public static List<SentenceBreaker.Sentence> cleanAndBreakSentences(String docText) {
// ACE IS EVIL
docText = docText.replaceAll("<\\S+>", "");
AlignedSub cleaner = AnalysisUtilities.cleanupDocument(docText);
List<SentenceBreaker.Sentence> sentences = SentenceBreaker.getSentences(cleaner);
return sentences;
}
private LabeledSentence generateSupersenseTaggingInput(Tree sentence) {
LabeledSentence res = new LabeledSentence();
List<Tree> leaves = sentence.getLeaves();
for (int i = 0; i < leaves.size(); i++) {
String word = leaves.get(i).label().toString();
Tree preterm = leaves.get(i).parent(sentence);
String pos = preterm.label().toString();
String stem = AnalysisUtilities.getInstance().getLemma(word, pos);
res.addToken(word, stem, pos, "0");
}
return res;
}