[SPARK-1406] Added a PMMLExportable interface

Restructured code in a new package mllib.pmml
Supported models implements the new PMMLExportable interface:
LogisticRegression, SVM, KMeansModel, LinearRegression, RidgeRegression,
Lasso

Author: selvinsource

mllib/src/main/scala/org/apache/spark/mllib/export/ModelExport.scala

@@ -0,0 +1,66 @@
+/*
+ * 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.pmml
+
+import java.io.File
+import java.io.OutputStream
+import java.io.StringWriter
+import javax.xml.transform.stream.StreamResult
+import org.jpmml.model.JAXBUtil
+import org.apache.spark.mllib.pmml.export.PMMLModelExport
+import org.apache.spark.mllib.pmml.export.PMMLModelExportFactory
+
+/**
+ * Export model to the PMML format
+ * Predictive Model Markup Language (PMML) in an XML-based file format
+ * developed by the Data Mining Group (www.dmg.org).
+ */
+trait PMMLExportable {
+  
+  /**
+  * Export the model to the stream result in PMML format 
+  */
+  private def toPMML(streamResult: StreamResult): Unit = {
+    val pmmlModelExport = PMMLModelExportFactory.createPMMLModelExport(this)
+    JAXBUtil.marshalPMML(pmmlModelExport.getPmml(), streamResult)
+  }
+  
+  /**
+  * Export the model to a local File in PMML format 
+  */
+  def toPMML(localPath: String): Unit = {
+    toPMML(new StreamResult(new File(localPath)))
+  }
+  
+  /**
+  * Export the model to the Outputtream in PMML format 
+  */
+  def toPMML(outputStream: OutputStream): Unit = {
+    toPMML(new StreamResult(outputStream))
+  }
+  
+  /**
+  * Export the model to a String in PMML format 
+  */
+  def toPMML(): String = {
+    var writer = new StringWriter();
+    toPMML(new StreamResult(writer))
+    return writer.toString();
+  }
+  
+}

mllib/src/main/scala/org/apache/spark/mllib/export/ModelExportType.scala

@@ -15,7 +15,7 @@
  * limitations under the License.
  */
 
-package org.apache.spark.mllib.export.pmml
+package org.apache.spark.mllib.pmml.export
 
 import org.dmg.pmml.DataDictionary
 import org.dmg.pmml.DataField
@@ -29,7 +29,6 @@ import org.dmg.pmml.NumericPredictor
 import org.dmg.pmml.OpType
 import org.dmg.pmml.RegressionModel
 import org.dmg.pmml.RegressionTable
-
 import org.apache.spark.mllib.regression.GeneralizedLinearModel
 
 /**

mllib/src/main/scala/org/apache/spark/mllib/export/ModelExporter.scala

@@ -15,7 +15,7 @@
  * limitations under the License.
  */
 
-package org.apache.spark.mllib.export.pmml
+package org.apache.spark.mllib.pmml.export
 
 import org.dmg.pmml.Array.Type
 import org.dmg.pmml.Cluster
@@ -35,7 +35,6 @@ import org.dmg.pmml.MiningFunctionType
 import org.dmg.pmml.MiningSchema
 import org.dmg.pmml.OpType
 import org.dmg.pmml.SquaredEuclidean
-
 import org.apache.spark.mllib.clustering.KMeansModel
 
 /**

mllib/src/main/scala/org/apache/spark/mllib/pmml/PMMLExportable.scala

@@ -15,7 +15,7 @@
  * limitations under the License.
  */
 
-package org.apache.spark.mllib.export.pmml
+package org.apache.spark.mllib.pmml.export
 
 import org.dmg.pmml.DataDictionary
 import org.dmg.pmml.DataField
@@ -29,8 +29,8 @@ import org.dmg.pmml.NumericPredictor
 import org.dmg.pmml.OpType
 import org.dmg.pmml.RegressionModel
 import org.dmg.pmml.RegressionTable
-import org.apache.spark.mllib.classification.LogisticRegressionModel
 import org.dmg.pmml.RegressionNormalizationMethodType
+import org.apache.spark.mllib.classification.LogisticRegressionModel
 
