Implementing a data selector for the deep learning module

src/export/swing/ExportProcessDialog.java

@@ -76,7 +76,6 @@ public class ExportProcessDialog {
 
 
 	public void createExportGroup() {
-
 		//clear current tasks. 
 		dlOfflineGroup.clearTasks();
 
@@ -88,7 +87,6 @@ public class ExportProcessDialog {
 				dlOfflineGroup.addTask(new ExportTask(dataBlocks.get(i), exportManager));
 			}
 		}
-		
 	}
 	////---Swing stuff----/// should not be here but this is how PG works. 
 
@@ -388,7 +386,7 @@ public class ExportProcessDialog {
 
 
 	/**
-	 * Export taks group. 
+	 * Export task
 	 */
 	class ExportTaskGroup extends OfflineTaskGroup {
 

src/rawDeepLearningClassifier/DLControl.java

@@ -622,8 +622,9 @@ public class DLControl extends PamControlledUnit implements PamSettings {
 	}
 	
 	/**
-	 * Get the classifier chooser. 
-	 * @return the classifier chooser.Take it
+	 * Get the classifier chooser. The classifier chooser chooses which classifier use
+	 * based on a selected file or URL. 
+	 * @return the classifier chooser.
 	 */
 	public DLClassifierChooser getDlClassifierChooser() {
 		return dlClassifierChooser;

src/rawDeepLearningClassifier/dataSelector/DLDataFilter.java

@@ -0,0 +1,35 @@
+package rawDeepLearningClassifier.dataSelector;
+
+import PamguardMVC.PamDataUnit;
+import PamguardMVC.dataSelector.DataSelectParams;
+
+/**
+ * Score a data unit with a deep learning annotation. Note that this could be an
+ * external data unit with an annotation e.g. a click, or deep learning
+ * detections generated by the deep learning module.
+ * 
+ * @author Jamie Macaulay
+ */
+public interface DLDataFilter {
+	
+	/**
+	 * Score a data unit with deep learning annotations 
+	 * @param dataUnit - the data unit to score
+	 * @return 0 to reject and >0 to accept. 
+	 */
+	public int scoreDLData(PamDataUnit dataUnit);
+	
+	/**
+	 * Get parameters for the data filter. 
+	 * @return parameters for the data selectors. 
+	 */
+	public DataSelectParams getParams(); 
+	
+	/**
+	 * Set the parameters for the data filter. 
+	 * @param params - the parameters to set. 
+	 */
+	public void setParams(DataSelectParams params); 
+
+
+}

src/rawDeepLearningClassifier/dataSelector/DLDataSelectCreator.java

@@ -0,0 +1,34 @@
+package rawDeepLearningClassifier.dataSelector;
+
+import PamguardMVC.PamDataBlock;
+import PamguardMVC.dataSelector.DataSelectParams;
+import PamguardMVC.dataSelector.DataSelector;
+import PamguardMVC.dataSelector.DataSelectorCreator;
+import clickDetector.dataSelector.ClickDataSelector;
+import rawDeepLearningClassifier.DLControl;
+
+/**
+ * Creates a data selector for the deep learning module. 
+ * 
+ * @author Jamie Macaulay
+ */
+public class DLDataSelectCreator extends DataSelectorCreator  {
+
+	private DLControl dlcontrol;
+
+	public DLDataSelectCreator(DLControl dlcontrol, PamDataBlock pamDataBlock) {
+		super(pamDataBlock);
+		this.dlcontrol = dlcontrol; 
+	}
+
+	@Override
+	public DataSelectParams createNewParams(String name) {
+		return new DLDataSelectorParams();
+	}
+
+	@Override
+	public DataSelector createDataSelector(String selectorName, boolean allowScores, String selectorType) {
+		return new DLDataSelector(dlcontrol, this.getPamDataBlock(), selectorName, allowScores, selectorType);
+	}
+
+}

