1
package org.apache.lucene.analysis.icu.segmentation;
4
* Licensed to the Apache Software Foundation (ASF) under one or more
5
* contributor license agreements. See the NOTICE file distributed with
6
* this work for additional information regarding copyright ownership.
7
* The ASF licenses this file to You under the Apache License, Version 2.0
8
* (the "License"); you may not use this file except in compliance with
9
* the License. You may obtain a copy of the License at
11
* http://www.apache.org/licenses/LICENSE-2.0
13
* Unless required by applicable law or agreed to in writing, software
14
* distributed under the License is distributed on an "AS IS" BASIS,
15
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
16
* See the License for the specific language governing permissions and
17
* limitations under the License.
20
import java.io.IOException;
21
import java.io.InputStream;
23
import org.apache.lucene.analysis.standard.StandardTokenizer;
25
import com.ibm.icu.lang.UScript;
26
import com.ibm.icu.text.BreakIterator;
27
import com.ibm.icu.text.RuleBasedBreakIterator;
28
import com.ibm.icu.util.ULocale;
31
* Default {@link ICUTokenizerConfig} that is generally applicable
34
* Generally tokenizes Unicode text according to UAX#29
35
* ({@link BreakIterator#getWordInstance(ULocale) BreakIterator.getWordInstance(ULocale.ROOT)}),
36
* but with the following tailorings:
38
* <li>Thai text is broken into words with a
39
* {@link com.ibm.icu.text.DictionaryBasedBreakIterator}
40
* <li>Lao, Myanmar, and Khmer text is broken into syllables
41
* based on custom BreakIterator rules.
42
* <li>Hebrew text has custom tailorings to handle special cases
43
* involving punctuation.
45
* @lucene.experimental
47
public class DefaultICUTokenizerConfig extends ICUTokenizerConfig {
48
/** Token type for words containing ideographic characters */
49
public static final String WORD_IDEO = StandardTokenizer.TOKEN_TYPES[StandardTokenizer.IDEOGRAPHIC];
50
/** Token type for words containing Japanese hiragana */
51
public static final String WORD_HIRAGANA = StandardTokenizer.TOKEN_TYPES[StandardTokenizer.HIRAGANA];
52
/** Token type for words containing Japanese katakana */
53
public static final String WORD_KATAKANA = StandardTokenizer.TOKEN_TYPES[StandardTokenizer.KATAKANA];
54
/** Token type for words containing Korean hangul */
55
public static final String WORD_HANGUL = StandardTokenizer.TOKEN_TYPES[StandardTokenizer.HANGUL];
56
/** Token type for words that contain letters */
57
public static final String WORD_LETTER = StandardTokenizer.TOKEN_TYPES[StandardTokenizer.ALPHANUM];
58
/** Token type for words that appear to be numbers */
59
public static final String WORD_NUMBER = StandardTokenizer.TOKEN_TYPES[StandardTokenizer.NUM];
62
* the default breakiterators in use. these can be expensive to
63
* instantiate, cheap to clone.
65
private static final BreakIterator rootBreakIterator =
66
readBreakIterator("Default.brk");
67
private static final BreakIterator thaiBreakIterator =
68
BreakIterator.getWordInstance(new ULocale("th_TH"));
69
private static final BreakIterator hebrewBreakIterator =
70
readBreakIterator("Hebrew.brk");
71
private static final BreakIterator khmerBreakIterator =
72
readBreakIterator("Khmer.brk");
73
private static final BreakIterator laoBreakIterator =
74
new LaoBreakIterator(readBreakIterator("Lao.brk"));
75
private static final BreakIterator myanmarBreakIterator =
76
readBreakIterator("Myanmar.brk");
79
public BreakIterator getBreakIterator(int script) {
81
case UScript.THAI: return (BreakIterator)thaiBreakIterator.clone();
82
case UScript.HEBREW: return (BreakIterator)hebrewBreakIterator.clone();
83
case UScript.KHMER: return (BreakIterator)khmerBreakIterator.clone();
84
case UScript.LAO: return (BreakIterator)laoBreakIterator.clone();
85
case UScript.MYANMAR: return (BreakIterator)myanmarBreakIterator.clone();
86
default: return (BreakIterator)rootBreakIterator.clone();
91
public String getType(int script, int ruleStatus) {
93
case RuleBasedBreakIterator.WORD_IDEO:
95
case RuleBasedBreakIterator.WORD_KANA:
96
return script == UScript.HIRAGANA ? WORD_HIRAGANA : WORD_KATAKANA;
97
case RuleBasedBreakIterator.WORD_LETTER:
98
return script == UScript.HANGUL ? WORD_HANGUL : WORD_LETTER;
99
case RuleBasedBreakIterator.WORD_NUMBER:
101
default: /* some other custom code */
106
private static RuleBasedBreakIterator readBreakIterator(String filename) {
108
DefaultICUTokenizerConfig.class.getResourceAsStream(filename);
110
RuleBasedBreakIterator bi =
111
RuleBasedBreakIterator.getInstanceFromCompiledRules(is);
114
} catch (IOException e) {
115
throw new RuntimeException(e);