Support data instance class for user usage, and add test cases for model serialization

samoa-api/src/main/java/org/apache/samoa/evaluation/EvaluatorProcessor.java

@@ -31,7 +31,7 @@
 
 import org.apache.samoa.core.ContentEvent;
 import org.apache.samoa.core.Processor;
-import org.apache.samoa.learners.ModelContentEvent;
+import org.apache.samoa.learners.ClassificationModelContentEvent;
 import org.apache.samoa.learners.ResultContentEvent;
 import org.apache.samoa.moa.core.Measurement;
 import org.apache.samoa.moa.evaluation.LearningCurve;
@@ -74,7 +74,7 @@ private EvaluatorProcessor(Builder builder) {
   @Override
   public boolean process(ContentEvent event) {
     // for serialize
-    if (event instanceof ModelContentEvent) {
+    if (event instanceof ClassificationModelContentEvent) {
       return false;
     }
 

samoa-api/src/main/java/org/apache/samoa/learners/ModelContentEvent.java → samoa-api/src/main/java/org/apache/samoa/learners/ClassificationModelContentEvent.java

@@ -21,25 +21,26 @@
  */
 
 import org.apache.samoa.core.ContentEvent;
+import org.apache.samoa.learners.classifiers.ClassificationModel;
 
-final public class ModelContentEvent implements ContentEvent {
+final public class ClassificationModelContentEvent implements ContentEvent {
     final private boolean isLast;
-    private Model model;
+    private ClassificationModel model;
     private long modelIndex;
     private long instanceIndex;
     private int classifierIndex;
     private int evaluationIndex;
 
-    public ModelContentEvent() {
+    public ClassificationModelContentEvent() {
         this.isLast = false;
     }
 
-    public ModelContentEvent(boolean isLast) {
+    public ClassificationModelContentEvent(boolean isLast) {
         this.isLast = isLast;
     }
 
-    public ModelContentEvent(boolean isLast, Model model, long modelIndex, long instanceIndex,
-                             int classifierIndex, int evaluationIndex) {
+    public ClassificationModelContentEvent(boolean isLast, ClassificationModel model, long modelIndex, long instanceIndex,
+                                           int classifierIndex, int evaluationIndex) {
         this.isLast = isLast;
         this.model = model;
         this.modelIndex = modelIndex;
@@ -71,11 +72,11 @@ public void setModelIndex(long modelIndex) {
         this.modelIndex = modelIndex;
     }
 
-    public Model getModel() {
+    public ClassificationModel getModel() {
         return model;
     }
 
-    public void setModel(Model model) {
+    public void setModel(ClassificationModel model) {
         this.model = model;
     }
 

