public void testDefaultsKuromojiAnalysis() throws IOException {
AnalysisService analysisService = createAnalysisService();
TokenizerFactory tokenizerFactory = analysisService.tokenizer("kuromoji_tokenizer");
assertThat(tokenizerFactory, instanceOf(KuromojiTokenizerFactory.class));
TokenFilterFactory filterFactory = analysisService.tokenFilter("kuromoji_part_of_speech");
assertThat(filterFactory, instanceOf(KuromojiPartOfSpeechFilterFactory.class));
filterFactory = analysisService.tokenFilter("kuromoji_readingform");
assertThat(filterFactory, instanceOf(KuromojiReadingFormFilterFactory.class));
filterFactory = analysisService.tokenFilter("kuromoji_baseform");
assertThat(filterFactory, instanceOf(KuromojiBaseFormFilterFactory.class));
filterFactory = analysisService.tokenFilter("kuromoji_stemmer");
assertThat(filterFactory, instanceOf(KuromojiKatakanaStemmerFactory.class));
filterFactory = analysisService.tokenFilter("ja_stop");
assertThat(filterFactory, instanceOf(JapaneseStopTokenFilterFactory.class));
NamedAnalyzer analyzer = analysisService.analyzer("kuromoji");
assertThat(analyzer.analyzer(), instanceOf(JapaneseAnalyzer.class));
analyzer = analysisService.analyzer("my_analyzer");
assertThat(analyzer.analyzer(), instanceOf(CustomAnalyzer.class));
assertThat(
analyzer.analyzer().tokenStream(null, new StringReader("")),
instanceOf(JapaneseTokenizer.class));
CharFilterFactory charFilterFactory = analysisService.charFilter("kuromoji_iteration_mark");
assertThat(charFilterFactory, instanceOf(KuromojiIterationMarkCharFilterFactory.class));
}
@Test
public void testArirangAnalysis() {
Index index = new Index("test");
Settings settings =
settingsBuilder()
.put("path.home", "/Users/hwjeong/temp/tmp")
.put(IndexMetaData.SETTING_VERSION_CREATED, Version.CURRENT)
.build();
Injector parentInjector =
new ModulesBuilder()
.add(
new SettingsModule(EMPTY_SETTINGS),
new EnvironmentModule(new Environment(settings)))
.createInjector();
Injector injector =
new ModulesBuilder()
.add(
new IndexSettingsModule(index, settings),
new IndexNameModule(index),
new AnalysisModule(
EMPTY_SETTINGS, parentInjector.getInstance(IndicesAnalysisService.class))
.addProcessor(new ArirangAnalysisBinderProcessor()))
.createChildInjector(parentInjector);
AnalysisService analysisService = injector.getInstance(AnalysisService.class);
TokenizerFactory tokenizerFactory = analysisService.tokenizer("arirang_tokenizer");
MatcherAssert.assertThat(tokenizerFactory, instanceOf(ArirangTokenizerFactory.class));
}
public void testKuromojiUserDict() throws IOException {
AnalysisService analysisService = createAnalysisService();
TokenizerFactory tokenizerFactory = analysisService.tokenizer("kuromoji_user_dict");
String source = "私は制限スピードを超える。";
String[] expected = new String[] {"私", "は", "制限スピード", "を", "超える"};
Tokenizer tokenizer = tokenizerFactory.create();
tokenizer.setReader(new StringReader(source));
assertSimpleTSOutput(tokenizer, expected);
}
// fix #59
public void testKuromojiEmptyUserDict() throws IOException {
AnalysisService analysisService = createAnalysisService();
TokenizerFactory tokenizerFactory = analysisService.tokenizer("kuromoji_empty_user_dict");
assertThat(tokenizerFactory, instanceOf(KuromojiTokenizerFactory.class));
}