
-      // Save model data: layers, weights
-      val data = Data(instance.layers, instance.weights)
+      // Save model data: weights
+      val data = Data(instance.weights)
       val dataPath = new Path(path, "data").toString
       sparkSession.createDataFrame(Seq(data)).repartition(1).write.parquet(dataPath)
     }
@@ -368,13 +361,21 @@ object MultilayerPerceptronClassificationModel
 
     override def load(path: String): MultilayerPerceptronClassificationModel = {
       val metadata = DefaultParamsReader.loadMetadata(path, sc, className)
+      val (majorVersion, _) = majorMinorVersion(metadata.sparkVersion)
 
       val dataPath = new Path(path, "data").toString
-      val data = sparkSession.read.parquet(dataPath).select("layers", "weights").head()
-      val layers = data.getAs[Seq[Int]](0).toArray
-      val weights = data.getAs[Vector](1)
-      val model = new MultilayerPerceptronClassificationModel(metadata.uid, layers, weights)
-
+      val df = sparkSession.read.parquet(dataPath)
+      val model = if (majorVersion < 3) { // model prior to 3.0.0
+        val data = df.select("layers", "weights").head()
+        val layers = data.getAs[Seq[Int]](0).toArray
+        val weights = data.getAs[Vector](1)
+        val model = new MultilayerPerceptronClassificationModel(metadata.uid, weights)
+        model.set("layers", layers)
+      } else {
+        val data = df.select("weights").head()
+        val weights = data.getAs[Vector](0)
+        new MultilayerPerceptronClassificationModel(metadata.uid, weights)
+      }
       metadata.getAndSetParams(model)
       model
     }

mllib/src/main/scala/org/apache/spark/ml/r/MultilayerPerceptronClassifierWrapper.scala

@@ -40,7 +40,7 @@ private[r] class MultilayerPerceptronClassifierWrapper private (
     pipeline.stages(1).asInstanceOf[MultilayerPerceptronClassificationModel]
 
   lazy val weights: Array[Double] = mlpModel.weights.toArray
-  lazy val layers: Array[Int] = mlpModel.layers
+  lazy val layers: Array[Int] = mlpModel.getLayers
 
   def transform(dataset: Dataset[_]): DataFrame = {
     pipeline.transform(dataset)

mllib/src/test/resources/ml-models/hashingTF-2.4.4/metadata/._SUCCESS.crc

@@ -0,0 +1 @@
+{"class":"org.apache.spark.ml.feature.HashingTF","timestamp":1577833408759,"sparkVersion":"2.4.4","uid":"hashingTF_f4565fe7f7da","paramMap":{"numFeatures":100,"outputCol":"features","inputCol":"words","binary":true},"defaultParamMap":{"numFeatures":262144,"outputCol":"hashingTF_f4565fe7f7da__output","binary":false}}

mllib/src/test/resources/ml-models/hashingTF-2.4.4/metadata/.part-00000.crc

@@ -0,0 +1 @@
+{"class":"org.apache.spark.ml.classification.MultilayerPerceptronClassificationModel","timestamp":1577833765310,"sparkVersion":"2.4.4","uid":"mlpc_30aa2f44dacc","paramMap":{},"defaultParamMap":{"rawPredictionCol":"rawPrediction","predictionCol":"prediction","probabilityCol":"probability","labelCol":"label","featuresCol":"features"}}

mllib/src/test/resources/test-data/hashingTF-pre3.0/metadata/_SUCCESS renamed to mllib/src/test/resources/ml-models/hashingTF-2.4.4/metadata/_SUCCESS

@@ -0,0 +1 @@
+{"class":"org.apache.spark.ml.feature.StringIndexerModel","timestamp":1577831053235,"sparkVersion":"2.4.4","uid":"myStringIndexerModel","paramMap":{"inputCol":"myInputCol","outputCol":"myOutputCol","handleInvalid":"skip"},"defaultParamMap":{"outputCol":"myStringIndexerModel__output","handleInvalid":"error"}}

mllib/src/test/resources/ml-models/hashingTF-2.4.4/metadata/part-00000

@@ -229,4 +229,17 @@ class MultilayerPerceptronClassifierSuite extends MLTest with DefaultReadWriteTe
         assert(expected.weights === actual.weights)
       }
   }
+
+  test("Load MultilayerPerceptronClassificationModel prior to Spark 3.0") {
+    val mlpPath = testFile("ml-models/mlp-2.4.4")
+    val model = MultilayerPerceptronClassificationModel.load(mlpPath)
+    val layers = model.getLayers
+    assert(layers(0) === 4)
+    assert(layers(1) === 5)
+    assert(layers(2) === 2)
+
+    val metadata = spark.read.json(s"$mlpPath/metadata")
+    val sparkVersionStr = metadata.select("sparkVersion").first().getString(0)
+    assert(sparkVersionStr == "2.4.4")
+  }
 }

mllib/src/test/resources/ml-models/mlp-2.4.4/data/._SUCCESS.crc

@@ -89,7 +89,7 @@ class HashingTFSuite extends MLTest with DefaultReadWriteTest {
   }
 