 /**
  * PMML Model Export for LogisticRegressionModel class

mllib/src/main/scala/org/apache/spark/mllib/export/pmml/GeneralizedLinearPMMLModelExport.scala renamed to mllib/src/main/scala/org/apache/spark/mllib/pmml/export/GeneralizedLinearPMMLModelExport.scala

@@ -15,21 +15,17 @@
  * limitations under the License.
  */
 
-package org.apache.spark.mllib.export.pmml
+package org.apache.spark.mllib.pmml.export
 
 import java.text.SimpleDateFormat
 import java.util.Date
-
 import scala.beans.BeanProperty
-
 import org.dmg.pmml.Application
 import org.dmg.pmml.Header
 import org.dmg.pmml.PMML
 import org.dmg.pmml.Timestamp
 
-import org.apache.spark.mllib.export.ModelExport
-
-private[mllib] trait PMMLModelExport extends ModelExport{
+private[mllib] trait PMMLModelExport {
 
   /**
    * Holder of the exported model in PMML format

mllib/src/main/scala/org/apache/spark/mllib/export/pmml/KMeansPMMLModelExport.scala renamed to mllib/src/main/scala/org/apache/spark/mllib/pmml/export/KMeansPMMLModelExport.scala

@@ -15,30 +15,23 @@
  * limitations under the License.
  */
 
-package org.apache.spark.mllib.export
+package org.apache.spark.mllib.pmml.export
 
 import org.apache.spark.mllib.classification.LogisticRegressionModel
 import org.apache.spark.mllib.classification.SVMModel
 import org.apache.spark.mllib.clustering.KMeansModel
-import org.apache.spark.mllib.export.ModelExportType.ModelExportType
-import org.apache.spark.mllib.export.ModelExportType.PMML
-import org.apache.spark.mllib.export.pmml.GeneralizedLinearPMMLModelExport
-import org.apache.spark.mllib.export.pmml.KMeansPMMLModelExport
-import org.apache.spark.mllib.export.pmml.LogisticRegressionPMMLModelExport
 import org.apache.spark.mllib.regression.LassoModel
 import org.apache.spark.mllib.regression.LinearRegressionModel
 import org.apache.spark.mllib.regression.RidgeRegressionModel
 
-private[mllib] object ModelExportFactory {
+private[mllib] object PMMLModelExportFactory {
 
   /**
-   * Factory object to help creating the necessary ModelExport implementation 
-   * taking as input the ModelExportType (for example PMML) 
-   * and the machine learning model (for example KMeansModel).
+   * Factory object to help creating the necessary PMMLModelExport implementation 
+   * taking as input the machine learning model (for example KMeansModel).
    */
-  def createModelExport(model: Any, exportType: ModelExportType): ModelExport = {
-    return exportType match{
-      case PMML => model match{
+  def createPMMLModelExport(model: Any): PMMLModelExport = {
+    return model match{
         case kmeans: KMeansModel => 
           new KMeansPMMLModelExport(kmeans)
         case linearRegression: LinearRegressionModel => 
@@ -54,10 +47,8 @@ private[mllib] object ModelExportFactory {
         case logisticRegression: LogisticRegressionModel => 
           new LogisticRegressionPMMLModelExport(logisticRegression, "logistic regression")
         case _ => 
-          throw new IllegalArgumentException("Export not supported for model: " + model.getClass)
-      }
-      case _ => throw new IllegalArgumentException("Export type not supported:" + exportType)
-      }
+          throw new IllegalArgumentException("PMML Export not supported for model: " + model.getClass)
+     }
   }
 
 }

mllib/src/main/scala/org/apache/spark/mllib/export/pmml/LogisticRegressionPMMLModelExport.scala renamed to mllib/src/main/scala/org/apache/spark/mllib/pmml/export/LogisticRegressionPMMLModelExport.scala

