edu.stanford.nlp.ling.IndexedWord Maven / Gradle / Ivy
Show all versions of stanford-parser Show documentation
package edu.stanford.nlp.ling;
import edu.stanford.nlp.util.logging.Redwood;
import java.util.Objects;
import java.util.Set;
import edu.stanford.nlp.util.StringUtils;
import edu.stanford.nlp.util.TypesafeMap;
/**
* This class provides a {@link CoreLabel} that uses its
* DocIDAnnotation, SentenceIndexAnnotation, and IndexAnnotation to implement
* Comparable/compareTo, hashCode, and equals. This means no other annotations,
* including the identity of the word, are taken into account when using these
* methods. Historically, this class was introduced for and is mainly used in
* the RTE package, and it provides a number of methods that are really specific
* to that use case. A second use case is now the Stanford Dependencies code,
* where this class directly implements the "copy nodes" of section 4.6 of the
* Stanford Dependencies Manual, rather than these being placed directly in the
* backing CoreLabel. This was so there can stay one CoreLabel per token, despite
* there being multiple IndexedWord nodes, additional ones representing copy
* nodes.
*
* The actual implementation is to wrap a {@code CoreLabel}.
* This avoids breaking the {@code equals()} and
* {@code hashCode()} contract and also avoids expensive copying
* when used to represent the same data as the original
* {@code CoreLabel}.
*
* @author rafferty
* @author John Bauer
* @author Sonal Gupta
*/
public class IndexedWord implements AbstractCoreLabel, Comparable {
/** A logger for this class */
private static Redwood.RedwoodChannels log = Redwood.channels(IndexedWord.class);
private static final long serialVersionUID = 3739633991145239829L;
/**
* The identifier that points to no word.
*/
public static final IndexedWord NO_WORD = new IndexedWord(null, -1, -1);
private final CoreLabel label;
private int copyCount; // = 0;
private int numCopies = 0;
private IndexedWord original = null;
/**
* Useful for specifying a fine-grained position when butchering parse trees.
* The canonical use case for this is resolving coreference in the OpenIE system, where
* we want to move nodes between sentences, but do not want to change their index annotation
* (plus, we need to have multiple nodes fit into the space of one pronoun).
*/
private double pseudoPosition = Double.NaN;
/**
* Default constructor; uses {@link CoreLabel} default constructor
*/
public IndexedWord() {
label = new CoreLabel();
}
/**
* Copy Constructor - relies on {@link CoreLabel} copy constructor
* It will set the value, and if the word is not set otherwise, set
* the word to the value.
*
* @param w A Label to initialize this IndexedWord from
*/
public IndexedWord(Label w) {
if (w instanceof CoreLabel) {
this.label = (CoreLabel) w;
} else {
label = new CoreLabel(w);
if (label.word() == null) {
label.setWord(label.value());
}
}
}
/**
* Construct an IndexedWord from a CoreLabel just as for a CoreMap.
* Implementation note: this is a the same as the constructor
* that takes a CoreMap, but is needed to ensure unique most specific
* type inference for selecting a constructor at compile-time.
*
* @param w A Label to initialize this IndexedWord from
*/
public IndexedWord(CoreLabel w) {
label = w;
}
/**
* Constructor for setting docID, sentenceIndex, and
* index without any other annotations.
