nordic-dev.net/PAMGuard
nordic-dev.net/PAMGuard
2
0
Fork 0
mirror of https://github.com/PAMGuard/PAMGuard.git synced 2024-11-21 14:42:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

Implement zip model for deep learning classifier.

Browse Source 
Implement zip models which allow packaging of deep learning models with metadata\
Fix TDdsiplayFX spinner bug
Bug fixes for deep learning module after testing
Implement a validator to warn if users have selected all or no classes for binary classifcation.
This commit is contained in:
Jamie Mac 2024-03-01 15:23:16 +00:00
parent 98e7d55034
commit 45bf890a62
25 changed files with 248 additions and 202 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

BIN
nullPamguardSettings_20171106_185953.psfx Normal file
View File

Binary file not shown.

17
pom.xml
View File

@ -240,23 +240,6 @@
<url>file://${project.basedir}/repo</url>
</repository>
<!-- Repo for Jamie's code -->
<!-- IMPORTANT: SOMEHOW THIS REPO INTERFERES WITH THE
JAVAFX DEPENDENCIES. Don't understand how, but it
does. If Maven is not downloading the JavaFX dependencies
and is instead throwing errors, try commenting out this
repository. Let Maven download the JavaFX files, and
then uncomment this repo again.
Same thing seems to be occurring if I try to update the postgresql library. -->
<!-- <repository> -->
<!-- <snapshots> -->
<!-- <enabled>false</enabled> -->
<!-- </snapshots> -->
<!-- <id>central</id> -->
<!-- <name>a0u0ltozdsehx-artifactory-primary-0-releases</name> -->
<!-- <url>https://jmachund.jfrog.io/artifactory/jpam2</url> -->
<!-- </repository> -->
<!-- Repo for netCDF -->
<repository>
<snapshots>

2
src/dataPlotsFX/TDControlFX.java
View File

@ -115,12 +115,12 @@ public class TDControlFX extends TDControl implements UserDisplayNodeFX {
System.out.println("TDControldFX: parent datablock null");
return dataBlocks;
}
for (int i=0; i<tdDisplayController.getUserDisplayProcess().getNumMuiltiplexDataBlocks(); i++){
dataBlock=this.tdDisplayController.getUserDisplayProcess().getMuiltiplexDataBlock(i);
if (TDDataProviderRegisterFX.getInstance().findDataProvider(dataBlock)!=null){
dataBlocks.add(dataBlock);
}
// System.out.println("TDControldFX: parent data block "+dataBlock.getDataName());
}
return dataBlocks;
}

2
src/dataPlotsFX/clickPlotFX/ClickPlotInfoFX.java
View File

@ -517,7 +517,7 @@ public class ClickPlotInfoFX extends TDDataInfoFX {
if (pamDataUnit.getSuperDetectionsCount()==0) return null;
ListIterator<PamDataUnit> iterator = this.getPamDataBlock().getListIterator(PamDataBlock.ITERATOR_END);
ListIterator<PamDataUnit> iterator = this.getDataBlock().getListIterator(PamDataBlock.ITERATOR_END);
PamDataUnit testDataUnit;
while (iterator.hasPrevious()) {

7
src/dataPlotsFX/data/TDDataInfoFX.java
View File

@ -731,13 +731,6 @@ public abstract class TDDataInfoFX {
return false;
}
/**
* @return the pamDataBlock
*/
protected PamDataBlock getPamDataBlock() {
return pamDataBlock;
}
/**
* @return the tdDataProvider
*/

1
src/dataPlotsFX/data/TDDataProviderRegisterFX.java
View File

@ -80,6 +80,7 @@ public class TDDataProviderRegisterFX {
*/
synchronized public TDDataProviderFX findDataProvider(PamDataBlock pamDataBlock) {
for (TDDataProviderFX aProvider:dataProviders) {
// System.out.println("Data providers: " + aProvider.getName() + " | " + aProvider.getDataBlock() + " | " +pamDataBlock);
if (aProvider.getDataBlock()==pamDataBlock) {
return aProvider;
}

2
src/dataPlotsFX/data/generic/GenericDataPlotInfo.java
View File

@ -294,7 +294,7 @@ public class GenericDataPlotInfo extends TDDataInfoFX {
* do some which may be associated with other annotations ?
*/
protected void updateAvailability() {
LocalisationInfo locInfo = getPamDataBlock().getLocalisationContents();
LocalisationInfo locInfo = getDataBlock().getLocalisationContents();
bearingScaleInfo.setAvailable(locInfo.hasLocContent(LocContents.HAS_BEARING));
slantScaleInfo.setAvailable(locInfo.hasLocContent(LocContents.HAS_BEARING));
}

7
src/dataPlotsFX/data/generic/GenericLinePlotInfo.java
View File

@ -23,8 +23,9 @@ import rawDeepLearningClassifier.dataPlotFX.LineInfo;
/**
* A Data Info which plots 1D line data i.e. usually used to plot continuous 1D data e.g. deep learning predictions,
* Ishmael data, click trigger data.
* A Data Info which plots 1D line data i.e. usually used to plot continuous 1D
* data e.g. deep learning predictions, Ishmael data, click trigger data.
*
* @author Jamie Macaulay
*
*/
@ -94,7 +95,7 @@ public abstract class GenericLinePlotInfo extends TDDataInfoFX {
}
//use the center of the window for plotting
double timeMillis = pamDataUnit.getTimeMilliseconds();
double timeMillis = pamDataUnit.getTimeMilliseconds()+pamDataUnit.getDurationInMilliseconds()/2;
double tC=tdProjector.getTimePix(timeMillis-scrollStart);