samoa-api/src/main/java/org/apache/samoa/learners/InstanceUtils.java

@@ -0,0 +1,142 @@
+package org.apache.samoa.learners;
+
+/*
+ * #%L
+ * SAMOA
+ * %%
+ * Copyright (C) 2014 - 2016 Apache Software Foundation
+ * %%
+ * Licensed 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.
+ * #L%
+ */
+
+import org.apache.samoa.instances.*;
+import org.apache.samoa.learners.classifiers.ClassificationDataInstance;
+import org.apache.samoa.learners.clusterers.ClusterDataInstance;
+import org.apache.samoa.moa.core.DataPoint;
+import org.apache.samoa.moa.core.FastVector;
+
+import java.util.ArrayList;
+import java.util.Arrays;
+
+public class InstanceUtils {
+
+    private static InstancesHeader getClusterInstanceHeader(ClusterDataInstance dataInstance) {
+        ArrayList<Attribute> attributes = new ArrayList<>();
+
+        for (int i = 0; i < dataInstance.getNumberFeatures(); i++) {
+            attributes.add(new Attribute("att" + (i + 1)));
+        }
+
+        // attributes.add(new Attribute("class", null));
+
+        InstancesHeader instancesHeader = new InstancesHeader(
+                new Instances(null, attributes, 0));
+        instancesHeader.setClassIndex(instancesHeader.numAttributes() - 1);
+
+        return instancesHeader;
+    }
+
+    private static InstancesHeader getClassificationInstanceHeader(ClassificationDataInstance dataInstance) {
+        FastVector<Attribute> attributes = new FastVector<>();
+
+        for (int i = 0; i < dataInstance.getNumberNominalFeatures(); i++) {
+            FastVector<String> nominalAttVals = new FastVector<>();
+            for (int j = 0; j < dataInstance.getNumberValsPerNominalFeature()[i]; j++) {
+                nominalAttVals.addElement("value" + (j + 1));
+            }
+            attributes.addElement(new Attribute("nominal" + (i + 1),
+                    nominalAttVals));
+        }
+
+        for (int i = 0; i < dataInstance.getNumberNumericFeatures(); i++) {
+            attributes.addElement(new Attribute("numeric" + (i + 1)));
+        }
+
+        FastVector<String> classLabels = new FastVector<>();
+        for (int i = 0; i < dataInstance.getNumberLabels(); i++) {
+            classLabels.addElement("class" + (i + 1));
+        }
+        attributes.addElement(new Attribute("class", classLabels));
+
+        InstancesHeader instancesHeader = new InstancesHeader(
+                new Instances(null, attributes, 0));
+        instancesHeader.setClassIndex(instancesHeader.numAttributes() - 1);
+
+        return instancesHeader;
+    }
+
+    /**
+     * convert ClassificationDataInstance to SAMOA Instance
+     */
+    public static Instance convertClassificationDataInstance(ClassificationDataInstance dataInstance) {
+        InstancesHeader header = InstanceUtils.getClassificationInstanceHeader(dataInstance);
+        Instance inst = new DenseInstance(header.numAttributes());
+
+        int numNomFeatures = dataInstance.getNumberNominalFeatures();
+        int numNumFeatures = dataInstance.getNumberNumericFeatures();
+
+        for (int i = 0; i < numNomFeatures + numNumFeatures; i++) {
+            if (i < numNomFeatures) {
+                inst.setValue(i, dataInstance.getNominalData()[i]);
+            } else {
+                inst.setValue(i, dataInstance.getNumericData()[i - numNomFeatures]);
+            }
+        }
+
+        inst.setDataset(header);
+        inst.setClassValue(dataInstance.getTrueLabel());
+
+        return inst;
+    }
+
+    /**
+     * convert SAMOA Instance to ClassificationDataInstance
+     */
+    public static ClassificationDataInstance reConvertClassificationDataInstance(
+            Instance inst, int numberNominalFeatures, int numberNumericFeatures) {
+        double[] nominalDataTmp = Arrays.copyOfRange(inst.toDoubleArray(), 0, numberNominalFeatures);
+        int[] nominalData = new int[nominalDataTmp.length];
+        for (int j = 0; j < nominalData.length; j++) {
+            nominalData[j] = (int) nominalDataTmp[j];
+        }
+
+        double[] numericData = Arrays.copyOfRange(
+                inst.toDoubleArray(), numberNominalFeatures,
+                inst.toDoubleArray().length - 1);
+
+        int[] numValsPerNominal = new int[nominalData.length];
+        Arrays.fill(numValsPerNominal, inst.attribute(0).numValues());
+
+        return new ClassificationDataInstance(
+                numberNumericFeatures, numericData, numberNominalFeatures,
+                numValsPerNominal, nominalData, inst.numClasses(), (int) inst.classValue());
+    }
+
+    /**
+     * convert ClusterDataInstance to SAMOA Instance
+     */
+    public static Instance convertClusterDataInstance(ClusterDataInstance dataInstance) {
+        Instance inst = new DenseInstance(1.0, dataInstance.getData());
+        inst.setDataset(InstanceUtils.getClusterInstanceHeader(dataInstance));
+        return new DataPoint(inst, dataInstance.getTimeStamp());
+    }
+
+    /**
+     * convert SAMOA Instance to ClusterDataInstance
+     */
+    public static ClusterDataInstance reConvertClusterDataInstance(DataPoint point) {
+        double[] data = point.toDoubleArray();
+        return new ClusterDataInstance(data.length, point.getTimestamp(), data);
+    }
+}

samoa-api/src/main/java/org/apache/samoa/learners/classifiers/ClassificationDataInstance.java