*
* @param docID The document ID (arbitrary string)
* @param sentenceIndex The sentence number in the document (normally 0-based)
* @param index The index of the word in the sentence (normally 0-based)
*/
public IndexedWord(String docID, int sentenceIndex, int index) {
label = new CoreLabel();
label.set(CoreAnnotations.DocIDAnnotation.class, docID);
label.set(CoreAnnotations.SentenceIndexAnnotation.class, sentenceIndex);
label.set(CoreAnnotations.IndexAnnotation.class, index);
}
public IndexedWord makeCopy(int count) {
CoreLabel labelCopy = new CoreLabel(label);
IndexedWord copy = new IndexedWord(labelCopy);
copy.setCopyCount(count);
return copy;
}
public IndexedWord makeCopy() {
return makeCopy(++numCopies);
}
public IndexedWord makeSoftCopy(int count) {
IndexedWord copy = new IndexedWord(label);
copy.setCopyCount(count);
copy.original = this;
return copy;
}
public IndexedWord makeSoftCopy() {
if (original != null) {
return original.makeSoftCopy();
} else {
return makeSoftCopy(++numCopies);
}
}
public IndexedWord getOriginal() {
return original;
}
/**
* TODO: get rid of this. Only used in two places in RTE (in rewriter code)
*/
public CoreLabel backingLabel() { return label; }
@Override
public VALUE get(Class> key) {
return label.get(key);
}
@Override
public boolean containsKey(Class> key) {
return label.containsKey(key);
}
@Override
public VALUE set(Class> key, VALUE value) {
return label.set(key, value);
}
@Override
public > String getString(Class key) {
return label.getString(key);
}
@Override
public > String getString(Class key, String def) {
return label.getString(key, def);
}
@Override
public VALUE remove(Class> key) {
return label.remove(key);
}
@Override
public Set> keySet() {
return label.keySet();
}
@Override
public int size() {
return label.size();
}
@Override
public String value() {
return label.value();
}
@Override
public void setValue(String value) {
label.setValue(value);
}
@Override
public String tag() {
return label.tag();
}
@Override
public void setTag(String tag) {
label.setTag(tag);
}
@Override
public String word() {
return label.word();
}
@Override
public void setWord(String word) {
label.setWord(word);
}
@Override
public String lemma() {
return label.lemma();
}
@Override
public void setLemma(String lemma) {
label.setLemma(lemma);
}
@Override
public String ner() {
return label.ner();
}
@Override
public void setNER(String ner) {
label.setNER(ner);
}
@Override
public String docID() {
return label.docID();
}
@Override
public void setDocID(String docID) {
label.setDocID(docID);
}
@Override
public int index() {
return label.index();
}
@Override
public void setIndex(int index) {
label.setIndex(index);
}
/**
* In most cases, this is just the index of the word.
* However, this should be the value used to sort nodes in
* a tree.
*
* @see IndexedWord#pseudoPosition
*/
public double pseudoPosition() {
if (!Double.isNaN(pseudoPosition)) {
return pseudoPosition;
} else {
return (double) index();
}
}
/**
* @see IndexedWord#pseudoPosition
*/
public void setPseudoPosition(double position) {
this.pseudoPosition = position;
}
@Override
public int sentIndex() {
return label.sentIndex();
}
@Override
public void setSentIndex(int sentIndex) {
label.setSentIndex(sentIndex);
}
@Override
public String originalText() {
return label.originalText();
}
@Override
public void setOriginalText(String originalText) {
label.setOriginalText(originalText);
}
@Override
public int beginPosition() {
return label.beginPosition();
}
@Override
public int endPosition() {
return label.endPosition();
}
@Override
public void setBeginPosition(int beginPos) {
label.setBeginPosition(beginPos);
}
@Override
public void setEndPosition(int endPos) {
label.setEndPosition(endPos);
}
public int copyCount() {
return copyCount;
}
public void setCopyCount(int count) {
this.copyCount = count;
}
public String toPrimes() {
return StringUtils.repeat('\'', copyCount);
}
public boolean isCopy(IndexedWord otherWord) {
Integer myInd = get(CoreAnnotations.IndexAnnotation.class);
Integer otherInd = otherWord.get(CoreAnnotations.IndexAnnotation.class);
if ( ! Objects.equals(myInd, otherInd)) {
return false;
}
Integer mySentInd = get(CoreAnnotations.SentenceIndexAnnotation.class);
Integer otherSentInd = otherWord.get(CoreAnnotations.SentenceIndexAnnotation.class);
if ( ! Objects.equals(mySentInd, otherSentInd)) {
return false;
}
String myDocID = getString(CoreAnnotations.DocIDAnnotation.class);
String otherDocID = otherWord.getString(CoreAnnotations.DocIDAnnotation.class);
if ( ! Objects.equals(myDocID, otherDocID)) {
return false;
}
if (copyCount() == 0 || otherWord.copyCount() != 0) {
return false;
}
return true;
}
/**
* This .equals is dependent only on docID, sentenceIndex, and index.
* It doesn't consider the actual word value, but assumes that it is
* validly represented by token position.
* All IndexedWords that lack these fields will be regarded as equal.
*/
@Override
public boolean equals(Object o) {
if (this == o) return true;
if (!(o instanceof IndexedWord)) return false;
//now compare on appropriate keys
final IndexedWord otherWord = (IndexedWord) o;
Integer myInd = get(CoreAnnotations.IndexAnnotation.class);
Integer otherInd = otherWord.get(CoreAnnotations.IndexAnnotation.class);
if ( ! Objects.equals(myInd, otherInd)) {
return false;
}
Integer mySentInd = get(CoreAnnotations.SentenceIndexAnnotation.class);
Integer otherSentInd = otherWord.get(CoreAnnotations.SentenceIndexAnnotation.class);
if ( ! Objects.equals(mySentInd, otherSentInd)) {
return false;
}
String myDocID = getString(CoreAnnotations.DocIDAnnotation.class);
String otherDocID = otherWord.getString(CoreAnnotations.DocIDAnnotation.class);
if ( ! Objects.equals(myDocID, otherDocID)) {
return false;
}
if (copyCount() != otherWord.copyCount()) {
return false;
}
// Compare pseudo-positions
if ( (!Double.isNaN(this.pseudoPosition) || !Double.isNaN(otherWord.pseudoPosition)) &&
this.pseudoPosition != otherWord.pseudoPosition) {
return false;
}
return true;
}
private int cachedHashCode = 0;
/**
* This hashCode uses only the docID, sentenceIndex, and index.