20
src/dataPlotsFX/scroller/TDAcousticScroller.java
View File

@ -144,10 +144,16 @@ public class TDAcousticScroller extends AcousticScrollerFX implements PamSetting
createTimeSpinner();
this.settingsPane.setParams(scrollerColourParams);
updateDataGramColours(scrollerColourParams);
}
private void setSpinnerValue(long millis) {
spinnerCall = true;
spinner.getValueFactory().setValue(millis);
spinnerCall = false;
}
boolean spinnerCall = false;
@ -177,9 +183,7 @@ public class TDAcousticScroller extends AcousticScrollerFX implements PamSetting
//add a listener so that the spinner changes of the visible amount chnages
super.getScrollBarPane().visibleAmountProperty().addListener((obsVal, oldVal, newVal)->{
spinnerCall = true;
spinner.getValueFactory().setValue(newVal.longValue());
spinnerCall = false;
setSpinnerValue(newVal.longValue());
});
//add a listener so the visible amount changes of the spinner changes value.
@ -213,6 +217,13 @@ public class TDAcousticScroller extends AcousticScrollerFX implements PamSetting
}
});
}
@Override
public void setVisibleMillis(long visibleAmount) {
super.setVisibleMillis(visibleAmount);
setSpinnerValue(visibleAmount);
}
@ -624,6 +635,7 @@ public class TDAcousticScroller extends AcousticScrollerFX implements PamSetting
}
@Override
public String getUnitName() {
// TODO Auto-generated method stub

1
src/pamViewFX/fxNodes/connectionPane/ConnectionLine.java
View File

@ -12,6 +12,7 @@ import javafx.scene.shape.Shape;
/**
* Line which connects a plug or socket to a ConnectionRectangle or other ConnectionLine.
*
* @author Jamie Macaulay
*
*/

2
src/rawDeepLearningClassifier/DLControl.java
View File

@ -275,7 +275,7 @@ public class DLControl extends PamControlledUnit implements PamSettings {
TDDataProviderRegisterFX.getInstance()
.registerDataInfo(new DLDetectionPlotProvider(this, dlClassifyProcess.getDLDetectionDatablock()));
TDDataProviderRegisterFX.getInstance()
.registerDataInfo(new DLPredictionProvider(this, dlClassifyProcess.getDLDetectionDatablock()));
.registerDataInfo(new DLPredictionProvider(this, dlClassifyProcess.getDLPredictionDataBlock()));
// register the DD display
DDPlotRegister.getInstance()

6
src/rawDeepLearningClassifier/dataPlotFX/DLPredictionPlotInfoFX.java
View File

@ -94,7 +94,7 @@ public class DLPredictionPlotInfoFX extends GenericLinePlotInfo {
if (getDlControl().getDLModel()!=null) {
DLClassName[] classNames = getDlControl().getDLModel().getClassNames();
System.out.println("Class names are: !!! " + (classNames == null ? "null" : classNames.length));
// System.out.println("Class names are: !!! " + (classNames == null ? "null" : classNames.length));
if (classNames!=null) {
@ -357,8 +357,8 @@ public class DLPredictionPlotInfoFX extends GenericLinePlotInfo {
updateClassNames();
break;
}
}
}

7
src/rawDeepLearningClassifier/dataPlotFX/DLPredictionProvider.java
View File

@ -5,6 +5,7 @@ import dataPlotsFX.data.TDDataProviderFX;
import dataPlotsFX.layout.TDGraphFX;
import rawDeepLearningClassifier.DLControl;
import rawDeepLearningClassifier.dlClassification.DLDetectionDataBlock;
import rawDeepLearningClassifier.dlClassification.DLModelDataBlock;
/**
* The DL detection plot provider.
@ -22,10 +23,10 @@ public class DLPredictionProvider extends TDDataProviderFX {
/**
* The DL detection plot provider.
* @param dlControl - reference to DL control.
* @param dlDetectionDataBlock - the dl detection data block.
* @param dlModelDataBlock - the dl detection data block.
*/
public DLPredictionProvider(DLControl dlControl, DLDetectionDataBlock dlDetectionDataBlock) {
super(dlDetectionDataBlock);
public DLPredictionProvider(DLControl dlControl, DLModelDataBlock dlModelDataBlock) {
super(dlModelDataBlock);
this.dlControl = dlControl;
}

37
src/rawDeepLearningClassifier/dlClassification/animalSpot/StandardModelPane.java
View File

