private boolean lookup() {
/**
* **** debugging code String thisLookup = word.toString(); boolean added =
* lookups.add(thisLookup); if (!added) { System.out.println("######extra lookup:" +
* thisLookup); // occaasional extra lookups aren't necessarily errors... could happen by diff
* manipulations // throw new RuntimeException("######extra lookup:" + thisLookup); } else { //
* System.out.println("new lookup:" + thisLookup); } ****
*/
matchedEntry = dict_ht.get(word.getArray(), 0, word.size());
return matchedEntry != null;
}
private DictEntry wordInDict() {
/**
* * if (matchedEntry != null) { if (dict_ht.get(word.getArray(), 0, word.size()) !=
* matchedEntry) { System.out.println("Uh oh... cached entry doesn't match"); } return
* matchedEntry; } *
*/
if (matchedEntry != null) return matchedEntry;
DictEntry e = dict_ht.get(word.getArray(), 0, word.length());
if (e != null && !e.exception) {
matchedEntry = e; // only cache if it's not an exception.
}
// lookups.add(word.toString());
return e;
}
private CharArrayMap convertPhraseSet(CharArraySet phraseSet) {
CharArrayMap<CharArraySet> phraseMap = new CharArrayMap(100, false);
Iterator<Object> phraseIt = phraseSet.iterator();
while (phraseIt != null && phraseIt.hasNext()) {
char[] phrase = (char[]) phraseIt.next();
Log.debug("'" + new String(phrase) + "'");
char[] firstTerm = getFirstTerm(phrase);
Log.debug("'" + new String(firstTerm) + "'");
CharArraySet itsPhrases = phraseMap.get(firstTerm, 0, firstTerm.length);
if (itsPhrases == null) {
itsPhrases = new CharArraySet(5, false);
phraseMap.put(new String(firstTerm), itsPhrases);
}
itsPhrases.add(phrase);
}
return phraseMap;
}
@Override
public boolean incrementToken() throws IOException {
if (!emitSingleTokens && unusedTokens.size() > 0) {
Log.debug("emitting unused phrases");
// emit these until the queue is empty before emitting any new stuff
Token aToken = unusedTokens.remove(0);
emit(aToken);
return true;
}
if (lastToken != null) {
emit(lastToken);
lastToken = null;
return true;
}
char[] nextToken = nextToken();
// if (nextToken != null) System.out.println( "nextToken: " + new String( nextToken ));
if (nextToken == null) {
if (lastValid != null) {
emit(lastValid);
lastValid = null;
return true;
}
if (emitSingleTokens && currentSetToCheck != null && currentSetToCheck.size() > 0) {
char[] phrase = getFirst(currentSetToCheck);
char[] lastTok = getCurrentBuffer(new char[0]);
if (phrase != null && endsWith(lastTok, phrase)) {
currentSetToCheck = remove(currentSetToCheck, phrase);
emit(phrase);
return true;
}
} else if (!emitSingleTokens && currentSetToCheck != null && currentSetToCheck.size() > 0) {
if (lastEmitted != null
&& !equals(fixWhitespace(lastEmitted), getCurrentBuffer(new char[0]))) {
discardCharTokens(currentPhrase, unusedTokens);
currentSetToCheck = null;
if (unusedTokens.size() > 0) {
Token aToken = unusedTokens.remove(0);
Log.debug("emitting putback token");
emit(aToken);
return true;
}
}
}
if (lastEmitted == null && (currentPhrase != null && currentPhrase.length() > 0)) {
char[] lastTok = getCurrentBuffer(new char[0]);
if (currentSetToCheck.contains(lastTok, 0, lastTok.length)) {
emit(lastTok);
currentPhrase.setLength(0);
return true;
} else if (!emitSingleTokens) {
discardCharTokens(currentPhrase, unusedTokens);
currentSetToCheck = null;
currentPhrase.setLength(0);
if (unusedTokens.size() > 0) {
Token aToken = unusedTokens.remove(0);
Log.debug("emitting putback token");
emit(aToken);
return true;
}
}
}
return false;
}
// if emitSingleToken, set lastToken = nextToken
if (emitSingleTokens) {
lastToken = nextToken;
}
if (currentSetToCheck == null || currentSetToCheck.size() == 0) {
Log.debug("Checking for phrase start on '" + new String(nextToken) + "'");
if (phraseMap.keySet().contains(nextToken, 0, nextToken.length)) {
// get the phrase set for this token, add it to currentSetTocheck
currentSetToCheck = phraseMap.get(nextToken, 0, nextToken.length);
if (currentPhrase == null) currentPhrase = new StringBuffer();
else currentPhrase.setLength(0);
currentPhrase.append(nextToken);
return incrementToken();
} else {
emit(nextToken);
// clear lastToken
lastToken = null;
return true;
}
} else {
// add token to the current string buffer.
char[] currentBuffer = getCurrentBuffer(nextToken);
if (currentSetToCheck.contains(currentBuffer, 0, currentBuffer.length)) {
// if its the only one valid, emit it
// if there is a longer one, wait to see if it will be matched
// if the longer one breaks on the next token, emit this one...
