Examples with AnalysisPlugin org.opensearch.plugins.AnalysisPlugin used on opensource projects

Example 1 with AnalysisPlugin

use of org.opensearch.plugins.AnalysisPlugin in project OpenSearch by opensearch-project.

the class AnalysisModuleTests method testRegisterHunspellDictionary.

public void testRegisterHunspellDictionary() throws Exception {
    Settings settings = Settings.builder().put(Environment.PATH_HOME_SETTING.getKey(), createTempDir().toString()).put(IndexMetadata.SETTING_VERSION_CREATED, Version.CURRENT).build();
    Environment environment = TestEnvironment.newEnvironment(settings);
    InputStream aff = getClass().getResourceAsStream("/indices/analyze/conf_dir/hunspell/en_US/en_US.aff");
    InputStream dic = getClass().getResourceAsStream("/indices/analyze/conf_dir/hunspell/en_US/en_US.dic");
    Dictionary dictionary;
    try (Directory tmp = new NIOFSDirectory(environment.tmpFile())) {
        dictionary = new Dictionary(tmp, "hunspell", aff, dic);
    }
    AnalysisModule module = new AnalysisModule(environment, singletonList(new AnalysisPlugin() {

        @Override
        public Map<String, Dictionary> getHunspellDictionaries() {
            return singletonMap("foo", dictionary);
        }
    }));
    assertSame(dictionary, module.getHunspellService().getDictionary("foo"));
}

Example 2 with AnalysisPlugin

use of org.opensearch.plugins.AnalysisPlugin in project OpenSearch by opensearch-project.

the class AnalysisModuleTests method testPluginPreConfiguredCharFilters.

/**
 * Tests that plugins can register pre-configured char filters that vary in behavior based on OpenSearch version, Lucene version,
 * and that do not vary based on version at all.
 */
public void testPluginPreConfiguredCharFilters() throws IOException {
    boolean noVersionSupportsMultiTerm = randomBoolean();
    boolean luceneVersionSupportsMultiTerm = randomBoolean();
    boolean opensearchVersionSupportsMultiTerm = randomBoolean();
    AnalysisRegistry registry = new AnalysisModule(TestEnvironment.newEnvironment(emptyNodeSettings), singletonList(new AnalysisPlugin() {

        @Override
        public List<PreConfiguredCharFilter> getPreConfiguredCharFilters() {
            return Arrays.asList(PreConfiguredCharFilter.singleton("no_version", noVersionSupportsMultiTerm, tokenStream -> new AppendCharFilter(tokenStream, "no_version")), PreConfiguredCharFilter.luceneVersion("lucene_version", luceneVersionSupportsMultiTerm, (tokenStream, luceneVersion) -> new AppendCharFilter(tokenStream, luceneVersion.toString())), PreConfiguredCharFilter.openSearchVersion("opensearch_version", opensearchVersionSupportsMultiTerm, (tokenStream, esVersion) -> new AppendCharFilter(tokenStream, esVersion.toString())));
        }

        @Override
        public Map<String, AnalysisProvider<TokenizerFactory>> getTokenizers() {
            // Need mock keyword tokenizer here, because alpha / beta versions are broken up by the dash.
            return singletonMap("keyword", (indexSettings, environment, name, settings) -> TokenizerFactory.newFactory(name, () -> new MockTokenizer(MockTokenizer.KEYWORD, false)));
        }
    })).getAnalysisRegistry();
    Version version = VersionUtils.randomVersion(random());
    IndexAnalyzers analyzers = getIndexAnalyzers(registry, Settings.builder().put("index.analysis.analyzer.no_version.tokenizer", "keyword").put("index.analysis.analyzer.no_version.char_filter", "no_version").put("index.analysis.analyzer.lucene_version.tokenizer", "keyword").put("index.analysis.analyzer.lucene_version.char_filter", "lucene_version").put("index.analysis.analyzer.opensearch_version.tokenizer", "keyword").put("index.analysis.analyzer.opensearch_version.char_filter", "opensearch_version").put(IndexMetadata.SETTING_VERSION_CREATED, version).build());
    assertTokenStreamContents(analyzers.get("no_version").tokenStream("", "test"), new String[] { "testno_version" });
    assertTokenStreamContents(analyzers.get("lucene_version").tokenStream("", "test"), new String[] { "test" + version.luceneVersion });
    assertTokenStreamContents(analyzers.get("opensearch_version").tokenStream("", "test"), new String[] { "test" + version });
    assertEquals("test" + (noVersionSupportsMultiTerm ? "no_version" : ""), analyzers.get("no_version").normalize("", "test").utf8ToString());
    assertEquals("test" + (luceneVersionSupportsMultiTerm ? version.luceneVersion.toString() : ""), analyzers.get("lucene_version").normalize("", "test").utf8ToString());
    assertEquals("test" + (opensearchVersionSupportsMultiTerm ? version.toString() : ""), analyzers.get("opensearch_version").normalize("", "test").utf8ToString());
}

Example 3 with AnalysisPlugin

use of org.opensearch.plugins.AnalysisPlugin in project OpenSearch by opensearch-project.

the class AnalysisModuleTests method testPluginPreConfiguredTokenFilters.