@ -2,12 +2,16 @@ package rawDeepLearningClassifier.dlClassification.animalSpot;
import java.io.File;
import org.controlsfx.control.CheckComboBox;
import org.controlsfx.control.IndexedCheckModel;
import org.controlsfx.control.PopOver;
import org.controlsfx.control.ToggleSwitch;
import PamController.SettingsPane;
import PamUtils.PamArrayUtils;
import PamView.dialog.warn.WarnOnce;
import javafx.beans.property.SimpleIntegerProperty;
import javafx.beans.value.ObservableValue;
import javafx.collections.ListChangeListener.Change;
import javafx.geometry.Insets;
import javafx.geometry.Pos;
import javafx.scene.Node;
@ -25,6 +29,7 @@ import pamViewFX.fxNodes.PamGridPane;
import pamViewFX.fxNodes.PamHBox;
import pamViewFX.fxNodes.PamSpinner;
import pamViewFX.fxNodes.PamVBox;
import pamViewFX.validator.PamValidator;
import rawDeepLearningClassifier.dlClassification.DLClassiferModel;
/**
@ -101,6 +106,8 @@ public abstract class StandardModelPane extends SettingsPane<StandardModelParams
* Model indicator.
*/
private ProgressIndicator modelLoadIndicator;
PamValidator validator;
@ -191,9 +198,35 @@ public abstract class StandardModelPane extends SettingsPane<StandardModelParams
gridPane.add(new Label(""), 2, 0);
speciesIDBox = new CheckComboBox<String>();
gridPane.add(speciesIDBox, 3, 0);
speciesIDBox.setMaxWidth(100);
speciesIDBox.setPrefWidth(100);
//speciesIDBox.setMaxWidth(100);
// speciesIDBox.setPrefWidth(100);
speciesIDBox.prefHeightProperty().bind(detectionSpinner.heightProperty());
validator = new PamValidator();
final SimpleIntegerProperty checkItemsCount = new SimpleIntegerProperty();
speciesIDBox.getCheckModel().getCheckedItems().addListener((Change<? extends String> c)->{
checkItemsCount.set(speciesIDBox.getCheckModel().getCheckedItems().size());
});
validator.createCheck()
.dependsOn("species_box",checkItemsCount)
.withMethod(c -> {
int nChecked = c.get("species_box");
if (nChecked==speciesIDBox.getItems().size()) {
c.warn("All output class are checked. If one of these classes is noise then PAMGuard will continually detect all sound data...");
}
if (nChecked==0) {
c.warn("No output classes are checked for binary classification. PAMGuard will save all prediction values but no detections will be generated");
}
})
.decorates(speciesIDBox)
.immediate();
;
vBoxHolder = new PamVBox();
vBoxHolder.setSpacing(5);

2
src/rawDeepLearningClassifier/dlClassification/archiveModel/ArchiveModelClassifier.java
View File

@ -167,7 +167,7 @@ public abstract class ArchiveModelClassifier extends StandardClassifierModel {
}
/**
* Create the parameters class for the model. This can be overridden for bespoke parameters.
*Create the parameters class for the model. This can be overridden for bespoke parameters.
*classes.
* @return a new parameters class object.
*/

66
src/rawDeepLearningClassifier/dlClassification/archiveModel/ArchiveModelWorker.java
View File