src/rawDeepLearningClassifier/dataSelector/DLDataSelector.java

@@ -0,0 +1,120 @@
+package rawDeepLearningClassifier.dataSelector;
+
+import java.util.ArrayList;
+import java.util.List;
+
+import PamView.dialog.PamDialogPanel;
+import PamguardMVC.PamDataBlock;
+import PamguardMVC.PamDataUnit;
+import PamguardMVC.dataSelector.DataSelectParams;
+import PamguardMVC.dataSelector.DataSelector;
+import pamViewFX.fxSettingsPanes.DynamicSettingsPane;
+import rawDeepLearningClassifier.DLControl;
+
+/**
+ * Data selector of DL data units. Note that data selectors are really data 
+ * unit specific and not annotation specific. 
+ * <p>
+ * The data selector can have different types of data selectors which can 
+ * depend on the classifer used and user choice. 
+ * <p>
+ * Note that this is slightly different from DLPredicitoDecision
+ * as it deals with data units that may have a more than one prediction. 
+ * i.e. 
+ */
+public class DLDataSelector extends DataSelector {
+
+
+	/**
+	 * Swing panel for the data selector. 
+	 */
+	private DLSelectPanel dlSelectPanel;
+
+
+	/**
+	 * FX panel for the data selector. 
+	 */
+	private DLSelectPaneFX dlSelectPaneFX;
+
+	/**
+	 * Data filter for filtering. 
+	 */
+	private List<DLDataFilter> dataFilters = new ArrayList<DLDataFilter> ();
+
+
+	private DLDataSelectorParams dlDataSelectParams;
+
+	/**
+	 * Constructor for the data selector. 
+	 * @param dlcontrol - reference to the DLControl.
+	 * @param pamDataBlock - the data block. 
+	 * @param selectorName - the selector name. 
+	 * @param allowScores - allow all the scores. 
+	 * @param selectorType - the selector type. 
+	 */
+	public DLDataSelector(DLControl dlcontrol, PamDataBlock pamDataBlock, String selectorName, boolean allowScores, String selectorType) {
+		super(pamDataBlock, selectorName, allowScores);
+		/****New data filters go here****/
+		dataFilters.add(new DLPredictionFilter(dlcontrol)); 
+	}
+
+	@Override
+	public PamDialogPanel getDialogPanel() {
+		if (dlSelectPanel == null) {
+			dlSelectPanel = new DLSelectPanel(this);
+		}
+		return dlSelectPanel;
+	}
+	
+	@Override
+	public DynamicSettingsPane<Boolean> getDialogPaneFX() {
+		if (dlSelectPaneFX == null) {
+			dlSelectPaneFX = new DLSelectPaneFX(this);
+		}
+		return dlSelectPaneFX;
+	}
+
+	@Override
+	public void setParams(DataSelectParams dataSelectParams) {
+		
+		if (dataSelectParams instanceof DLDataSelectorParams) {
+			dlDataSelectParams = (DLDataSelectorParams) dataSelectParams; 
+
+			checkDataFilterParamsSize(dlDataSelectParams); 
+
+			//set paramters for all data filters. 
+			for (int i=0; i<dlDataSelectParams.dataSelectorParams.length; i++) {
+				dataFilters.get(i).setParams((DLDataSelectorParams) dataSelectParams); 
+			}
+		}
+
+	}
+
+	/**
+	 * Check that the data filters are the corret size.
+	 * @param dataSelectParams - the apramters to set. 
+	 */
+	private void checkDataFilterParamsSize(DLDataSelectorParams dataSelectParams) {
+		if (dataSelectParams.dataSelectorParams==null || dataSelectParams.dataSelectorParams.length!=dataFilters.size()) {
+			dataSelectParams.dataSelectorParams = new DataSelectParams[dataFilters.size()]; 
+			for (int i=0; i<dataSelectParams.dataSelectorParams.length; i++) {
+				dataSelectParams.dataSelectorParams[i] = dataFilters.get(i).getParams();
+			}
+		}
+	}
+
+	@Override
+	public DataSelectParams getParams() {
+		for (int i=0; i<dlDataSelectParams.dataSelectorParams.length; i++) {
+			dlDataSelectParams.dataSelectorParams[i] = dataFilters.get(i).getParams();
+		}
+		return dlDataSelectParams;
+	}
+
+	@Override
+	public double scoreData(PamDataUnit pamDataUnit) {
+		int score = dataFilters.get(dlDataSelectParams.dataSelectorIndex).scoreDLData(pamDataUnit);
+		return score>=0 ? 1 : 0;
+	}
+
+}

src/rawDeepLearningClassifier/dataSelector/DLDataSelectorParams.java

@@ -0,0 +1,22 @@
+package rawDeepLearningClassifier.dataSelector;
+
+import java.io.Serializable;
+
+import PamguardMVC.dataSelector.DataSelectParams;
+
+/**
+ * Paramters for the DL data seelctor	
+ */
+public class DLDataSelectorParams extends DataSelectParams implements Serializable, Cloneable{
+	
+	private static final long serialVersionUID = 1L;
+
+	public int dataSelectorIndex = 0;
+	
+	/**
+	 * List of data selector parameters for different deep learning selectors. 
+	 */
+	public DataSelectParams[] dataSelectorParams;
+	
+	
+}

