/*
* Licensed to the Apache Software Foundation (ASF) under one or more
* contributor license agreements. See the NOTICE file distributed with
* this work for additional information regarding copyright ownership.
* The ASF licenses this file to You under the Apache License, Version 2.0
* (the "License"); you may not use this file except in compliance with
* the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.apache.lucene.analysis.pattern;
import java.io.IOException;
import org.apache.lucene.analysis.Tokenizer;
import org.apache.lucene.analysis.tokenattributes.CharTermAttribute;
import org.apache.lucene.analysis.tokenattributes.OffsetAttribute;
import org.apache.lucene.util.ArrayUtil;
import org.apache.lucene.util.AttributeFactory;
import org.apache.lucene.util.automaton.Automaton;
import org.apache.lucene.util.automaton.CharacterRunAutomaton;
import org.apache.lucene.util.automaton.Operations;
import org.apache.lucene.util.automaton.RegExp;
/**
* This tokenizer uses a Lucene {@link RegExp} or (expert usage) a pre-built determinized {@link Automaton}, to locate tokens.
* The regexp syntax is more limited than {@link PatternTokenizer}, but the tokenization is quite a bit faster. The provided
* regex should match valid token characters (not token separator characters, like {@code String.split}). The matching is greedy:
* the longest match at a given start point will be the next token. Empty string tokens are never produced.
*
* @lucene.experimental
*/
// TODO: the matcher here is naive and does have N^2 adversarial cases that are unlikely to arise in practice, e.g. if the pattern is
// aaaaaaaaaab and the input is aaaaaaaaaaa, the work we do here is N^2 where N is the number of a's. This is because on failing to match
// a token, we skip one character forward and try again. A better approach would be to compile something like this regexp
// instead: .* | <pattern>, because that automaton would not "forget" all the as it had already seen, and would be a single pass
// through the input. I think this is the same thing as Aho/Corasick's algorithm (http://en.wikipedia.org/wiki/Aho%E2%80%93Corasick_string_matching_algorithm).
// But we cannot implement this (I think?) until/unless Lucene regexps support sub-group capture, so we could know
// which specific characters the pattern matched. SynonymFilter has this same limitation.
public final class SimplePatternTokenizer extends Tokenizer {
private final CharTermAttribute termAtt = addAttribute(CharTermAttribute.class);
private final OffsetAttribute offsetAtt = addAttribute(OffsetAttribute.class);
private final CharacterRunAutomaton runDFA;
// TODO: we could likely use a single rolling buffer instead of two separate char buffers here. We could also use PushBackReader but I
// suspect it's slowish:
private char[] pendingChars = new char[8];
private int pendingLimit;
private int pendingUpto;
private int offset;
private int tokenUpto;
private final char[] buffer = new char[1024];
private int bufferLimit;
private int bufferNextRead;
/** See {@link RegExp} for the accepted syntax. */
public SimplePatternTokenizer(String regexp) {
this(DEFAULT_TOKEN_ATTRIBUTE_FACTORY, regexp, Operations.DEFAULT_MAX_DETERMINIZED_STATES);
}
/** Runs a pre-built automaton. */
public SimplePatternTokenizer(Automaton dfa) {
this(DEFAULT_TOKEN_ATTRIBUTE_FACTORY, dfa);
}
/** See {@link RegExp} for the accepted syntax. */
public SimplePatternTokenizer(AttributeFactory factory, String regexp, int maxDeterminizedStates) {
this(factory, new RegExp(regexp).toAutomaton());
}
/** Runs a pre-built automaton. */
public SimplePatternTokenizer(AttributeFactory factory, Automaton dfa) {