@ -14,15 +14,18 @@ import org.jamdev.jdl4pam.ArchiveModel;
import org.jamdev.jdl4pam.genericmodel.GenericModelParams;
import org.jamdev.jdl4pam.transforms.DLTransform;
import org.jamdev.jdl4pam.transforms.DLTransformsFactory;
import org.jamdev.jdl4pam.transforms.FreqTransform;
import org.jamdev.jdl4pam.transforms.jsonfile.DLTransformParser2;
import org.jamdev.jdl4pam.transforms.jsonfile.DLTransformsParser;
import org.json.JSONObject;
import PamUtils.PamArrayUtils;
import PamView.dialog.warn.WarnOnce;
import ai.djl.MalformedModelException;
import rawDeepLearningClassifier.DLControl;
import rawDeepLearningClassifier.dlClassification.animalSpot.StandardModelParams;
import rawDeepLearningClassifier.dlClassification.genericModel.DLModelWorker;
import rawDeepLearningClassifier.dlClassification.genericModel.GenericModelWorker;
import rawDeepLearningClassifier.dlClassification.genericModel.GenericPrediction;
/**
@ -33,7 +36,7 @@ import rawDeepLearningClassifier.dlClassification.genericModel.GenericPrediction
* @author Jamie Macaulay
*
*/
public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
public class ArchiveModelWorker extends GenericModelWorker {
/**
@ -57,6 +60,7 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
/**
* Prepare the model
*/
@Override
public void prepModel(StandardModelParams dlParams, DLControl dlControl) {
//ClassLoader origCL = Thread.currentThread().getContextClassLoader();
try {
@ -96,10 +100,11 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
//read the JSON string from the the file.
String jsonString = DLTransformsParser.readJSONString(new File(dlModel.getAudioReprFile()));
//convert the JSON string to a parameters object.
GenericModelParams modelParams = makeModelParams( jsonString);
//important to add this for Ketos models because the JSON string does not necessarily contain and output shape.
//System.out.println("----Default output shape: " + ketosParams.defaultOutputShape + " " + ketosModel.getOutShape());
if (modelParams.defaultOutputShape==null) {
@ -108,6 +113,7 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
//HACK there seems to be some sort of bug in ketos where the params input shape is correct but the model input shape is wrong.
if (dlModel.getInputShape()==null || !dlModel.getInputShape().equals(modelParams.defaultInputShape)) {
System.out.println("Model input shape: " + modelParams.defaultInputShape);
WarnOnce.showWarning("Model shape", "The model shape does not match the model metadata. \n Metadata shape will be used used.", WarnOnce.OK_OPTION);
dlModel.setInputShape(modelParams.defaultInputShape);
}
@ -131,26 +137,35 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
//only load new transforms if defaults are selected
if (getModelTransforms()==null || dlParams.dlTransfroms==null || dlParams.useDefaultTransfroms) {
System.out.println(" " + transforms);
//System.out.println("SET MODEL TRANSFORMS: " + ketosDLParams.dlTransfroms + " " + ketosDLParams.useDefaultTransfroms);
//only set the transforms if they are null - otherwise handled elsewhere.
setModelTransforms(transforms);
dlParams.useDefaultTransfroms = true;
dlParams.dlTransfroms=transforms;
}
else {
//System.out.println("SET CURRENT TRANSFORMS: " + ketosDLParams.dlTransfroms + " " + ketosDLParams.useDefaultTransfroms);
//use the old transforms.
setModelTransforms(dlParams.dlTransfroms);
}
//set whether a wave or spectrogram model
//this is important for setting the input stack into the model.
setWaveFreqModel(dlParams);
// //enable softmax? - TODO
// this.setEnableSoftMax(true);
//ketosDLParams.dlTransfroms = transforms; //this is done after prep model in the settings pane.
dlParams.defaultSegmentLen = modelParams.segLen*1000.; //the segment length in microseconds.
//ketosParams.classNames = new String[] {"Noise", "Right Whale"}; // FIXME;
dlParams.defaultSegmentLen = modelParams.segLen; //the segment length in microseconds.
if (modelParams.classNames!=null) {
dlParams.numClasses = modelParams.classNames.length;
}
else {
dlParams.numClasses = (int) modelParams.defaultOutputShape.get(1);
}
//ok 0 the other values are not user selectable but this is. If we relaod the same model we probably want to keep it....
//So this is a little bt of a hack but will probably be OK in most cases.
//So this is a little bit of a hack but will probably be OK in most cases.
if (dlParams.binaryClassification==null || dlParams.binaryClassification.length!=dlParams.numClasses) {
dlParams.binaryClassification = new boolean[dlParams.numClasses];
for (int i=0; i<dlParams.binaryClassification.length; i++) {
@ -165,19 +180,6 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
//set the number of class names from the default output shape
dlParams.numClasses = (int) modelParams.defaultOutputShape.get(1);
}
// if (dlParams.classNames!=null) {
// for (int i = 0; i<dlParams.classNames.length; i++) {
// System.out.println("Class name " + i + " " + dlParams.classNames[i]);
// }
// }
// ketosDLParams.classNames = dlControl.getClassNameManager().makeClassNames(ketosParams.classNames);
// if (ketosParams.classNames!=null) {
// for (int i = 0; i<ketosDLParams.classNames.length; i++) {
// System.out.println("Class name " + i + " " + ketosDLParams.classNames[i].className + " ID " + ketosDLParams.classNames[i].ID );
// }
// }
}
catch (Exception e) {
dlModel=null;
@ -195,13 +197,6 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
* @throws IOException
*/
public ArchiveModel loadModel(String currentPath2) throws MalformedModelException, IOException {
//note the the model should have been check for compatibility beforehand
File file = new File(currentPath2);
String model = getZipFilePath(file, ".py");
if (model==null) model = getZipFilePath(file, ".pb");
String settings = getZipFilePath(file, ".pdtf");
return new SimpleArchiveModel(new File(currentPath2));
}
@ -225,11 +220,11 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
@Override
public float[] runModel(float[][][] transformedDataStack) {
System.out.println("Model input: " + transformedDataStack.length + " " + transformedDataStack[0].length + " " + transformedDataStack[0][0].length);
return dlModel.runModel(transformedDataStack);
}
// @Override
// public float[] runModel(float[][][] transformedDataStack) {
// System.out.println("Model input: " + transformedDataStack.length + " " + transformedDataStack[0].length + " " + transformedDataStack[0][0].length);
// return dlModel.runModel(transformedDataStack);
// }
@Override
@ -252,6 +247,7 @@ public class ArchiveModelWorker extends DLModelWorker<GenericPrediction> {
* Get the currently loaded mode.
* @return - the currently loaded mode.
*/
@Override
public ArchiveModel getModel() {
return dlModel;
}

2
src/rawDeepLearningClassifier/dlClassification/archiveModel/PamZipModelClassifier.java
View File

@ -65,7 +65,7 @@ public class PamZipModelClassifier extends ArchiveModelClassifier {
}
/**
* Find the first file within a zip folder that matches a pattern.
* Find the first file within a zip folder that matches a pattern. This peaks into the zip file instead of decompressing it.
* @param zipUri - uri to the zip file
* @param filePattern - the file pattern to match - the file must contain this string.
* @return null if no file found and the file pqth if the file is founf

10
src/rawDeepLearningClassifier/dlClassification/archiveModel/SimpleArchiveModel.java
View File

@ -28,7 +28,7 @@ public class SimpleArchiveModel extends ArchiveModel {
@Override
public String getAudioReprRelPath(String zipFolder) {
try {
System.out.println("SETTINGS PATH: " + getRelFilePath(zipFolder, ".pdtf"));
//System.out.println("SETTINGS PATH: " + getRelFilePath(zipFolder, ".pdtf"));
return getRelFilePath(zipFolder, ".pdtf");
} catch (IOException e) {
// TODO Auto-generated catch block
@ -43,7 +43,7 @@ public class SimpleArchiveModel extends ArchiveModel {
String model = null;
model = getRelFilePath(zipFolder, ".pb");
if (model==null) model = getRelFilePath(zipFolder, ".py");
System.out.println("MODEL PATH: " +model);
//System.out.println("MODEL PATH: " +model);
return model;
} catch (IOException e) {
@ -64,14 +64,16 @@ public class SimpleArchiveModel extends ArchiveModel {
try (Stream<Path> walk = Files.walk(Paths.get(zipFolder))) {
List<String> result = walk
.filter(p -> !Files.isDirectory(p)) // not a directory
.map(p -> p.toString().toLowerCase()) // convert path to string
.map(p -> p.toString()) // convert path to string
.filter(f -> f.endsWith(fileEnd)) // check end with
.collect(Collectors.toList()); // collect all matched to a List
if (result.size()>0) {
String firstFile = result.get(0);
String relative = new File(zipFolder).toURI().relativize(new File(firstFile).toURI()).getPath();
//System.out.println("First file: " +firstFile);
String relative = firstFile.replace(zipFolder, "");
return relative;
}

