/**
* @return a list of forms of @param verb which match @param expectedVerbPOS (person:number)
* @param toUppercase true when the suggestions should be capitalized
*/
private List<String> getVerbSuggestions(
final AnalyzedTokenReadings verb, final String expectedVerbPOS, final boolean toUppercase) {
// find the first verb reading
AnalyzedToken verbToken = new AnalyzedToken("", "", "");
for (AnalyzedToken token : verb.getReadings()) {
if (token.getPOSTag().startsWith("VER:")) {
verbToken = token;
break;
}
}
try {
String[] synthesized =
german.getSynthesizer().synthesize(verbToken, "VER.*:" + expectedVerbPOS + ".*", true);
// remove duplicates
Set<String> suggestionSet = new HashSet<>();
suggestionSet.addAll(Arrays.asList(synthesized));
List<String> suggestions = new ArrayList<>();
suggestions.addAll(suggestionSet);
if (toUppercase) {
for (int i = 0; i < suggestions.size(); ++i) {
suggestions.set(i, StringTools.uppercaseFirstChar(suggestions.get(i)));
}
}
Collections.sort(suggestions);
return suggestions;
} catch (IOException e) {
throw new RuntimeException(e);
}
}
@Override
protected boolean isTagged(AnalyzedTokenReadings tokenReadings) {
for (AnalyzedToken token : tokenReadings.getReadings()) {
String posTag = token.getPOSTag();
if (isGoodPosTag(posTag)) {
return true;
}
}
return false;
}
/**
* @return true if the verb @param token (if it is a verb) matches @param person and @param
* number, and matches no other person/number
*/
private boolean hasUnambiguouslyPersonAndNumber(
final AnalyzedTokenReadings tokenReadings, final String person, final String number) {
if (tokenReadings.getToken().length() == 0
|| (Character.isUpperCase(tokenReadings.getToken().charAt(0))
&& !(tokenReadings.getStartPos() == 0))
|| !tokenReadings.hasPartialPosTag("VER")) return false;
for (AnalyzedToken analyzedToken : tokenReadings) {
final String postag = analyzedToken.getPOSTag();
if (postag.contains("_END")) // ignore SENT_END and PARA_END
continue;
if (!postag.contains(":" + person + ":" + number)) return false;
} // for each reading
return true;
}
@Override
public List<AnalyzedToken> additionalTags(String word) {
final IStemmer dictLookup;
try {
dictLookup = new DictionaryLookup(getDictionary());
} catch (IOException e) {
throw new RuntimeException("Could not load Catalan dictionary from " + getFileName(), e);
}
List<AnalyzedToken> additionalTaggedTokens = new ArrayList<>();
// Any well-formed adverb with suffix -ment is tagged as an adverb (RG)
// Adjectiu femení singular o participi femení singular + -ment
if (word.endsWith("ment")) {
final String lowerWord = word.toLowerCase(conversionLocale);
final String possibleAdj = lowerWord.replaceAll("^(.+)ment$", "$1");
List<AnalyzedToken> taggerTokens;
taggerTokens = asAnalyzedTokenList(possibleAdj, dictLookup.lookup(possibleAdj));
for (AnalyzedToken taggerToken : taggerTokens) {
final String posTag = taggerToken.getPOSTag();
if (posTag != null) {
final Matcher m = ADJ_PART_FS.matcher(posTag);
if (m.matches()) {
additionalTaggedTokens.add(new AnalyzedToken(word, "RG", lowerWord));
return additionalTaggedTokens;
}
}
}
}
// Any well-formed verb with prefixes is tagged as a verb copying the original tags
Matcher matcher = PREFIXES_FOR_VERBS.matcher(word);
if (matcher.matches()) {
final String possibleVerb = matcher.group(2).toLowerCase();
List<AnalyzedToken> taggerTokens;
taggerTokens = asAnalyzedTokenList(possibleVerb, dictLookup.lookup(possibleVerb));
for (AnalyzedToken taggerToken : taggerTokens) {
final String posTag = taggerToken.getPOSTag();
if (posTag != null) {
final Matcher m = VERB.matcher(posTag);
if (m.matches()) {
String lemma = matcher.group(1).toLowerCase().concat(taggerToken.getLemma());
additionalTaggedTokens.add(new AnalyzedToken(word, posTag, lemma));
}
}
}
return additionalTaggedTokens;
}
// Any well-formed noun with prefix ex- is tagged as a noun copying the original tags
if (word.startsWith("ex")) {
final String lowerWord = word.toLowerCase(conversionLocale);
final String possibleNoun = lowerWord.replaceAll("^ex(.+)$", "$1");
List<AnalyzedToken> taggerTokens;
taggerTokens = asAnalyzedTokenList(possibleNoun, dictLookup.lookup(possibleNoun));
for (AnalyzedToken taggerToken : taggerTokens) {
final String posTag = taggerToken.getPOSTag();
if (posTag != null) {
final Matcher m = NOUN.matcher(posTag);
if (m.matches()) {
String lemma = "ex".concat(taggerToken.getLemma());
additionalTaggedTokens.add(new AnalyzedToken(word, posTag, lemma));
}
}
}
return additionalTaggedTokens;
}
// Interpret deprecated characters of "ela geminada"
// U+013F LATIN CAPITAL LETTER L WITH MIDDLE DOT
// U+0140 LATIN SMALL LETTER L WITH MIDDLE DOT
if (word.contains("\u0140") || word.contains("\u013f")) {
final String lowerWord = word.toLowerCase(conversionLocale);
final String possibleWord = lowerWord.replaceAll("\u0140", "l·");
List<AnalyzedToken> taggerTokens = asAnalyzedTokenList(word, dictLookup.lookup(possibleWord));
return taggerTokens;
}
return null;
}