src/rawDeepLearningClassifier/dataSelector/DLPredictionFilter.java

@@ -0,0 +1,107 @@
+package rawDeepLearningClassifier.dataSelector;
+
+
+import PamguardMVC.PamDataUnit;
+import PamguardMVC.dataSelector.DataSelectParams;
+import rawDeepLearningClassifier.DLControl;
+import rawDeepLearningClassifier.dlClassification.DLDetection;
+import rawDeepLearningClassifier.dlClassification.PredictionResult;
+
+/**
+ * A data filter which filters data by the maximum prediction value
+ * for different classes. 
+ * 
+ * @author Jamie Macaulay
+ */
+public class DLPredictionFilter implements DLDataFilter {
+
+	/**
+	 * Reference to the DLControl
+	 */
+	private DLControl dlcontrol;
+	
+	/**
+	 * The filter parameters
+	 */
+	private DLPredictionFilterParams filterParams = new DLPredictionFilterParams(); 
+
+	public DLPredictionFilter(DLControl dlcontrol) {
+		this.dlcontrol = dlcontrol; 
+		checkParamsClass() ;
+	}
+	
+	
+	@Override
+	public int scoreDLData(PamDataUnit dataUnit) {
+
+		DLDetection dlDetection = (DLDetection) dataUnit; 
+	
+		//iterate through all results and check that at least one class passes data selection. 
+		float[] results;
+		int maxClassIndex = -1; 
+		int maxPred = -1; 
+
+		//get the maximum prediction index which passes the minimum threshold
+		for (PredictionResult modelResult: dlDetection.getModelResults()) {
+			results = modelResult.getPrediction();
+			for (int j=0; j<results.length; j++) {
+				if (filterParams.classSelect[j] && results[j]>filterParams.minClassPredicton[j]) {
+					if (results[j]> maxPred) maxClassIndex = j;
+				}
+			}
+		}
+		
+		return maxClassIndex;
+	}
+
+	@Override
+	public DataSelectParams getParams() {
+		 checkParamsClass() ;
+		return filterParams;
+	}
+
+	@Override
+	public void setParams(DataSelectParams params) {
+		this.filterParams = ((DLPredictionFilterParams) params).clone();
+		 checkParamsClass();
+		
+	}
+	
+	private void checkParamsClass() {
+		if (filterParams.classSelect==null ||  dlcontrol.getNumClasses()!=filterParams.classSelect.length) {
+			filterParams.classSelect = new boolean[dlcontrol.getNumClasses()];
+			filterParams.minClassPredicton = new double[dlcontrol.getNumClasses()];
+			
+			//set default so data selector does not always prevent all classes from showing. 
+			for (int i=0; i<filterParams.classSelect.length; i++) {
+				filterParams.classSelect[i]=true;
+				filterParams.minClassPredicton[i]=0.4;
+			}
+
+		}
+	}
+	
+//	/**
+//	 * Get the index of the highest prediciton value a list of results. 
+//	 * @param predictions - index of the highest prediction within a matrix of predicitons. 
+//	 * @return an index of the hghest predictions. 
+//	 */
+//	public static int[] getBestClass(List<PredictionResult> predictions) {
+//				
+//		float[][] results = new float[predictions.size()][]; 
+//		
+//		//A detection might have multiple prediction results, i.e. predictions are a matrix. Need 
+//		//to iterate through all the predictions and then work out whihc is the maximum. That index is then then]
+//		//class colour. 
+//		int i=0;
+//		for (PredictionResult modelResult: predictions) {
+//			results[i] = modelResult.getPrediction();
+//			i++;
+//		}
+//		
+//		int[] indexBest = PamArrayUtils.maxPos(results); 
+//		
+//		return indexBest;
+//	}
+
+}