1
src/rawDeepLearningClassifier/dlClassification/genericModel/GenericAdvPane.java
View File

@ -184,7 +184,6 @@ public class GenericAdvPane extends SettingsPane<GenericModelParams> {
defualtShapeSwitch = new PamToggleSwitch("Use model default shape");
defualtShapeSwitch.selectedProperty().addListener((obsval, oldval, newval)->{
System.out.println("Hello: deafult shape: " + newval);
PamArrayUtils.printArray( currentInput.defaultShape);
if (newval) {

15
src/rawDeepLearningClassifier/dlClassification/genericModel/GenericDLClassifier.java
View File

@ -121,13 +121,12 @@ public class GenericDLClassifier extends StandardClassifierModel {
genericModelParams.dlTransfromParams=dlTransformParams;
if (genericModelParams.dlTransfromParams!=null) {
System.out.println("Generic settings have been saved. : " + genericModelParams.dlTransfromParams.size());
}
else {
System.out.println("Generic settings have been saved. : " + null);
}
// if (genericModelParams.dlTransfromParams!=null) {
// System.out.println("Generic settings have been saved. : " + genericModelParams.dlTransfromParams.size());
// }
// else {
// System.out.println("Generic settings have been saved. : " + null);
// }
return genericModelParams;
}
@ -211,6 +210,8 @@ public class GenericDLClassifier extends StandardClassifierModel {
*/
public static boolean isBinaryResult(GenericPrediction 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;

30
src/rawDeepLearningClassifier/dlClassification/genericModel/GenericModelWorker.java
View File

@ -33,19 +33,21 @@ public class GenericModelWorker extends DLModelWorker<GenericPrediction> {
@Override
public float[] runModel(float[][][] transformedDataStack) {
//System.out.println("RUN GENERIC MODEL: " + transformedDataStack.length + " " + transformedDataStack[0].length + " " + transformedDataStack[0][0].length);
//System.out.println("RUN GENERIC MODEL: " + transformedDataStack[0][0][0]);
// System.out.println("RUN GENERIC MODEL: " + transformedDataStack[0][0][0]);
float[] results;
if (freqTransform)
results = genericModel.runModel(transformedDataStack);
results = getModel().runModel(transformedDataStack);
else {
//run a model if it is waveform info.
float[][] waveStack = new float[transformedDataStack.length][];
for (int i=0; i<waveStack.length; i++) {
waveStack[i] = transformedDataStack[i][0];
}
results = genericModel.runModel(waveStack);
//System.out.println("RUN GENERIC MODEL WAVE: " + transformedDataStack.length + " " + transformedDataStack[0].length);
results = getModel().runModel(waveStack);
}
System.out.println("GENERIC MODEL RESULTS: " + results== null ? null : results.length);
//System.out.println("GENERIC MODEL RESULTS: " + results== null ? null : results.length);
return results;
}
@ -97,13 +99,7 @@ public class GenericModelWorker extends DLModelWorker<GenericPrediction> {
setModelTransforms(genericParams.dlTransfroms);
//is this a waveform or a spectrogram model?
DLTransform transform = genericParams.dlTransfroms.get(genericParams.dlTransfroms.size()-1);
if (transform instanceof FreqTransform) {
freqTransform = true;
}
else {
freqTransform = false;
}
setWaveFreqModel(genericParams);
//use softmax or not?
String extension = FilenameUtils.getExtension(genericParams.modelPath);
@ -131,6 +127,18 @@ public class GenericModelWorker extends DLModelWorker<GenericPrediction> {
//Thread.currentThread().setContextClassLoader(origCL);
}
protected void setWaveFreqModel(StandardModelParams genericParams) {
//is this a waveform or a spectrogram model?
DLTransform transform = genericParams.dlTransfroms.get(genericParams.dlTransfroms.size()-1);
if (transform instanceof FreqTransform) {
freqTransform = true;
}
else {
freqTransform = false;
}
}
/**
* Convert a long[] array to a Long[] array.
* @param inArr - the input long[] array.

4
src/rawDeepLearningClassifier/dlClassification/ketos/KetosWorker2.java
View File

@ -32,6 +32,8 @@ public class KetosWorker2 extends ArchiveModelWorker {
public GenericModelParams makeModelParams(String jsonString) {
//ketos parameters are non standard and need a bit of extra work to get right.
//This also deal with legacy paramters.
return new KetosParams(jsonString);
KetosParams params = new KetosParams(jsonString);
// System.out.println(params);
return params;
}
}