* See compareTo for more info.
*/
@Override
public int hashCode() {
if (cachedHashCode != 0) {
return cachedHashCode;
}
boolean sensible = false;
int result = 0;
if (get(CoreAnnotations.DocIDAnnotation.class) != null) {
result = get(CoreAnnotations.DocIDAnnotation.class).hashCode();
sensible = true;
}
if (containsKey(CoreAnnotations.SentenceIndexAnnotation.class)) {
result = 29 * result + get(CoreAnnotations.SentenceIndexAnnotation.class).hashCode();
sensible = true;
}
if (containsKey(CoreAnnotations.IndexAnnotation.class)) {
result = 29 * result + get(CoreAnnotations.IndexAnnotation.class).hashCode();
sensible = true;
}
if ( ! sensible) {
log.info("WARNING!!! You have hashed an IndexedWord with no docID, sentIndex or wordIndex. You will almost certainly lose");
}
cachedHashCode = result;
return result;
}
/**
* NOTE: This compareTo is based on and made to be compatible with the one
* from IndexedFeatureLabel. You must have a DocIDAnnotation,
* SentenceIndexAnnotation, and IndexAnnotation for this to make sense and
* be guaranteed to work properly. Currently, it won't error out and will
* try to return something sensible if these are not defined, but that really
* isn't proper usage!
*
* This compareTo method is based not by value elements like the word(),
* but on passage position. It puts NO_WORD elements first, and then orders
* by document, sentence, and word index. If these do not differ, it
* returns equal.
*
* @param w The IndexedWord to compare with
* @return Whether this is less than w or not in the ordering
*/
@Override
public int compareTo(IndexedWord w) {
if (this.equals(IndexedWord.NO_WORD)) {
if (w.equals(IndexedWord.NO_WORD)) {
return 0;
} else {
return -1;
}
}
if (w.equals(IndexedWord.NO_WORD)) {
return 1;
}
// Override the default comparator if pseudo-positions are set.
// This is needed for splicing trees together awkwardly in OpenIE.
if (!Double.isNaN(w.pseudoPosition) || !Double.isNaN(this.pseudoPosition)) {
double val = this.pseudoPosition() - w.pseudoPosition();
if (val < 0) { return -1; }
if (val > 0) { return 1; }
else { return 0; }
}
// Otherwise, compare using the normal doc/sentence/token index hierarchy
String docID = this.getString(CoreAnnotations.DocIDAnnotation.class);
int docComp = docID.compareTo(w.getString(CoreAnnotations.DocIDAnnotation.class));
if (docComp != 0) return docComp;
int sentComp = sentIndex() - w.sentIndex();
if (sentComp != 0) return sentComp;
int indexComp = index() - w.index();
if (indexComp != 0) return indexComp;
return copyCount() - w.copyCount();
}
/**
* Returns the value-tag of this label.
*/
@Override
public String toString() {
return toString(CoreLabel.OutputFormat.VALUE_TAG);
}
public String toString(CoreLabel.OutputFormat format) {
return label.toString(format) + toPrimes();
}
/**
* {@inheritDoc}
*/
@Override
public void setFromString(String labelStr) {
throw new UnsupportedOperationException("Cannot set from string");
}
public static LabelFactory factory() {
return new LabelFactory() {
@Override
public Label newLabel(String labelStr) {
CoreLabel coreLabel = new CoreLabel();
coreLabel.setValue(labelStr);
return new IndexedWord(coreLabel);
}
@Override
public Label newLabel(String labelStr, int options) {
return newLabel(labelStr);
}
@Override
public Label newLabel(Label oldLabel) {
return new IndexedWord(oldLabel);
}
@Override
public Label newLabelFromString(String encodedLabelStr) {
throw new UnsupportedOperationException("This code branch left blank" +
" because we do not understand what this method should do.");
}
};
}
/**
* {@inheritDoc}
*/
@Override
public LabelFactory labelFactory() {
return IndexedWord.factory();
}
}