/**
* Remove traces and non-terminal decorations (e.g., "-SUBJ" in "NP-SUBJ") from a Penn
* Treebank-style tree.
*
* @param inputTree
*/
public void normalizeTree(Tree inputTree) {
inputTree.label().setFromString("ROOT");
List<Pair<TregexPattern, TsurgeonPattern>> ops =
new ArrayList<Pair<TregexPattern, TsurgeonPattern>>();
List<TsurgeonPattern> ps = new ArrayList<TsurgeonPattern>();
String tregexOpStr;
TregexPattern matchPattern;
TsurgeonPattern p;
TregexMatcher matcher;
tregexOpStr = "/\\-NONE\\-/=emptynode";
matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
matcher = matchPattern.matcher(inputTree);
ps.add(Tsurgeon.parseOperation("prune emptynode"));
matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
p = Tsurgeon.collectOperations(ps);
ops.add(new Pair<TregexPattern, TsurgeonPattern>(matchPattern, p));
Tsurgeon.processPatternsOnTree(ops, inputTree);
Label nonterminalLabel;
tregexOpStr = "/.+\\-.+/=nonterminal < __";
matchPattern = TregexPatternFactory.getPattern(tregexOpStr);
matcher = matchPattern.matcher(inputTree);
while (matcher.find()) {
nonterminalLabel = matcher.getNode("nonterminal");
if (nonterminalLabel == null) continue;
nonterminalLabel.setFromString(tlp.basicCategory(nonterminalLabel.value()));
}
}
/**
* Tries to match a pattern against a tree. If it succeeds, apply the surgical operations
* contained in a {@link TsurgeonPattern}.
*
* @param matchPattern A {@link TregexPattern} to be matched against a {@link Tree}.
* @param p A {@link TsurgeonPattern} to apply.
* @param t the {@link Tree} to match against and perform surgery on.
* @return t, which has been surgically modified.
*/
public static Tree processPattern(TregexPattern matchPattern, TsurgeonPattern p, Tree t) {
TregexMatcher m = matchPattern.matcher(t);
while (m.find()) {
t = p.evaluate(t, m);
if (t == null) {
break;
}
m = matchPattern.matcher(t);
}
return t;
}
public static boolean filter(PhraseInfo phrase) {
if (phrase == null || phrase.getTree() == null) return false;
Tree phraseTree = phrase.getTree();
// __ < ( NP < (PRP !< it|them) ) | < ( NP < ( NP < (PRP !< it|them) ) )
String invalidPronounPattern =
" ( NP < (PRP=prp !< " + Utils.wordsConjuction(Rules.VALID_PRONOUNS) + ") ) ";
String filterPattern =
"__ < " + invalidPronounPattern + " | < ( NP <" + invalidPronounPattern + ")";
TregexPattern tregexPattern = TregexPattern.compile(filterPattern);
TregexMatcher matcher = tregexPattern.matcher(phraseTree);
if (matcher.matches()) {
Proof proof = new Proof(ProofType.FAIL_PERSONAL_PRONOUN);
Tree evdTree = matcher.getNode("prp");
proof.setEvidenceTree(evdTree);
phrase.addProof(proof);
return false;
}
return true;
}
// todo: add an option to only print each tree once, regardless. Most useful in conjunction
// with -w
public void visitTree(Tree t) {
treeNumber++;
if (printTree) {
pw.print(treeNumber + ":");
pw.println("Next tree read:");
tp.printTree(t, pw);
}
TregexMatcher match = p.matcher(t);
if (printNonMatchingTrees) {
if (match.find()) numMatches++;
else tp.printTree(t, pw);
return;
}
Tree lastMatchingRootNode = null;
while (match.find()) {
if (oneMatchPerRootNode) {
if (lastMatchingRootNode == match.getMatch()) continue;
else lastMatchingRootNode = match.getMatch();
}
numMatches++;
if (printFilename && treebank instanceof DiskTreebank) {
DiskTreebank dtb = (DiskTreebank) treebank;
pw.print("# ");
pw.println(dtb.getCurrentFilename());
}
if (printSubtreeCode) {
pw.print(treeNumber);
pw.print(':');
pw.println(match.getMatch().nodeNumber(t));
}
if (printMatches) {
if (reportTreeNumbers) {
pw.print(treeNumber);
pw.print(": ");
}
if (printTree) {
pw.println("Found a full match:");
}
if (printWholeTree) {
tp.printTree(t, pw);
} else if (handles != null) {
if (printTree) {
pw.println("Here's the node you were interested in:");
}
for (String handle : handles) {
Tree labeledNode = match.getNode(handle);
if (labeledNode == null) {
System.err.println(
"Error!! There is no matched node \""
+ handle
+ "\"! Did you specify such a label in the pattern?");
} else {
tp.printTree(labeledNode, pw);
}
}
} else {
tp.printTree(match.getMatch(), pw);
}
// pw.println(); // TreePrint already puts a blank line in
} // end if (printMatches)
} // end while match.find()
} // end visitTree