3
src/rawDeepLearningClassifier/layoutFX/DLModelSelectPane.java
View File

@ -377,6 +377,7 @@ public class DLModelSelectPane extends PamBorderPane {
//we are loading model from a file - anything can happen so put in a try catch.
DLStatus status = currentClassifierModel.setModel(file);
if (status.isError()) {
System.err.println("Model load failed: " + currentClassifierModel.getModelStatus());
currentClassifierModel=null;
@ -437,7 +438,7 @@ public class DLModelSelectPane extends PamBorderPane {
pathLabel.setText("No classifier model loaded: Select model");
pathLabel.setTooltip(new Tooltip("Use the browse button/ URI botton to select a model or select a default model"));
}
if (currentClassifierModel.getModelStatus().isError()) {
else if (currentClassifierModel.getModelStatus().isError()) {
pathLabel.setGraphic(createErrorIcon(currentClassifierModel.getModelStatus()));
pathLabel.setText(currentClassifierModel.getModelStatus().getName());
pathLabel.setTooltip(new Tooltip(currentClassifierModel.getModelStatus().getDescription()));

204
src/rawDeepLearningClassifier/layoutFX/DLSettingsPane.java
View File

@ -20,8 +20,10 @@ import javafx.scene.control.Label;
import javafx.scene.control.Spinner;
import javafx.scene.control.Tooltip;
import javafx.scene.control.Alert.AlertType;
import javafx.scene.layout.ColumnConstraints;
import javafx.scene.layout.HBox;
import javafx.scene.layout.Pane;
import javafx.scene.layout.Priority;
import pamViewFX.PamGuiManagerFX;
import pamViewFX.fxGlyphs.PamGlyphDude;
import pamViewFX.fxNodes.PamBorderPane;
@ -48,7 +50,7 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
public static double MAX_WIDTH = 270;
/**
* The source for the FFT data source.
*/
@ -64,10 +66,10 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
*/
private PamBorderPane mainPane;
// /**
// * Combo box which allows users to select model.
// */
// private ComboBox<String> dlModelBox;
// /**
// * Combo box which allows users to select model.
// */
// private ComboBox<String> dlModelBox;
/**
* The window length spinner for the segmenter process
@ -112,9 +114,9 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
private Label infoLabel;
private DLModelSelectPane modelSelectPane;
public DLSettingsPane(DLControl dlControl){
super(null);
this.dlControl=dlControl;
@ -126,19 +128,19 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
// stage.sizeToScene();
// });
// this.setTop(newButton);
mainPane=new PamBorderPane();
mainPane.setCenter(createDLPane());
mainPane.setPadding(new Insets(5,5,5,5));
if (!PamGUIManager.isFX()){
mainPane.setMinHeight(430);
mainPane.setMaxWidth(MAX_WIDTH);
mainPane.setPrefWidth(MAX_WIDTH);
}
//this.getAdvPane().setMaxWidth(MAX_WIDTH);
//mainPane.getStylesheets().add(PamStylesManagerFX.getPamStylesManagerFX().getCurStyle().getDialogCSS());
@ -162,14 +164,14 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
vBox.getChildren().add(sourcePane);
sourcePane.prefWidthProperty().bind(vBox.widthProperty());
sourcePane.setMaxWidth(Double.MAX_VALUE);
sourcePane.getDataBlockBox().setOnAction((action)->{
//need to create a data selector if one exists.
this.dlControl.createDataSelector(getSelectedParentDataBlock());
//enable the controls to show a data selector or not.
enableControls();
});
//create the detection
vBox.getChildren().add(createDataSelectorPane());
@ -193,14 +195,14 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
hopLength.valueProperty().addListener((obsVal, oldVal, newVal)->{
setSegInfoLabel();
});
reMergeSeg = new PamSpinner<Integer>(0, Integer.MAX_VALUE, 1, 1);
reMergeSeg.getStyleClass().add(Spinner.STYLE_CLASS_SPLIT_ARROWS_HORIZONTAL);
reMergeSeg.setEditable(true);
//button to set default hop size
Button defaultButton = new Button();
// defaultButton.setGraphic(PamGlyphDude.createPamGlyph(MaterialDesignIcon.REFRESH, PamGuiManagerFX.iconSize-3));
// defaultButton.setGraphic(PamGlyphDude.createPamGlyph(MaterialDesignIcon.REFRESH, PamGuiManagerFX.iconSize-3));
defaultButton.setGraphic(PamGlyphDude.createPamIcon("mdi2r-refresh", PamGuiManagerFX.iconSize-3));
defaultButton.setTooltip(new Tooltip("Set default hop size"));
defaultButton.setOnAction((action)->{
@ -208,6 +210,14 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
});
PamGridPane segmenterGridPane = new PamGridPane();
segmenterGridPane.setHgap(5);
ColumnConstraints col1 = new ColumnConstraints();
col1.setHgrow( Priority.SOMETIMES );
ColumnConstraints col2 = new ColumnConstraints();
col2.setHgrow( Priority.ALWAYS );
segmenterGridPane.getColumnConstraints().addAll(col1, col2 );
segmenterGridPane.add(new Label("Window length"), 0, 0);
segmenterGridPane.add(windowLength, 1, 0);
segmenterGridPane.add(new Label("samples"), 2, 0);
@ -222,40 +232,40 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
segmenterGridPane.add(new Label("segments"), 2, 2);
vBox.getChildren().add(segmenterGridPane);
vBox.getChildren().add(infoLabel = new Label());
Label label2 = new Label("Deep Learning Model");
label2.setPadding(new Insets(5,0,0,0));
PamGuiManagerFX.titleFont2style(label2);
vBox.getChildren().add(label2);
/**
* Pane which allows users to select a model type.
*/
modelSelectPane = new DLModelSelectPane(this);
