opennlp.tools.ml.model.AbstractDataIndexer Maven / Gradle / Ivy
/*
* 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 opennlp.tools.ml.model;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
import java.util.Comparator;
import java.util.HashMap;
import java.util.List;
import java.util.Map;
import java.util.Objects;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import opennlp.tools.util.InsufficientTrainingDataException;
import opennlp.tools.util.ObjectStream;
import opennlp.tools.util.TrainingParameters;
/**
* Abstract {@link DataIndexer} implementation for collecting
* event and context counts used in training.
*
* @see DataIndexer
*/
public abstract class AbstractDataIndexer implements DataIndexer {
private static final Logger logger = LoggerFactory.getLogger(AbstractDataIndexer.class);
public static final String SORT_PARAM = "sort";
public static final boolean SORT_DEFAULT = true;
protected TrainingParameters trainingParameters;
protected Map reportMap;
/**
* {@inheritDoc}
*/
@Override
public void init(TrainingParameters indexingParameters, Map reportMap) {
this.reportMap = reportMap;
if (this.reportMap == null) reportMap = new HashMap<>();
trainingParameters = indexingParameters;
}
private int numEvents;
/** The integer contexts associated with each unique event. */
protected int[][] contexts;
/** The integer outcome associated with each unique event. */
protected int[] outcomeList;
/** The number of times an event occurred in the training data. */
protected int[] numTimesEventsSeen;
/** The predicate/context names. */
protected String[] predLabels;
/** The names of the outcomes. */
protected String[] outcomeLabels;
/** The number of times each predicate occurred. */
protected int[] predCounts;
/**
* {@inheritDoc}
*/
@Override
public int[][] getContexts() {
return contexts;
}
/**
* {@inheritDoc}
*/
@Override
public int[] getNumTimesEventsSeen() {
return numTimesEventsSeen;
}
/**
* {@inheritDoc}
*/
@Override
public int[] getOutcomeList() {
return outcomeList;
}
/**
* {@inheritDoc}
*/
@Override
public String[] getPredLabels() {
return predLabels;
}
/**
* {@inheritDoc}
*/
@Override
public String[] getOutcomeLabels() {
return outcomeLabels;
}
/**
* {@inheritDoc}
*/
@Override
public int[] getPredCounts() {
return predCounts;
}
/**
* {@inheritDoc}
*/
@Override
public int getNumEvents() {
return numEvents;
}
/**
* Sorts and uniques the array of comparable events and return the number of unique events.
* This method will alter the {@code eventsToCompare} list.
*
* It does an in place sort, followed by an in place edit to remove duplicates.
*
* @param eventsToCompare The {@link List} events used as input.
* @param sort Whether to use sorting, or not.
*
* @return The number of unique events in the specified list.
* @throws InsufficientTrainingDataException Thrown if not enough events are provided
* @since maxent 1.2.6
*/
protected int sortAndMerge(List eventsToCompare, boolean sort)
throws InsufficientTrainingDataException {
int numUniqueEvents = 1;
numEvents = eventsToCompare.size();
if (sort && eventsToCompare.size() > 0) {
Collections.sort(eventsToCompare);
ComparableEvent ce = eventsToCompare.get(0);
for (int i = 1; i < numEvents; i++) {
ComparableEvent ce2 = eventsToCompare.get(i);
if (ce.compareTo(ce2) == 0) {
ce.seen++; // increment the seen count
eventsToCompare.set(i, null); // kill the duplicate
}
else {
ce = ce2; // a new champion emerges...
numUniqueEvents++; // increment the # of unique events
}
}
}
else {
numUniqueEvents = eventsToCompare.size();
}
if (numUniqueEvents == 0) {
throw new InsufficientTrainingDataException("Insufficient training data to create model.");
}
if (sort) logger.info("done. Reduced {} events to {}.", numEvents, numUniqueEvents);
contexts = new int[numUniqueEvents][];
outcomeList = new int[numUniqueEvents];
numTimesEventsSeen = new int[numUniqueEvents];
for (int i = 0, j = 0; i < numEvents; i++) {
ComparableEvent evt = eventsToCompare.get(i);
if (null == evt) {
continue; // this was a dupe, skip over it.
}
numTimesEventsSeen[j] = evt.seen;
outcomeList[j] = evt.outcome;
contexts[j] = evt.predIndexes;
++j;
}
return numUniqueEvents;
}
/**
* Performs the data indexing.
*
* Note:
* Make sure the {@link #init(TrainingParameters, Map)} method is called first.
*
* @param events A {@link ObjectStream} of events used as input.
* @param predicateIndex A {@link Map} providing the data of a predicate index.
*
* @throws IOException Thrown if IO errors occurred during indexing.
*/
protected List index(ObjectStream events,
Map predicateIndex) throws IOException {
Map omap = new HashMap<>();
List eventsToCompare = new ArrayList<>();
Event ev;
while ((ev = events.read()) != null) {
omap.putIfAbsent(ev.getOutcome(), omap.size());
int[] cons = Arrays.stream(ev.getContext())
.map(predicateIndex::get)
.filter(Objects::nonNull)
.mapToInt(i -> i).toArray();
// drop events with no active features
if (cons.length > 0) {
int ocID = omap.get(ev.getOutcome());
eventsToCompare.add(new ComparableEvent(ocID, cons, ev.getValues()));
} else {
logger.info("Dropped event {}:{}", ev.getOutcome(), Arrays.asList(ev.getContext()));
}
}
outcomeLabels = toIndexedStringArray(omap);
predLabels = toIndexedStringArray(predicateIndex);
return eventsToCompare;
}
/**
* Updates the {@link Map} of predicates and counter with the specified event contexts.
*
* @param ec The contexts/features which occur in an event.
* @param counter The predicate counters in form of a {@link Map}.
*/
protected static void update(String[] ec, Map counter) {
for (String s : ec) {
counter.merge(s, 1, (value, one) -> value + one);
}
}
/**
* Utility method for creating a {@code String[]} from a map whose
* keys are labels (Strings) to be stored in the array and whose
* values are the indices (Integers) at which the corresponding
* labels should be inserted.
*
* @param labelToIndexMap A {@link Map} that holds labels to index positions.
* @return The resulting {@code String[]}.
*/
protected static String[] toIndexedStringArray(Map labelToIndexMap) {
return labelToIndexMap.entrySet().stream()
.sorted(Comparator.comparingInt(Map.Entry::getValue))
.map(Map.Entry::getKey).toArray(String[]::new);
}
@Override
public float[][] getValues() {
return null;
}
}