continue; final ChunkList chunks = chunker.chunk(Arrays.asList(s.split(" "))); PhraseTableImpl pt = new PhraseTableImpl(srcLang, trgLang, "std-table"); for (Chunk chunk : chunks) {
int i = 0; for (EntityLabel sp : corpus) { final ChunkList chunkList = setup.chunker(sp.entity).chunk(Arrays.asList(sp.srcLabel.split(" "))); final PhraseTableImpl pt = new PhraseTableImpl(setup.sourceLanguage(), setup.targetLanguage(), "mert_table"); for (Chunk chunk : chunkList) {
private List<Translation> doDecoding(TranslatorSetup setup, Decoder decoder, EntityLabel el, int options) { final ChunkList chunkList = setup.chunker(el.entity).chunk(tokenizer.tokenize(el.srcLabel)); final PhraseTableImpl pt = new PhraseTableImpl(setup.sourceLanguage(), setup.targetLanguage(), "mert_table"); for (Chunk chunk : chunkList) { for (TranslationSource source : setup.sources()) { pt.addAll(source.candidates(chunk)); } } PhraseTable rerankedTable = pt; for (TranslationFeaturizer featurizer : setup.featurizers(el.entity)) { try { rerankedTable = featurizer.featurize(rerankedTable, el.entity); } catch (Exception x) { } } final List<Translation> decoded = (options & OntologyTranslator.DECODE_FAST) == 0 ? decoder.decode(Arrays.asList(el.srcLabel.split("\\s+")), rerankedTable, setup.featureNames(), nBest) : decoder.decodeFast(Arrays.asList(el.srcLabel.split("\\s+")), rerankedTable, setup.featureNames(), nBest); return decoded; }