modelSelectPane = new DLModelSelectPane(this);
// //add the possible deep learning models.
// dlModelBox= new ComboBox<String>();
// for (int i=0; i<dlControl.getDLModels().size(); i++) {
// dlModelBox.getItems().add(dlControl.getDLModels().get(i).getName());
// }
// dlModelBox.prefWidthProperty().bind(vBox.widthProperty());
//
// dlModelBox.setOnAction((action)->{
// setClassifierPane();
// if (mainPane!=null) {
// if (mainPane.getScene().getWindow() instanceof Stage) {
// Stage stage = (Stage) mainPane.getScene().getWindow();
// stage.sizeToScene();
// }
// }
// //this.dlControl.getAnnotationType().getSymbolModifier(symbolChooser).
// });
//
// vBox.getChildren().add(dlModelBox);
// //add the possible deep learning models.
// dlModelBox= new ComboBox<String>();
// for (int i=0; i<dlControl.getDLModels().size(); i++) {
// dlModelBox.getItems().add(dlControl.getDLModels().get(i).getName());
// }
// dlModelBox.prefWidthProperty().bind(vBox.widthProperty());
//
// dlModelBox.setOnAction((action)->{
// setClassifierPane();
// if (mainPane!=null) {
// if (mainPane.getScene().getWindow() instanceof Stage) {
// Stage stage = (Stage) mainPane.getScene().getWindow();
// stage.sizeToScene();
// }
// }
// //this.dlControl.getAnnotationType().getSymbolModifier(symbolChooser).
// });
//
// vBox.getChildren().add(dlModelBox);
classifierPane = new PamBorderPane();
@ -265,7 +275,7 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
return vBox;
}
/**
* Create the data selector.
* @return the data selector.
@ -274,13 +284,13 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
dataSelectorPane = new PamHBox();
dataSelectorPane.setSpacing(5);
dataSelectorPane.setAlignment(Pos.CENTER_LEFT);
dataSelectorCheckBox = new PamToggleSwitch("Detection Selector");
dataSelectorCheckBox.selectedProperty().addListener((obsval, oldval, newval)->{
enableControls();
});
dataSelectorButton = new PamButton();
// dataSelectorButton.setGraphic(PamGlyphDude.createPamGlyph(MaterialIcon.SETTINGS, PamGuiManagerFX.iconSize));
// dataSelectorButton.setGraphic(PamGlyphDude.createPamGlyph(MaterialIcon.SETTINGS, PamGuiManagerFX.iconSize));
dataSelectorButton.setGraphic(PamGlyphDude.createPamIcon("mdi2c-cog", PamGuiManagerFX.iconSize));
dataSelectorButton.setOnAction((action)->{
showAdvPane();
@ -288,7 +298,7 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
dataSelectorPane.getChildren().addAll(dataSelectorCheckBox, dataSelectorButton);
return dataSelectorPane;
}
/**
* Set extra information in the info label.
*/
@ -306,7 +316,7 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
}
infoLabel.setText(text);
}
/**
* Creates pane allowing the user to change fine scale things such as error limits.
* @return the pop over pane.
@ -327,7 +337,7 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
popOver.show(dataSelectorButton);
}
/**
* Enable the controls.
*/
@ -338,10 +348,10 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
//^bit messy but cannot think of a better way to do it.
this.dataSelectorPane.setVisible(false);
}
dataSelectorButton.setDisable(!dataSelectorCheckBox.isSelected());
}
/**
@ -368,9 +378,9 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
protected void setClassifierPane() {
//set the classifier Pane.class
if (modelSelectPane.currentClassifierModel!=null && modelSelectPane.currentClassifierModel.getModelUI()!=null) {
classifierPane.setCenter(modelSelectPane.currentClassifierModel.getModelUI().getSettingsPane().getContentNode());
if (modelSelectPane.currentClassifierModel!=null) {
modelSelectPane.currentClassifierModel.getModelUI().setParams();
}
@ -400,8 +410,8 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
}
sourcePane.getParams(currParams.groupedSourceParams);
// currParams.modelSelection = dlModelBox.getSelectionModel().getSelectedIndex();
// currParams.modelSelection = dlModelBox.getSelectionModel().getSelectedIndex();
if (windowLength.getValue() == 0 || hopLength.getValue()==0){
Platform.runLater(()->{
@ -413,8 +423,8 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
currParams.rawSampleSize = windowLength.getValue();
currParams.sampleHop = hopLength.getValue();
currParams.maxMergeHops = reMergeSeg.getValue();
if (modelSelectPane.currentClassifierModel == null) {
currParams.modelSelection = -1;
}
@ -423,28 +433,28 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
}
// //update any changes
// if (this.dlControl.getDLModels().get(dlModelBox.getSelectionModel().getSelectedIndex()).getModelUI()!=null){
// this.dlControl.getDLModels().get(dlModelBox.getSelectionModel().getSelectedIndex()).getModelUI().getParams();
//
// //display any warnings from the settings.
// ArrayList<PamWarning> warnings = this.dlControl.getDLModels().get(dlModelBox.getSelectionModel().getSelectedIndex()).checkSettingsOK();
// showWarnings(warnings);
//
// for (int i=0; i<warnings.size(); i++) {
// if (warnings.get(i).getWarnignLevel()>1) {
// //Serious error. Do not close dialog.
// return null;
// }
// }
// }
// //update any changes
// if (this.dlControl.getDLModels().get(dlModelBox.getSelectionModel().getSelectedIndex()).getModelUI()!=null){
// this.dlControl.getDLModels().get(dlModelBox.getSelectionModel().getSelectedIndex()).getModelUI().getParams();