src/rawDeepLearningClassifier/dataSelector/DLPredictionFilterParams.java

@@ -0,0 +1,35 @@
+package rawDeepLearningClassifier.dataSelector;
+
+import PamguardMVC.dataSelector.DataSelectParams;
+
+/**
+ * Parameters for filtering by the minimum prediciton value. 
+ */
+public class DLPredictionFilterParams extends DataSelectParams {
+
+	private static final long serialVersionUID = 1L;
+	
+	/**
+	 * An array indicating which classes are to be used in data selection
+	 */
+	public boolean[] classSelect; 
+	
+	/**
+	 * The minimum class prediction. 
+	 */
+	public double[] minClassPredicton;
+	
+	/**
+	 * Clone the parameters. 
+	 */
+	public DLPredictionFilterParams clone() {
+		try {
+			return (DLPredictionFilterParams) super.clone();
+		} catch (CloneNotSupportedException e) {
+			// TODO Auto-generated catch block
+			e.printStackTrace();
+		}
+		return null;
+	}
+
+}

src/rawDeepLearningClassifier/dataSelector/DLSelectPaneFX.java

@@ -0,0 +1,47 @@
+package rawDeepLearningClassifier.dataSelector;
+
+import javafx.scene.Node;
+import javafx.scene.control.Label;
+import pamViewFX.fxSettingsPanes.DynamicSettingsPane;
+
+/**
+ * JavaFX pane for the deep learning data selector. 
+ * 
+ * @author Jamie Macaulay
+ */
+public class DLSelectPaneFX extends DynamicSettingsPane<Boolean>{
+
+	public DLSelectPaneFX(Object ownerWindow) {
+		super(ownerWindow);
+		// TODO Auto-generated constructor stub
+	}
+
+	@Override
+	public Boolean getParams(Boolean currParams) {
+		// TODO Auto-generated method stub
+		return currParams;
+	}
+
+	@Override
+	public void setParams(Boolean input) {
+		// TODO Auto-generated method stub
+		
+	}
+
+	@Override
+	public String getName() {
+		return "Deep Learning Data Selector:";
+	}
+
+	@Override
+	public Node getContentNode() {
+		return new Label("Hello data selector");
+	}
+
+	@Override
+	public void paneInitialized() {
+		// TODO Auto-generated method stub
+		
+	}
+
+}

src/rawDeepLearningClassifier/dataSelector/DLSelectPanel.java

@@ -0,0 +1,34 @@
+package rawDeepLearningClassifier.dataSelector;
+
+import javax.swing.JComponent;
+
+import PamView.dialog.PamDialogPanel;
+
+/**
+ * Swing panel for the deep learning data selector. 
+ */
+public class DLSelectPanel implements PamDialogPanel {
+
+	public DLSelectPanel(DLDataSelector dlDataSelector) {
+		// TODO Auto-generated constructor stub
+	}
+
+	@Override
+	public JComponent getDialogComponent() {
+		// TODO Auto-generated method stub
+		return null;
+	}
+
+	@Override
+	public void setParams() {
+		// TODO Auto-generated method stub
+		
+	}
+
+	@Override
+	public boolean getParams() {
+		// TODO Auto-generated method stub
+		return false;
+	}
+
+}

src/rawDeepLearningClassifier/dlClassification/DLClassiferModel.java

@@ -115,7 +115,7 @@ public interface DLClassiferModel {
 	
 	/**
 	 * Get the parameters which can be serialized  from  transforms. 
-	 * @param dlTransfroms- the dl transforms. 
+	 * @param dlTransfroms- the DL transforms. 
 	 */
 	public static ArrayList<DLTransfromParams> getDLTransformParams(ArrayList<DLTransform> dlTransfroms) {
 		ArrayList<DLTransfromParams> dlTransformParams = new ArrayList<DLTransfromParams>(); 

src/rawDeepLearningClassifier/dlClassification/DLClassifyProcess.java

@@ -415,7 +415,7 @@ public class DLClassifyProcess extends PamInstantProcess {
 				else {
 					/****Add annotation to existing data unit (e.g. click, clip or other RawDataHolder)****/
 					//Need to go by the parent data unit for merging data not the segments. Note that we may still add multiple
-					//predicitions to a single data unit depending on how many segments it contains. 
+					//predictions to a single data unit depending on how many segments it contains. 
 
