@Override
protected TokenStreamComponents createComponents(final String fieldName, final Reader reader) {
final StreamLemmasFilter src =
new StreamLemmasFilter(reader, hebMorphLemmatizer, commonWords, lemmaFilter);
src.setAlwaysSaveMarkedOriginal(alwaysSaveMarkedOriginal);
src.setSuffixForExactMatch(suffixForExactMatch);
TokenStream tok = new SynonymFilter(src, acronymMergingMap, false);
if (commonWords != null && commonWords.size() > 0)
tok = new CommonGramsFilter(matchVersion, tok, commonWords, false);
tok = new SuffixKeywordFilter(tok, '$');
return new TokenStreamComponents(src, tok) {
@Override
protected void setReader(final Reader reader) throws IOException {
super.setReader(reader);
}
};
}
@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();
}
}