/**
 * Tests that plugins can register pre-configured token filters that vary in behavior based on OpenSearch version, Lucene version,
 * and that do not vary based on version at all.
 */
public void testPluginPreConfiguredTokenFilters() throws IOException {
    boolean noVersionSupportsMultiTerm = randomBoolean();
    boolean luceneVersionSupportsMultiTerm = randomBoolean();
    boolean opensearchVersionSupportsMultiTerm = randomBoolean();
    AnalysisRegistry registry = new AnalysisModule(TestEnvironment.newEnvironment(emptyNodeSettings), singletonList(new AnalysisPlugin() {

        @Override
        public List<PreConfiguredTokenFilter> getPreConfiguredTokenFilters() {
            return Arrays.asList(PreConfiguredTokenFilter.singleton("no_version", noVersionSupportsMultiTerm, tokenStream -> new AppendTokenFilter(tokenStream, "no_version")), PreConfiguredTokenFilter.luceneVersion("lucene_version", luceneVersionSupportsMultiTerm, (tokenStream, luceneVersion) -> new AppendTokenFilter(tokenStream, luceneVersion.toString())), PreConfiguredTokenFilter.openSearchVersion("opensearch_version", opensearchVersionSupportsMultiTerm, (tokenStream, esVersion) -> new AppendTokenFilter(tokenStream, esVersion.toString())));
        }
    })).getAnalysisRegistry();
    Version version = VersionUtils.randomVersion(random());
    IndexAnalyzers analyzers = getIndexAnalyzers(registry, Settings.builder().put("index.analysis.analyzer.no_version.tokenizer", "standard").put("index.analysis.analyzer.no_version.filter", "no_version").put("index.analysis.analyzer.lucene_version.tokenizer", "standard").put("index.analysis.analyzer.lucene_version.filter", "lucene_version").put("index.analysis.analyzer.opensearch_version.tokenizer", "standard").put("index.analysis.analyzer.opensearch_version.filter", "opensearch_version").put(IndexMetadata.SETTING_VERSION_CREATED, version).build());
    assertTokenStreamContents(analyzers.get("no_version").tokenStream("", "test"), new String[] { "testno_version" });
    assertTokenStreamContents(analyzers.get("lucene_version").tokenStream("", "test"), new String[] { "test" + version.luceneVersion });
    assertTokenStreamContents(analyzers.get("opensearch_version").tokenStream("", "test"), new String[] { "test" + version });
    assertEquals("test" + (noVersionSupportsMultiTerm ? "no_version" : ""), analyzers.get("no_version").normalize("", "test").utf8ToString());
    assertEquals("test" + (luceneVersionSupportsMultiTerm ? version.luceneVersion.toString() : ""), analyzers.get("lucene_version").normalize("", "test").utf8ToString());
    assertEquals("test" + (opensearchVersionSupportsMultiTerm ? version.toString() : ""), analyzers.get("opensearch_version").normalize("", "test").utf8ToString());
}

Example 4 with AnalysisPlugin

use of org.opensearch.plugins.AnalysisPlugin in project OpenSearch by opensearch-project.

the class AnalysisModule method setupPreConfiguredTokenFilters.

static Map<String, PreConfiguredTokenFilter> setupPreConfiguredTokenFilters(List<AnalysisPlugin> plugins) {
    NamedRegistry<PreConfiguredTokenFilter> preConfiguredTokenFilters = new NamedRegistry<>("pre-configured token_filter");
    // Add filters available in lucene-core
    preConfiguredTokenFilters.register("lowercase", PreConfiguredTokenFilter.singleton("lowercase", true, LowerCaseFilter::new));
    // Add "standard" for old indices (bwc)
    preConfiguredTokenFilters.register("standard", PreConfiguredTokenFilter.openSearchVersion("standard", true, (reader, version) -> {
        // until version 7_5_2
        if (version.before(LegacyESVersion.V_7_6_0)) {
            deprecationLogger.deprecate("standard_deprecation", "The [standard] token filter is deprecated and will be removed in a future version.");
        } else {
            throw new IllegalArgumentException("The [standard] token filter has been removed.");
        }
        return reader;
    }));
    for (AnalysisPlugin plugin : plugins) {
        for (PreConfiguredTokenFilter filter : plugin.getPreConfiguredTokenFilters()) {
            preConfiguredTokenFilters.register(filter.getName(), filter);
        }
    }
    return unmodifiableMap(preConfiguredTokenFilters.getRegistry());
}

Example 5 with AnalysisPlugin

use of org.opensearch.plugins.AnalysisPlugin in project OpenSearch by opensearch-project.

the class AnalysisModule method setupPreConfiguredTokenizers.

static Map<String, PreConfiguredTokenizer> setupPreConfiguredTokenizers(List<AnalysisPlugin> plugins) {
    NamedRegistry<PreConfiguredTokenizer> preConfiguredTokenizers = new NamedRegistry<>("pre-configured tokenizer");
    // Temporary shim to register old style pre-configured tokenizers
    for (PreBuiltTokenizers tokenizer : PreBuiltTokenizers.values()) {
        String name = tokenizer.name().toLowerCase(Locale.ROOT);
        PreConfiguredTokenizer preConfigured;
        switch(tokenizer.getCachingStrategy()) {
            case ONE:
                preConfigured = PreConfiguredTokenizer.singleton(name, () -> tokenizer.create(Version.CURRENT));
                break;
            default:
                throw new UnsupportedOperationException("Caching strategy unsupported by temporary shim [" + tokenizer + "]");
        }
        preConfiguredTokenizers.register(name, preConfigured);
    }
    for (AnalysisPlugin plugin : plugins) {
        for (PreConfiguredTokenizer tokenizer : plugin.getPreConfiguredTokenizers()) {
            preConfiguredTokenizers.register(tokenizer.getName(), tokenizer);
        }
    }
    return unmodifiableMap(preConfiguredTokenizers.getRegistry());
}