@@ -0,0 +1,80 @@
+package org.apache.samoa.learners.classifiers;
+
+/*
+ * #%L
+ * SAMOA
+ * %%
+ * Copyright (C) 2014 - 2016 Apache Software Foundation
+ * %%
+ * Licensed 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.
+ * #L%
+ */
+
+import java.io.Serializable;
+
+/**
+ * DataInstance for classification problem
+ * There may be tow types of feature in feature vector: numeric feature and nominal feature
+ */
+public class ClassificationDataInstance implements Serializable {
+    private int numberNumericFeatures;
+    private double[] numericData;
+
+    private int numberNominalFeatures;
+    private int[] numberValsPerNominalFeature;
+    private int[] nominalData;
+
+    private int numberLabels;
+    private int trueLabel;
+
+    public ClassificationDataInstance(int numberNumericFeatures,
+                                      double[] numericData, int numberNominalFeatures,
+                                      int[] numberValsPerNominalFeature, int[] nominalData,
+                                      int numberLabels, int trueLabel) {
+        this.numberNumericFeatures = numberNumericFeatures;
+        this.numericData = numericData;
+        this.numberNominalFeatures = numberNominalFeatures;
+        this.numberValsPerNominalFeature = numberValsPerNominalFeature;
+        this.nominalData = nominalData;
+        this.numberLabels = numberLabels;
+        this.trueLabel = trueLabel;
+    }
+
+    public int getNumberNumericFeatures() {
+        return numberNumericFeatures;
+    }
+
+    public double[] getNumericData() {
+        return numericData;
+    }
+
+    public int getNumberNominalFeatures() {
+        return numberNominalFeatures;
+    }
+
+    public int[] getNumberValsPerNominalFeature() {
+        return numberValsPerNominalFeature;
+    }
+
+    public int[] getNominalData() {
+        return nominalData;
+    }
+
+    public int getNumberLabels() {
+        return numberLabels;
+    }
+
+    public int getTrueLabel() {
+        return trueLabel;
+    }
+}

samoa-api/src/main/java/org/apache/samoa/learners/classifiers/ClassificationModel.java

@@ -0,0 +1,30 @@
+package org.apache.samoa.learners.classifiers;
+
+/*
+ * #%L
+ * SAMOA
+ * %%
+ * Copyright (C) 2014 - 2016 Apache Software Foundation
+ * %%
+ * Licensed 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.
+ * #L%
+ */
+
+import java.io.Serializable;
+
+/**
+ * Model for classification problem
+ */
+public interface ClassificationModel extends Serializable {
+    double[] predict(ClassificationDataInstance dataInstance);
+}

samoa-api/src/main/java/org/apache/samoa/learners/classifiers/ensemble/BoostingPredictionCombinerProcessor.java

@@ -30,7 +30,7 @@
 import org.apache.samoa.core.ContentEvent;
 import org.apache.samoa.instances.Instance;
 import org.apache.samoa.learners.InstanceContentEvent;
-import org.apache.samoa.learners.ModelContentEvent;
+import org.apache.samoa.learners.ClassificationModelContentEvent;
 import org.apache.samoa.learners.ResultContentEvent;
 import org.apache.samoa.moa.core.DoubleVector;
 import org.apache.samoa.moa.core.Utils;
@@ -59,8 +59,8 @@ public class BoostingPredictionCombinerProcessor extends PredictionCombinerProce
   @Override
   public boolean process(ContentEvent event) {
     // for serialize
-    if (event instanceof ModelContentEvent) {
-      return this.processModel((ModelContentEvent) event);
+    if (event instanceof ClassificationModelContentEvent) {
+      return this.processModel((ClassificationModelContentEvent) event);
     }
 
     ResultContentEvent inEvent = (ResultContentEvent) event;