@@ -15,14 +15,11 @@
  * limitations under the License.
  */
 
-package org.apache.spark.mllib.export.pmml
+package org.apache.spark.mllib.pmml.export
 
 import org.dmg.pmml.RegressionModel
 import org.scalatest.FunSuite
-
 import org.apache.spark.mllib.classification.SVMModel
-import org.apache.spark.mllib.export.ModelExportFactory
-import org.apache.spark.mllib.export.ModelExportType
 import org.apache.spark.mllib.regression.LassoModel
 import org.apache.spark.mllib.regression.LinearRegressionModel
 import org.apache.spark.mllib.regression.RidgeRegressionModel
@@ -41,7 +38,7 @@ class GeneralizedLinearPMMLModelExportSuite extends FunSuite{
     val svmModel = new SVMModel(linearInput(0).features, linearInput(0).label);
 
     //act by exporting the model to the PMML format
-    val linearModelExport = ModelExportFactory.createModelExport(linearRegressionModel, ModelExportType.PMML)         
+    val linearModelExport = PMMLModelExportFactory.createPMMLModelExport(linearRegressionModel)         
     //assert that the PMML format is as expected
     assert(linearModelExport.isInstanceOf[PMMLModelExport])
     var pmml = linearModelExport.asInstanceOf[PMMLModelExport].getPmml()
@@ -54,7 +51,7 @@ class GeneralizedLinearPMMLModelExportSuite extends FunSuite{
      .getRegressionTables().get(0).getNumericPredictors().size() === linearRegressionModel.weights.size)
 
     //act
-    val ridgeModelExport = ModelExportFactory.createModelExport(ridgeRegressionModel, ModelExportType.PMML)         
+    val ridgeModelExport = PMMLModelExportFactory.createPMMLModelExport(ridgeRegressionModel)         
     //assert that the PMML format is as expected
     assert(ridgeModelExport.isInstanceOf[PMMLModelExport])
     pmml = ridgeModelExport.asInstanceOf[PMMLModelExport].getPmml()
@@ -67,7 +64,7 @@ class GeneralizedLinearPMMLModelExportSuite extends FunSuite{
      .getRegressionTables().get(0).getNumericPredictors().size() === ridgeRegressionModel.weights.size)
 
     //act
-    val lassoModelExport = ModelExportFactory.createModelExport(lassoModel, ModelExportType.PMML)         
+    val lassoModelExport = PMMLModelExportFactory.createPMMLModelExport(lassoModel)         
     //assert that the PMML format is as expected
     assert(lassoModelExport.isInstanceOf[PMMLModelExport])
     pmml = lassoModelExport.asInstanceOf[PMMLModelExport].getPmml()
@@ -80,7 +77,7 @@ class GeneralizedLinearPMMLModelExportSuite extends FunSuite{
      .getRegressionTables().get(0).getNumericPredictors().size() === lassoModel.weights.size)
 
     //act
-    val svmModelExport = ModelExportFactory.createModelExport(svmModel, ModelExportType.PMML)         
+    val svmModelExport = PMMLModelExportFactory.createPMMLModelExport(svmModel)         
     //assert that the PMML format is as expected
     assert(svmModelExport.isInstanceOf[PMMLModelExport])
     pmml = svmModelExport.asInstanceOf[PMMLModelExport].getPmml()
@@ -93,10 +90,10 @@ class GeneralizedLinearPMMLModelExportSuite extends FunSuite{
      .getRegressionTables().get(0).getNumericPredictors().size() === svmModel.weights.size)
 
     //manual checking
-    //ModelExporter.toPMML(linearRegressionModel,"/tmp/linearregression.xml")
-    //ModelExporter.toPMML(ridgeRegressionModel,"/tmp/ridgeregression.xml")
-    //ModelExporter.toPMML(lassoModel,"/tmp/lassoregression.xml")
-    //ModelExporter.toPMML(svmModel,"/tmp/linearsvm.xml")
+    //linearRegressionModel.toPMML("/tmp/linearregression.xml")
+    //ridgeRegressionModel.toPMML("/tmp/ridgeregression.xml")
+    //lassoModel.toPMML("/tmp/lassoregression.xml")
+    //svmModel.toPMML("/tmp/linearsvm.xml")
 
    }