/*
* Encog(tm) Core v2.5 - Java Version
* http://www.heatonresearch.com/encog/
* http://code.google.com/p/encog-java/
* Copyright 2008-2010 Heaton Research, Inc.
*
* 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.
*
* For more information on Heaton Research copyrights, licenses
* and trademarks visit:
* http://www.heatonresearch.com/copyright
*/
package org.encog.persist.persistors;
import org.encog.EncogError;
import org.encog.engine.network.activation.ActivationFunction;
import org.encog.neural.NeuralNetworkError;
import org.encog.neural.networks.layers.BasicLayer;
import org.encog.parse.tags.read.ReadXML;
import org.encog.parse.tags.write.WriteXML;
import org.encog.persist.EncogPersistedCollection;
import org.encog.persist.EncogPersistedObject;
import org.encog.persist.Persistor;
import org.encog.util.csv.CSVFormat;
import org.encog.util.csv.NumberList;
import org.encog.util.obj.ReflectionUtil;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
/**
* Provides basic functions that many of the persistors will need.
*
*
* @author jheaton
*/
public class BasicLayerPersistor implements Persistor {
/**
* The activation function tag.
*/
public static final String TAG_ACTIVATION = "activation";
/**
* The neurons property.
*/
public static final String PROPERTY_NEURONS = "neurons";
/**
* The bias property, stores the bias weights.
*/
public static final String PROPERTY_THRESHOLD = "threshold";
/**
* The bias activation.
*/
public static final String PROPERTY_BIAS_ACTIVATION = "biasActivation";
/**
* The x-coordinate to place this object at.
*/
public static final String PROPERTY_X = "x";
/**
* The y-coordinate to place this object at.
*/
public static final String PROPERTY_Y = "y";
/**
* The logging object.
*/
@SuppressWarnings("unused")
private final Logger logger = LoggerFactory.getLogger(this.getClass());
/**
* Load the specified Encog object from an XML reader.
*
* @param in
* The XML reader to use.
* @return The loaded object.
*/
public EncogPersistedObject load(final ReadXML in) {
double biasActivation = 1;
int neuronCount = 0;
int x = 0;
int y = 0;
String threshold = null;
ActivationFunction activation = null;
final String end = in.getTag().getName();
while (in.readToTag()) {
if (in.is(BasicLayerPersistor.TAG_ACTIVATION, true)) {
in.readToTag();
final String type = in.getTag().getName();
activation = loadActivation(type, in);
} else if (in.is(BasicLayerPersistor.PROPERTY_NEURONS, true)) {
neuronCount = in.readIntToTag();
} else if (in.is(BasicLayerPersistor.PROPERTY_THRESHOLD, true)) {
threshold = in.readTextToTag();
} else if (in.is(BasicLayerPersistor.PROPERTY_X, true)) {
x = in.readIntToTag();
} else if (in.is(BasicLayerPersistor.PROPERTY_Y, true)) {
y = in.readIntToTag();
} else if (in
.is(BasicLayerPersistor.PROPERTY_BIAS_ACTIVATION, true)) {
biasActivation = Double.parseDouble(in.readTextToTag());
} else if (in.is(end, false)) {
break;
}
}
if (neuronCount > 0) {
BasicLayer layer;
if (threshold == null) {
layer = new BasicLayer(activation, false, neuronCount);
} else {
final double[] t = NumberList.fromList(CSVFormat.EG_FORMAT,
threshold);
layer = new BasicLayer(activation, true, neuronCount);
for (int i = 0; i < t.length; i++) {
layer.setBiasWeight(i, t[i]);
}
layer.setBiasActivation(biasActivation);
}
layer.setX(x);
layer.setY(y);
return layer;
}
return null;
}
/**
* Save the specified Encog object to an XML writer.
*
* @param obj
* The object to save.
* @param out
* The XML writer to save to.
*/
public void save(final EncogPersistedObject obj, final WriteXML out) {
PersistorUtil.beginEncogObject(
EncogPersistedCollection.TYPE_BASIC_LAYER, out, obj, false);
final BasicLayer layer = (BasicLayer) obj;
out.addProperty(BasicLayerPersistor.PROPERTY_NEURONS, layer
.getNeuronCount());
out.addProperty(BasicLayerPersistor.PROPERTY_X, layer.getX());
out.addProperty(BasicLayerPersistor.PROPERTY_Y, layer.getY());
if (layer.hasBias()) {
final StringBuilder result = new StringBuilder();
NumberList.toList(CSVFormat.EG_FORMAT, result, layer
.getBiasWeights());
out.addProperty(BasicLayerPersistor.PROPERTY_THRESHOLD, result
.toString());
}
out.addProperty(BasicLayerPersistor.PROPERTY_BIAS_ACTIVATION, layer
.getBiasActivation());
saveActivationFunction(layer.getActivationFunction(), out);
out.endTag();
}
public static void saveActivationFunction(
ActivationFunction activationFunction, WriteXML out) {
if (activationFunction != null) {
out.beginTag(BasicLayerPersistor.TAG_ACTIVATION);
String[] names = activationFunction.getParamNames();
for (int i = 0; i < names.length; i++) {
String str = names[i];
double d = activationFunction.getParams()[i];
out.addAttribute(str, "" + CSVFormat.EG_FORMAT.format(d, 10));
}
out.beginTag(activationFunction.getClass().getSimpleName());
out.endTag();
out.endTag();
}
}
public static ActivationFunction loadActivation(String type, ReadXML in) {
try {
Class<?> clazz = ReflectionUtil.resolveEncogClass(type);
if( clazz==null ) {
throw new NeuralNetworkError("Unknown activation function type: " + type);
}
ActivationFunction result = (ActivationFunction) clazz
.newInstance();
for (String key : in.getTag().getAttributes().keySet()) {
int index = -1;
for (int i = 0; i < result.getParamNames().length; i++) {
if (key.equalsIgnoreCase(result.getParamNames()[i])) {
index = i;
break;
}
}
if (index != -1) {
String str = in.getTag().getAttributeValue(key);
double d = CSVFormat.EG_FORMAT.parse(str);
result.setParam(index, d);
}
}
return result;
} catch (InstantiationException e) {
throw new EncogError(e);
} catch (IllegalAccessException e) {
throw new EncogError(e);
}
}
}