   test("SPARK-23469: Load HashingTF prior to Spark 3.0") {
-    val hashingTFPath = testFile("test-data/hashingTF-pre3.0")
+    val hashingTFPath = testFile("ml-models/hashingTF-2.4.4")
     val loadedHashingTF = HashingTF.load(hashingTFPath)
     val mLlibHashingTF = new MLlibHashingTF(100)
     assert(loadedHashingTF.indexOf("a") === mLlibHashingTF.indexOf("a"))
@@ -99,7 +99,7 @@ class HashingTFSuite extends MLTest with DefaultReadWriteTest {
 
     val metadata = spark.read.json(s"$hashingTFPath/metadata")
     val sparkVersionStr = metadata.select("sparkVersion").first().getString(0)
-    assert(sparkVersionStr == "2.3.0-SNAPSHOT")
+    assert(sparkVersionStr == "2.4.4")
   }
 
   test("read/write") {

mllib/src/test/resources/ml-models/mlp-2.4.4/data/.part-00000-fa18aaf6-d8df-4b90-8231-eb5f6ac12138-c000.snappy.parquet.crc

@@ -459,13 +459,13 @@ class StringIndexerSuite extends MLTest with DefaultReadWriteTest {
   }
 
   test("Load StringIndexderModel prior to Spark 3.0") {
-    val modelPath = testFile("test-data/strIndexerModel")
+    val modelPath = testFile("ml-models/strIndexerModel-2.4.4")
 
     val loadedModel = StringIndexerModel.load(modelPath)
     assert(loadedModel.labelsArray === Array(Array("b", "c", "a")))
 
     val metadata = spark.read.json(s"$modelPath/metadata")
     val sparkVersionStr = metadata.select("sparkVersion").first().getString(0)
-    assert(sparkVersionStr == "2.4.1-SNAPSHOT")
+    assert(sparkVersionStr == "2.4.4")
   }
 }

mllib/src/test/resources/ml-models/mlp-2.4.4/data/_SUCCESS

@@ -332,6 +332,10 @@ object MimaExcludes {
     // [SPARK-26457] Show hadoop configurations in HistoryServer environment tab
     ProblemFilters.exclude[DirectMissingMethodProblem]("org.apache.spark.status.api.v1.ApplicationEnvironmentInfo.this"),
 
+    // [SPARK-30144][ML] Make MultilayerPerceptronClassificationModel extend MultilayerPerceptronParams
+    ProblemFilters.exclude[IncompatibleResultTypeProblem]("org.apache.spark.ml.classification.MultilayerPerceptronClassificationModel.layers"),
+    ProblemFilters.exclude[DirectMissingMethodProblem]("org.apache.spark.ml.classification.MultilayerPerceptronClassificationModel.this"),
+
     // Data Source V2 API changes
     (problem: Problem) => problem match {
       case MissingClassProblem(cls) =>

mllib/src/test/resources/ml-models/mlp-2.4.4/data/part-00000-fa18aaf6-d8df-4b90-8231-eb5f6ac12138-c000.snappy.parquet

@@ -2201,7 +2201,9 @@ class MultilayerPerceptronClassifier(JavaProbabilisticClassifier, _MultilayerPer
     >>> model = mlp.fit(df)
     >>> model.setFeaturesCol("features")
     MultilayerPerceptronClassificationModel...
-    >>> model.layers
+    >>> model.getMaxIter()
+    100
+    >>> model.getLayers()
     [2, 2, 2]
     >>> model.weights.size
     12
@@ -2230,15 +2232,15 @@ class MultilayerPerceptronClassifier(JavaProbabilisticClassifier, _MultilayerPer
     >>> model_path = temp_path + "/mlp_model"
     >>> model.save(model_path)
     >>> model2 = MultilayerPerceptronClassificationModel.load(model_path)
-    >>> model.layers == model2.layers
+    >>> model.getLayers() == model2.getLayers()
     True
     >>> model.weights == model2.weights
     True
     >>> mlp2 = mlp2.setInitialWeights(list(range(0, 12)))
     >>> model3 = mlp2.fit(df)
     >>> model3.weights != model2.weights
     True
-    >>> model3.layers == model.layers
+    >>> model3.getLayers() == model.getLayers()
     True
 
     .. versionadded:: 1.6.0
@@ -2334,22 +2336,15 @@ def setSolver(self, value):
         return self._set(solver=value)
 
 
-class MultilayerPerceptronClassificationModel(JavaProbabilisticClassificationModel, JavaMLWritable,
+class MultilayerPerceptronClassificationModel(JavaProbabilisticClassificationModel,
+                                              _MultilayerPerceptronParams, JavaMLWritable,
                                               JavaMLReadable):
     """
     Model fitted by MultilayerPerceptronClassifier.
 
     .. versionadded:: 1.6.0
     """
 
-    @property
-    @since("1.6.0")
-    def layers(self):
-        """
-        array of layer sizes including input and output layers.
-        """
-        return self._call_java("javaLayers")
-
     @property
     @since("2.0.0")
     def weights(self):