 					//System.out.println("New model data " + pamRawData.getParentDataUnit().getUID() + " " + groupDataBuffer[i].size() + " " + modelResultDataBuffer[i].size()); 
 

src/rawDeepLearningClassifier/dlClassification/DLDetectionDataBlock.java

@@ -3,7 +3,10 @@ package rawDeepLearningClassifier.dlClassification;
 import PamView.GroupedDataSource;
 import PamView.GroupedSourceParameters;
 import PamguardMVC.AcousticDataBlock;
+import PamguardMVC.dataSelector.DataSelectorCreator;
+import clickTrainDetector.dataselector.CTDataSelectCreator;
 import rawDeepLearningClassifier.DLControl;
+import rawDeepLearningClassifier.dataSelector.DLDataSelectCreator;
 import rawDeepLearningClassifier.tethys.DLSpeciesManager;
 import rawDeepLearningClassifier.tethys.DLTethysDataProvider;
 import tethys.TethysControl;
@@ -18,11 +21,31 @@ import tethys.species.DataBlockSpeciesManager;
  */
 public class DLDetectionDataBlock extends AcousticDataBlock<DLDetection> implements GroupedDataSource {
 
+	/**
+	 * Reference to the deep learning classifier process. 
+	 */
 	private DLClassifyProcess dlClassifyProcess;
+	
+	/**
+	 * Reference to the Tethys data provider for the deep learning module (provides standardised metadata). 
+	 */
 	private DLTethysDataProvider dlTethysDataProvider;
+	
+	/**
+	 * A species manager for metadata. 
+	 */
 	private DLSpeciesManager dlSpeciesManager;
+	
+	/**
+	 * Reference to the deep learning control. 
+	 */
 	private DLControl dlControl;
 	
+	/**
+	 * Data selector for the deep learning detections. 
+	 */
+	private DLDataSelectCreator dlDataSelectCreator;
+
 	public DLDetectionDataBlock(String dataName, DLClassifyProcess parentProcess, int channelMap) {
 		super(DLDetection.class, dataName, parentProcess, channelMap);
 		this.dlClassifyProcess = parentProcess; 
@@ -51,4 +74,17 @@ public class DLDetectionDataBlock extends AcousticDataBlock<DLDetection> impleme
 	}
 	
 	
+	/* (non-Javadoc)
+	 * @see PamguardMVC.PamDataBlock#getDataSelectCreator()
+	 */
+	@Override
+	public synchronized  DataSelectorCreator getDataSelectCreator() {
+		if (dlDataSelectCreator == null) {
+			dlDataSelectCreator = new DLDataSelectCreator(dlControl, this);
+		}
+		return dlDataSelectCreator;
+	
+	}
+
+
 }

src/rawDeepLearningClassifier/dlClassification/DLPredictionDecision.java

@@ -0,0 +1,26 @@
+package rawDeepLearningClassifier.dlClassification;
+
+/**
+ * Decides whether a prediction has passed a threshold to be used to create a new
+ * data unit. 
+ * <p>
+ * Note that the majority of the time this will be a simple test of the value of 
+ * predictions of a model but there will be cases when a classifier implements a 
+ * more complex system. For example, a implementation could save a buffer of predictions 
+ * so that previous predictions inform the latest prediction. Or results may include 
+ * some sort of object detection components and frequency bounds etc could be used for 
+ * classification. 
+ * 
+ * 
+ */
+public interface DLPredictionDecision {
+	
+	/**
+	 * Check whether a single prediction passes a binary classifier. Prediction which pass decision will be
+	 * passed on to create new data units. 
+	 * @param result - the prediciton result to test. 
+	 * @return true if the result is passed. 
+	 */
+	public boolean isBinaryResult(PredictionResult result); 
+
+}

src/rawDeepLearningClassifier/dlClassification/SimpleDLDecision.java

@@ -0,0 +1,49 @@
+package rawDeepLearningClassifier.dlClassification;
+
+import rawDeepLearningClassifier.dlClassification.animalSpot.StandardModelParams;
+
+/**
+ * Make a decision based on a simple binary threshold for a prediction. 
+ */
+public class SimpleDLDecision implements DLPredictionDecision {
+	
+	/**
+	 * Reference to the parameters. 
+	 */
+	private StandardModelParams params;
+
+
+	@Override
+	public boolean isBinaryResult(PredictionResult modelResult) {
+		return  isBinaryResult(modelResult, getParams()) ;
+	}
+	
+	
+	/**
+	 * Check whether a model passes a binary test...
+	 * @param modelResult - the model results
+	 * @return the model results. 
+	 */
+	private static boolean isBinaryResult(PredictionResult modelResult, StandardModelParams genericModelParams) {
+		for (int i=0; i<modelResult.getPrediction().length; i++) {
+						//System.out.println("Binary Classification: "  + genericModelParams.binaryClassification.length); 
+
+			if (modelResult.getPrediction()[i]>genericModelParams.threshold && genericModelParams.binaryClassification[i]) {
+				//				System.out.println("SoundSpotClassifier: prediciton: " + i + " passed threshold with val: " + modelResult.getPrediction()[i]); 
+				return true; 
+			}
+		}
+		return  false;
+	}
+	
+	
+	public StandardModelParams getParams() {
+		return params;
+	}
+	
+	
+	public void setParams( StandardModelParams genericModelParams) {
+		this.params = genericModelParams;
+	}
+
+}