// emit the current phrase
currentSetToCheck = remove(currentSetToCheck, currentBuffer);
if (currentSetToCheck.size() == 0) {
emit(currentBuffer);
lastValid = null;
--positionIncr;
} else {
if (emitSingleTokens) {
lastToken = currentBuffer;
return true;
}
lastValid = currentBuffer;
}
if (phraseMap.keySet().contains(nextToken, 0, nextToken.length)) {
// get the phrase set for this token, add it to currentPhrasesTocheck
currentSetToCheck = phraseMap.get(nextToken, 0, nextToken.length);
if (currentPhrase == null) currentPhrase = new StringBuffer();
else currentPhrase.setLength(0);
currentPhrase.append(nextToken);
}
return (lastValid != null) ? incrementToken() : true;
}
if (phraseMap.keySet().contains(nextToken, 0, nextToken.length)) {
// get the phrase set for this token, add it to currentPhrasesTocheck
// System.out.println( "starting new phrase with " + new String( nextToken ) );
// does this add all of the set? if not need iterator loop
CharArraySet newSet = phraseMap.get(nextToken, 0, nextToken.length);
Iterator<Object> phraseIt = newSet.iterator();
while (phraseIt != null && phraseIt.hasNext()) {
char[] phrase = (char[]) phraseIt.next();
currentSetToCheck.add(phrase);
}
}
// for each phrase in currentSetToCheck -
// if there is a phrase prefix match, get the next token recursively
Iterator<Object> phraseIt = currentSetToCheck.iterator();
while (phraseIt != null && phraseIt.hasNext()) {
char[] phrase = (char[]) phraseIt.next();
if (startsWith(phrase, currentBuffer)) {
return incrementToken();
}
}
if (lastValid != null) {
emit(lastValid);
lastValid = null;
return true;
}
if (!emitSingleTokens) {
// current phrase didn't match fully: put the tokens back
// into the unusedTokens list
discardCharTokens(currentPhrase, unusedTokens);
currentPhrase.setLength(0);
currentSetToCheck = null;
if (unusedTokens.size() > 0) {
Token aToken = unusedTokens.remove(0);
Log.debug("emitting putback token");
emit(aToken);
return true;
}
}
currentSetToCheck = null;
Log.debug("returning at end.");
return incrementToken();
}
}
/** Stems the text in the token. Returns true if changed. */
boolean stem(char[] term, int len) {
result = null;
k = len - 1;
if ((k <= 1) || (k >= MaxWordLen - 1)) {
return false; // don't stem
}
// first check the stemmer dictionaries, and avoid using the
// cache if it's in there.
DictEntry entry = dict_ht.get(term, 0, len);
if (entry != null) {
if (entry.root != null) {
result = entry.root;
return true;
}
return false;
}
/**
* * caching off is normally faster if (cache == null) initializeStemHash();
*
* <p>// now check the cache, before we copy chars to "word" if (cache != null) { String val =
* cache.get(term, 0, len); if (val != null) { if (val != SAME) { result = val; return true; }
* return false; } } *
*/
word.reset();
// allocate enough space so that an expansion is never needed
word.reserve(len + 10);
for (int i = 0; i < len; i++) {
char ch = term[i];
if (!isAlpha(ch)) return false; // don't stem
// don't lowercase... it's a requirement that lowercase filter be
// used before this stemmer.
word.unsafeWrite(ch);
}
matchedEntry = null;
/** * lookups.clear(); lookups.add(word.toString()); * */
/*
* This while loop will never be executed more than one time; it is here
* only to allow the break statement to be used to escape as soon as a word
* is recognized
*/
while (true) {
// YCS: extra lookup()s were inserted so we don't need to
// do an extra wordInDict() here.
plural();
if (matched()) break;
pastTense();
if (matched()) break;
aspect();
if (matched()) break;
ityEndings();
if (matched()) break;
nessEndings();
if (matched()) break;
ionEndings();
if (matched()) break;
erAndOrEndings();
if (matched()) break;
lyEndings();
if (matched()) break;
alEndings();
if (matched()) break;
entry = wordInDict();
iveEndings();
if (matched()) break;
izeEndings();
if (matched()) break;
mentEndings();
if (matched()) break;
bleEndings();
if (matched()) break;
ismEndings();
if (matched()) break;
icEndings();
if (matched()) break;
ncyEndings();
if (matched()) break;
nceEndings();
matched();
break;
}
/*
* try for a direct mapping (allows for cases like `Italian'->`Italy' and
* `Italians'->`Italy')
*/
entry = matchedEntry;
if (entry != null) {
result = entry.root; // may be null, which means that "word" is the stem
}
/**
* * caching off is normally faster if (cache != null && cache.size() < maxCacheSize) { char[]
* key = new char[len]; System.arraycopy(term, 0, key, 0, len); if (result != null) {
* cache.put(key, result); } else { cache.put(key, word.toString()); } } *
*/
/**
* * if (entry == null) { if (!word.toString().equals(new String(term,0,len))) {
* System.out.println("CASE:" + word.toString() + "," + new String(term,0,len));
*
* <p>} } *
*/
// no entry matched means result is "word"
return true;
}