edu.stanford.nlp.parser.lexparser.CollinsPuncTransformer Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of stanford-corenlp Show documentation
Show all versions of stanford-corenlp Show documentation
Stanford CoreNLP provides a set of natural language analysis tools which can take raw English language text input and give the base forms of words, their parts of speech, whether they are names of companies, people, etc., normalize dates, times, and numeric quantities, mark up the structure of sentences in terms of phrases and word dependencies, and indicate which noun phrases refer to the same entities. It provides the foundational building blocks for higher level text understanding applications.
package edu.stanford.nlp.parser.lexparser;
import java.util.*;
import edu.stanford.nlp.trees.TreeTransformer;
import edu.stanford.nlp.trees.TreebankLanguagePack;
import edu.stanford.nlp.trees.Tree;
import edu.stanford.nlp.trees.TreeFactory;
/** This class manipulates punctuation in trees (used with training trees)
* in the same manner that Collins manipulated punctuation in trees when
* building his parsing model. This is the same punctuation that is
* the punctuation ignored in the standard EvalB evaluation is promoted
* as high in the tree as possible.
*
* @author Dan Klein
* @author Christopher Manning
*/
public class CollinsPuncTransformer implements TreeTransformer {
private TreebankLanguagePack tlp;
boolean isPunc(Tree t) {
if (t.isPreTerminal()) {
String s = t.label().value();
if (tlp.isEvalBIgnoredPunctuationTag(s)) {
return true;
}
}
return false;
}
static LinkedList preTerms(Tree t) {
LinkedList l = new LinkedList<>();
preTermHelper(t, l);
return l;
}
static void preTermHelper(Tree t, List l) {
if (t.isLeaf()) {
return;
}
if (t.isPreTerminal()) {
l.add(t);
return;
}
Tree[] children = t.children();
for (Tree child : children) {
preTermHelper(child, l);
}
}
Tree transformRoot(Tree tree, TreeFactory tf) {
// XXXX TODO: use tlp and don't assume 1 daughter of ROOT!
// leave the root intact
// if (tlp.isStartSymbol(tlp.basicCategory(tree.label().value())))
if (tree.label().toString().startsWith("ROOT")) {
return tf.newTreeNode(tree.label(), Collections.singletonList(transformNode(tree.children()[0], tf)));
}
return transformNode(tree, tf);
}
Tree transformNode(Tree tree, TreeFactory tf) {
if (tree.isLeaf()) {
return tf.newLeaf(tree.label());
}
if (tree.isPreTerminal()) {
return tf.newTreeNode(tree.label(), Collections.singletonList(tf.newLeaf(tree.children()[0].label())));
}
List children = tree.getChildrenAsList();
LinkedList newChildren = new LinkedList<>();
// promote lower punctuation
for (Tree child : children) {
LinkedList preTerms = preTerms(child);
while (!preTerms.isEmpty() && isPunc(preTerms.getFirst())) {
newChildren.add(preTerms.getFirst());
preTerms.removeFirst();
}
Tree newChild = transformNode(child, tf);
LinkedList temp = new LinkedList<>();
if (newChild.children().length > 0) {
newChildren.add(newChild);
}
while (!preTerms.isEmpty() && isPunc(preTerms.getLast())) {
temp.addFirst(preTerms.getLast());
preTerms.removeLast();
}
newChildren.addAll(temp);
}
// remove local punctuation
while (!newChildren.isEmpty() && isPunc(newChildren.getFirst())) {
newChildren.removeFirst();
}
while (!newChildren.isEmpty() && isPunc(newChildren.getLast())) {
newChildren.removeLast();
}
return tf.newTreeNode(tree.label(), newChildren);
}
// public Tree transformTree(Tree tree) {
// //System.out.println("PUNCTUATION TRANSFORM:");
// //tree.pennPrint();
// //System.out.println("BECOMES:");
// //transformRoot(tree, tf).pennPrint();
// return transformRoot(tree, tf);
// }
public Tree transformTree(Tree tree) {
return transformRoot(tree, tree.treeFactory());
}
public CollinsPuncTransformer(TreebankLanguagePack tlp) {
this.tlp = tlp;
}
}