super(factory);
// we require user to do this up front because it is a possibly very costly operation, and user may be creating us frequently, not
// realizing this ctor is otherwise trappy
if (dfa.isDeterministic() == false) {
throw new IllegalArgumentException("please determinize the incoming automaton first");
}
runDFA = new CharacterRunAutomaton(dfa, Operations.DEFAULT_MAX_DETERMINIZED_STATES);
}
@Override
public boolean incrementToken() throws IOException {
clearAttributes();
tokenUpto = 0;
while (true) {
int offsetStart = offset;
// The runDFA operates in Unicode space, not UTF16 (java's char):
int ch = nextCodePoint();
if (ch == -1) {
return false;
}
int state = runDFA.step(0, ch);
if (state != -1) {
// a token just possibly started; keep scanning to see if the token is accepted:
int lastAcceptLength = -1;
do {
if (runDFA.isAccept(state)) {
// record that the token matches here, but keep scanning in case a longer match also works (greedy):
lastAcceptLength = tokenUpto;
}
ch = nextCodePoint();
if (ch == -1) {
break;
}
state = runDFA.step(state, ch);
} while (state != -1);
if (lastAcceptLength != -1) {
// we found a token
int extra = tokenUpto - lastAcceptLength;
if (extra != 0) {
pushBack(extra);
}
termAtt.setLength(lastAcceptLength);
offsetAtt.setOffset(correctOffset(offsetStart), correctOffset(offsetStart+lastAcceptLength));
return true;
} else if (ch == -1) {
return false;
} else {
// false alarm: there was no token here; push back all but the first character we scanned
pushBack(tokenUpto-1);
tokenUpto = 0;
}
} else {
tokenUpto = 0;
}
}
}
@Override
public void end() throws IOException {
super.end();
final int ofs = correctOffset(offset + pendingLimit - pendingUpto);
offsetAtt.setOffset(ofs, ofs);
}
@Override
public void reset() throws IOException {
super.reset();
offset = 0;
pendingUpto = 0;
pendingLimit = 0;
tokenUpto = 0;
bufferNextRead = 0;
bufferLimit = 0;
}
/** Pushes back the last {@code count} characters in current token's buffer. */
private void pushBack(int count) {
if (pendingLimit == 0) {
if (bufferLimit != -1 && bufferNextRead >= count) {
// optimize common case when the chars we are pushing back are still in the buffer
bufferNextRead -= count;
} else {
if (count > pendingChars.length) {
pendingChars = ArrayUtil.grow(pendingChars, count);
}
System.arraycopy(termAtt.buffer(), tokenUpto - count, pendingChars, 0, count);
pendingLimit = count;
}
} else {
// we are pushing back what is already in our pending buffer
pendingUpto -= count;
assert pendingUpto >= 0;
}
offset -= count;
}
private void appendToToken(char ch) {
char[] buffer = termAtt.buffer();
if (tokenUpto == buffer.length) {
buffer = termAtt.resizeBuffer(tokenUpto + 1);
}
buffer[tokenUpto++] = ch;
}
private int nextCodeUnit() throws IOException {
int result;
if (pendingUpto < pendingLimit) {
result = pendingChars[pendingUpto++];
if (pendingUpto == pendingLimit) {
// We used up the pending buffer
pendingUpto = 0;
pendingLimit = 0;
}
appendToToken((char) result);
offset++;
} else if (bufferLimit == -1) {
return -1;
} else {
assert bufferNextRead <= bufferLimit: "bufferNextRead=" + bufferNextRead + " bufferLimit=" + bufferLimit;
if (bufferNextRead == bufferLimit) {
bufferLimit = input.read(buffer, 0, buffer.length);
if (bufferLimit == -1) {
return -1;
}
bufferNextRead = 0;
}
result = buffer[bufferNextRead++];
offset++;
appendToToken((char) result);
}
return result;
}
private int nextCodePoint() throws IOException {
int ch = nextCodeUnit();
if (ch == -1) {
return ch;
}
if (Character.isHighSurrogate((char) ch)) {
return Character.toCodePoint((char) ch, (char) nextCodeUnit());
} else {
return ch;
}
}
}