//
// //display any warnings from the settings.
// ArrayList<PamWarning> warnings = this.dlControl.getDLModels().get(dlModelBox.getSelectionModel().getSelectedIndex()).checkSettingsOK();
// showWarnings(warnings);
//
// for (int i=0; i<warnings.size(); i++) {
// if (warnings.get(i).getWarnignLevel()>1) {
// //Serious error. Do not close dialog.
// return null;
// }
// }
// }
currParams.useDataSelector = dataSelectorCheckBox.isSelected();
if (dlControl.getDataSelector()!=null) {
dlControl.getDataSelector().getDialogPaneFX().getParams(true);
}
//need to make sure we call get params for the current model when the oK button is pressed.
if (this.modelSelectPane.currentClassifierModel!=null) {
if (this.modelSelectPane.currentClassifierModel.getModelUI()!=null) {
@ -453,16 +463,16 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
}
currParams.modelURI = this.modelSelectPane.currentSelectedFile;
return currParams;
}
public static PamWarning statusToWarnings(DLStatus dlStatus) {
PamWarning pamWarning = new PamWarning(dlStatus.getName(), dlStatus.getDescription(), dlStatus.isError() ? 2 : 1);
return pamWarning;
}
/**
* Show a warning dialog for the status
* @param the status to show
@ -472,7 +482,7 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
dlWarnings.add(statusToWarnings(dlWarning));
showWarning(dlWarnings);
}
/**
* Show a warning dialog.
* @param the warning to show.
@ -489,12 +499,12 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
* @param dlWarnings - list of warnings - the most important will be shown.
*/
public void showWarning(ArrayList<PamWarning> dlWarnings) {
if (dlWarnings==null || dlWarnings.size()<1) return;
String warnings ="";
boolean error = false;
for (int i=0; i<dlWarnings.size(); i++) {
warnings += dlWarnings.get(i).getWarningMessage() + "\n\n";
@ -502,14 +512,14 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
error=true;
}
}
final String warningsF = warnings;
final boolean errorF = error;
Platform.runLater(()->{
WarnOnce.showWarningFX(null, "Deep Learning Settings Warning", warningsF , errorF ? AlertType.ERROR : AlertType.WARNING);
// WarnOnce.showWarning( "Deep Learning Settings Warning", warningsF , WarnOnce.WARNING_MESSAGE);
// WarnOnce.showWarning( "Deep Learning Settings Warning", warningsF , WarnOnce.WARNING_MESSAGE);
});
//user presses OK - these warnings are just a message - they do not prevent running the module.
}
@ -518,35 +528,35 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
sourcePane.setParams(currParams.groupedSourceParams);
sourcePane.sourceChanged();
dlControl.createDataSelector(sourcePane.getSource());
//set the classifier model.
if (currParams.modelURI !=null) {
modelSelectPane.currentClassifierModel = dlControl.getDLModel();
}
// dlModelBox.getSelectionModel().select(currParams.modelSelection);
// dlModelBox.getSelectionModel().select(currParams.modelSelection);
windowLength.getValueFactory().setValue(currParams.rawSampleSize);
hopLength.getValueFactory().setValue(currParams.sampleHop);
reMergeSeg.getValueFactory().setValue(currParams.maxMergeHops);
dataSelectorCheckBox.setSelected(currParams.useDataSelector);
setClassifierPane();
enableControls();
setSegInfoLabel();
// //set up the model and the custom pane if necessary.
// //set up the model and the custom pane if necessary.
this.modelSelectPane.loadNewModel(currParams.modelURI);
//this.modelSelectPane.updatePathLabel();
this.setClassifierPane();
//For some reason, in the FX GUI, this causes a root used in multiple scenes exceptions...not sure why.
Platform.runLater(()->{
sourcePane.getChannelValidator().validate();
@ -598,23 +608,23 @@ public class DLSettingsPane extends SettingsPane<RawDLParams>{
*/
public void setSegmentLength(Double defaultSegmentLen) {
if (defaultSegmentLen==null) return;
double sR = getDLControl().getSettingsPane().getSelectedParentDataBlock().getSampleRate();
System.out.println("Set the segment length: " + defaultSegmentLen + " sR " + sR);
//automatically set the default segment length.
getDLControl().getSettingsPane().getSegmentLenSpinner().getValueFactory().setValue((int) (sR*defaultSegmentLen/1000.));
}
/**
* Convenience class to set the hop length in samples from milliseconds
* @param defaultSegmentLen - the segment length in milliseconds.
*/
public void setHopLength(Double hopLength) {
if (hopLength==null) return;
double sR = getDLControl().getSettingsPane().getSelectedParentDataBlock().getSampleRate();
//automatically set the default segment length.
getDLControl().getSettingsPane().getHopLenSpinner().getValueFactory().setValue((int) (sR*hopLength/1000.));

2
src/rawDeepLearningClassifier/logging/DLDataUnitDatagram.java
View File

@ -31,6 +31,8 @@ public class DLDataUnitDatagram implements DatagramProvider {
public int addDatagramData(PamDataUnit dataUnit, float[] dataGramLine) {
DLDataUnit dlDataUnit = (DLDataUnit) dataUnit;
if (dataGramLine==null || dataGramLine.length==0) return 0;
if (dlDataUnit.getPredicitionResult().getPrediction()!=null) {
for (int i=0; i<dlDataUnit.getPredicitionResult().getPrediction().length; i++) {
dataGramLine[i] += (float) dlDataUnit.getPredicitionResult().getPrediction()[i];