/* Copyright (C) 2002 Univ. of Massachusetts Amherst, Computer Science Dept. This file is part of "MALLET" (MAchine Learning for LanguagE Toolkit). http://www.cs.umass.edu/~mccallum/mallet This software is provided under the terms of the Common Public License, version 1.0, as published by http://www.opensource.org. For further information, see the file `LICENSE' included with this distribution. */ package cc.mallet.types; import java.util.Arrays; import java.io.*; /** * An implementation of {@link Sequence} that ensures that every * Object in the sequence has the same class. Feature sequences are * mutable, and will expand as new objects are added. * * @author Andrew McCallum <a href="mailto:mccallum@cs.umass.edu">mccallum@cs.umass.edu</a> */ public class FeatureSequence implements Sequence, Serializable, AlphabetCarrying { Alphabet dictionary; int[] features; int length; /** * Creates a FeatureSequence given all of the objects in the * sequence. * * @param dict A dictionary that maps objects in the sequence * to numeric indices. * @param features An array where features[i] gives the index * in dict of the ith element of the sequence. */ public FeatureSequence (Alphabet dict, int[] features) { this(dict, features.length); for (int i = 0; i < features.length; i++) add(features[i]); } public FeatureSequence (Alphabet dict, int[] features, int len) { this(dict, len); for (int i = 0; i < len; i++) add(features[i]); } public FeatureSequence (Alphabet dict, int capacity) { dictionary = dict; features = new int[capacity > 2 ? capacity : 2]; length = 0; } public FeatureSequence (Alphabet dict) { this (dict, 2); } public int[] getFeatures() { return features ;} public Alphabet getAlphabet () { return dictionary; } public Alphabet[] getAlphabets() { return new Alphabet[] {getAlphabet()}; } public boolean alphabetsMatch (AlphabetCarrying object) { return getAlphabet().equals (object.getAlphabet()); } public final int getLength () { return length; } public final int size () { return length; } public final int getIndexAtPosition (int pos) { return features[pos]; } public Object getObjectAtPosition (int pos) { return dictionary.lookupObject (features[pos]); } // xxx This method name seems a bit ambiguous? public Object get (int pos) { return dictionary.lookupObject (features[pos]); } public String toString () { StringBuffer sb = new StringBuffer (); for (int fsi = 0; fsi < length; fsi++) { Object o = dictionary.lookupObject(features[fsi]); sb.append (fsi); sb.append (": "); sb.append (o.toString()); sb.append (" ("); sb.append (features[fsi]); sb.append (")\n"); } return sb.toString(); } protected void growIfNecessary () { if (length == features.length) { int[] newFeatures = new int[features.length * 2]; System.arraycopy (features, 0, newFeatures, 0, length); features = newFeatures; } } public void add (int featureIndex) { growIfNecessary (); assert (featureIndex < dictionary.size()); features[length++] = featureIndex; } public void add (Object key) { int fi = dictionary.lookupIndex (key); if (fi >= 0) add (fi); // gdruck@cs.umass.edu // With the exception below, it is not possible to pipe data // when growth of the alphabet is stopped. We want to be // able to do this, for example to process new data using // an old Pipe (for example from a fixed, cached classifier // that we want to apply to new data.). //else // xxx Should we raise an exception if the appending doesn't happen? "yes" -akm, added 1/2008 // throw new IllegalStateException ("Object cannot be added to FeatureSequence because its Alphabet is frozen."); } public void addFeatureWeightsTo (double[] weights) { for (int i = 0; i < length; i++) weights[features[i]]++; } public void addFeatureWeightsTo (double[] weights, double scale) { for (int i = 0; i < length; i++) weights[features[i]] += scale; } public int[] toFeatureIndexSequence () { int[] feats = new int[length]; System.arraycopy (features, 0, feats, 0, length); return feats; } public int[] toSortedFeatureIndexSequence () { int[] feats = this.toFeatureIndexSequence (); java.util.Arrays.sort (feats); return feats; } /** * Remove features from the sequence that occur fewer than * <code>cutoff</code> times in the corpus, as indicated by * the provided counts. Also swap in the new, reduced alphabet. * This method alters the instance in place; it is not appropriate * if the original instance will be needed. */ public void prune (double[] counts, Alphabet newAlphabet, int cutoff) { // The goal is to replace the sequence of features in place, by // creating a new array and then swapping it in. // First: figure out how long the new array will have to be int newLength = 0; for (int i = 0; i < length; i++) { if (counts[features[i]] >= cutoff) { newLength++; } } // Second: allocate a new features array int[] newFeatures = new int[newLength]; // Third: fill the new array int newIndex = 0; for (int i = 0; i < length; i++) { if (counts[features[i]] >= cutoff) { Object feature = dictionary.lookupObject(features[i]); newFeatures[newIndex] = newAlphabet.lookupIndex(feature); newIndex++; } } // Fourth: swap out the arrays features = newFeatures; length = newLength; dictionary = newAlphabet; } // Serialization private static final long serialVersionUID = 1; private static final int CURRENT_SERIAL_VERSION = 0; private static final int NULL_INTEGER = -1; private void writeObject (ObjectOutputStream out) throws IOException { out.writeInt (CURRENT_SERIAL_VERSION); out.writeObject (dictionary); out.writeInt (features.length); for (int i = 0; i < features.length; i++) out.writeInt (features[i]); out.writeInt (length); } private void readObject (ObjectInputStream in) throws IOException, ClassNotFoundException { int featuresLength; int version = in.readInt (); dictionary = (Alphabet) in.readObject (); featuresLength = in.readInt(); features = new int[featuresLength]; for (int i = 0; i < featuresLength; i++) features[i] = in.readInt (); length = in.readInt (); } }