src/rawDeepLearningClassifier/dlClassification/StandardClassifierModel.java

@@ -45,6 +45,12 @@ public abstract class StandardClassifierModel implements DLClassiferModel, PamSe
 	 */
 	private TaskThread workerThread;
 	
+	/**
+	 * Makes a binary decision on whether a prediction result should go on
+	 * to be part of a data unit. 
+	 */
+	private SimpleDLDecision simpleDLDecision = new SimpleDLDecision();
+
 
 	public StandardClassifierModel(DLControl dlControl) {
 		this.dlControl=dlControl; 
@@ -84,7 +90,7 @@ public abstract class StandardClassifierModel implements DLClassiferModel, PamSe
 			
 			for (int i =0; i<modelResult.size(); i++) {
 				modelResult.get(i).setClassNameID(GenericDLClassifier.getClassNameIDs(getDLParams())); 
-				modelResult.get(i).setBinaryClassification(GenericDLClassifier.isBinaryResult(modelResult.get(i), getDLParams())); 
+				modelResult.get(i).setBinaryClassification(isDecision(modelResult.get(i), getDLParams())); 
 				modelResult.get(i).setTimeMillis(groupedRawData.get(i).getTimeMilliseconds());
 
 			}
@@ -106,7 +112,7 @@ public abstract class StandardClassifierModel implements DLClassiferModel, PamSe
 	
 	@Override
 	public void prepModel() {
-		System.out.println("STANDARD CLASSIFIER MODEL PREP MODEL! !!!: " +  getDLParams().modelPath);
+//		System.out.println("STANDARD CLASSIFIER MODEL PREP MODEL! !!!: " +  getDLParams().modelPath);
 //		StandardModelParams oldParams = getDLParams().clone();
 		
 		getDLWorker().prepModel(getDLParams(), dlControl);
@@ -221,7 +227,8 @@ public abstract class StandardClassifierModel implements DLClassiferModel, PamSe
 	 * @return true if a threshold has been met. 
 	 */
 	public boolean isDecision(StandardPrediction modelResult, StandardModelParams modelParmas) {
-		return isBinaryResult(modelResult, modelParmas);
+		simpleDLDecision.setParams(modelParmas);
+		return simpleDLDecision.isBinaryResult(modelResult);
 	}
 
 	
@@ -241,22 +248,22 @@ public abstract class StandardClassifierModel implements DLClassiferModel, PamSe
 
 
 
-	/**
-	 * Check whether a model passes a binary test...
-	 * @param modelResult - the model results
-	 * @return the model results. 
-	 */
-	public static boolean isBinaryResult(StandardPrediction modelResult, StandardModelParams genericModelParams) {
-		for (int i=0; i<modelResult.getPrediction().length; i++) {
-						//System.out.println("Binary Classification: "  + genericModelParams.binaryClassification.length); 
-
-			if (modelResult.getPrediction()[i]>genericModelParams.threshold && genericModelParams.binaryClassification[i]) {
-				//				System.out.println("SoundSpotClassifier: prediciton: " + i + " passed threshold with val: " + modelResult.getPrediction()[i]); 
-				return true; 
-			}
-		}
-		return  false;
-	}
+//	/**
+//	 * Check whether a model passes a binary test...
+//	 * @param modelResult - the model results
+//	 * @return the model results. 
+//	 */
+//	public static boolean isBinaryResult(StandardPrediction modelResult, StandardModelParams genericModelParams) {
+//		for (int i=0; i<modelResult.getPrediction().length; i++) {
+//						//System.out.println("Binary Classification: "  + genericModelParams.binaryClassification.length); 
+//
+//			if (modelResult.getPrediction()[i]>genericModelParams.threshold && genericModelParams.binaryClassification[i]) {
+//				//				System.out.println("SoundSpotClassifier: prediciton: " + i + " passed threshold with val: " + modelResult.getPrediction()[i]); 
+//				return true; 
+//			}
+//		}
+//		return  false;
+//	}
 	
 	
 	@Override