Examples with StandardTokenizer org.apache.lucene.analysis.standard.StandardTokenizer used on opensource projects

Example 41 with StandardTokenizer

use of org.apache.lucene.analysis.standard.StandardTokenizer in project omegat by omegat-org.

the class HunspellTokenizer method getTokenStream.

@Override
protected TokenStream getTokenStream(final String strOrig, final boolean stemsAllowed, final boolean stopWordsAllowed) throws IOException {
    StandardTokenizer tokenizer = new StandardTokenizer();
    tokenizer.setReader(new StringReader(strOrig));
    if (stemsAllowed) {
        Dictionary dictionary = getDict();
        if (dictionary == null) {
            return tokenizer;
        }
        return new HunspellStemFilter(tokenizer, dictionary);
    // / TODO: implement stop words checks
    } else {
        return tokenizer;
    }
}

Example 42 with StandardTokenizer

use of org.apache.lucene.analysis.standard.StandardTokenizer in project elasticsearch by elastic.

the class KeepTypesFilterFactoryTests method testKeepTypes.

public void testKeepTypes() throws IOException {
    Settings settings = Settings.builder().put(Environment.PATH_HOME_SETTING.getKey(), createTempDir().toString()).put("index.analysis.filter.keep_numbers.type", "keep_types").putArray("index.analysis.filter.keep_numbers.types", new String[] { "<NUM>", "<SOMETHINGELSE>" }).build();
    ESTestCase.TestAnalysis analysis = AnalysisTestsHelper.createTestAnalysisFromSettings(settings);
    TokenFilterFactory tokenFilter = analysis.tokenFilter.get("keep_numbers");
    assertThat(tokenFilter, instanceOf(KeepTypesFilterFactory.class));
    String source = "Hello 123 world";
    String[] expected = new String[] { "123" };
    Tokenizer tokenizer = new StandardTokenizer();
    tokenizer.setReader(new StringReader(source));
    assertTokenStreamContents(tokenFilter.create(tokenizer), expected, new int[] { 2 });
}

Example 43 with StandardTokenizer

use of org.apache.lucene.analysis.standard.StandardTokenizer in project elasticsearch by elastic.

the class CJKFilterFactoryTests method testNoFlags.

public void testNoFlags() throws IOException {
    ESTestCase.TestAnalysis analysis = AnalysisTestsHelper.createTestAnalysisFromClassPath(createTempDir(), RESOURCE);
    TokenFilterFactory tokenFilter = analysis.tokenFilter.get("cjk_no_flags");
    String source = "多くの学生が試験に落ちた。";
    String[] expected = new String[] { "多く", "くの", "の学", "学生", "生が", "が試", "試験", "験に", "に落", "落ち", "ちた" };
    Tokenizer tokenizer = new StandardTokenizer();
    tokenizer.setReader(new StringReader(source));
    assertTokenStreamContents(tokenFilter.create(tokenizer), expected);
}

Example 44 with StandardTokenizer

use of org.apache.lucene.analysis.standard.StandardTokenizer in project elasticsearch by elastic.

the class NoisyChannelSpellCheckerTests method testNgram.

public void testNgram() throws IOException {
    RAMDirectory dir = new RAMDirectory();
    Map<String, Analyzer> mapping = new HashMap<>();
    mapping.put("body_ngram", new Analyzer() {

        @Override
        protected TokenStreamComponents createComponents(String fieldName) {
            Tokenizer t = new StandardTokenizer();
            ShingleFilter tf = new ShingleFilter(t, 2, 3);
            tf.setOutputUnigrams(false);
            return new TokenStreamComponents(t, new LowerCaseFilter(tf));
        }
    });
    mapping.put("body", new Analyzer() {

        @Override
        protected TokenStreamComponents createComponents(String fieldName) {
            Tokenizer t = new StandardTokenizer();
            return new TokenStreamComponents(t, new LowerCaseFilter(t));
        }
    });
    PerFieldAnalyzerWrapper wrapper = new PerFieldAnalyzerWrapper(new WhitespaceAnalyzer(), mapping);
    IndexWriterConfig conf = new IndexWriterConfig(wrapper);
    IndexWriter writer = new IndexWriter(dir, conf);
    String[] strings = new String[] { "Xorr the God-Jewel", "Grog the God-Crusher", "Xorn", "Walter Newell", "Wanda Maximoff", "Captain America", "American Ace", "USA Hero", "Wundarr the Aquarian", "Will o' the Wisp", "Xemnu the Titan", "Fantastic Four", "Quasar", "Quasar II" };
    for (String line : strings) {
        Document doc = new Document();
        doc.add(new Field("body", line, TextField.TYPE_NOT_STORED));
        doc.add(new Field("body_ngram", line, TextField.TYPE_NOT_STORED));
        writer.addDocument(doc);
    }
    DirectoryReader ir = DirectoryReader.open(writer);
    WordScorer wordScorer = new LaplaceScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.95d, new BytesRef(" "), 0.5f);
    NoisyChannelSpellChecker suggester = new NoisyChannelSpellChecker();
    DirectSpellChecker spellchecker = new DirectSpellChecker();
    spellchecker.setMinQueryLength(1);
    DirectCandidateGenerator generator = new DirectCandidateGenerator(spellchecker, "body", SuggestMode.SUGGEST_MORE_POPULAR, ir, 0.95, 5);
    Result result = suggester.getCorrections(wrapper, new BytesRef("american ame"), generator, 1, 1, ir, "body", wordScorer, 1, 2);
    Correction[] corrections = result.corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(space).utf8ToString(), equalTo("american ace"));
    assertThat(corrections[0].join(space, preTag, postTag).utf8ToString(), equalTo("american <em>ace</em>"));
    assertThat(result.cutoffScore, greaterThan(0d));
    result = suggester.getCorrections(wrapper, new BytesRef("american ame"), generator, 1, 1, ir, "body", wordScorer, 0, 1);
    corrections = result.corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(space).utf8ToString(), equalTo("american ame"));
    assertThat(corrections[0].join(space, preTag, postTag).utf8ToString(), equalTo("american ame"));
    assertThat(result.cutoffScore, equalTo(Double.MIN_VALUE));
    suggester = new NoisyChannelSpellChecker(0.85);
    wordScorer = new LaplaceScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.85d, new BytesRef(" "), 0.5f);
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor the Got-Jewel"), generator, 0.5f, 4, ir, "body", wordScorer, 0, 2).corrections;
    assertThat(corrections.length, equalTo(4));
    assertThat(corrections[0].join(space).utf8ToString(), equalTo("xorr the god jewel"));
    assertThat(corrections[1].join(space).utf8ToString(), equalTo("xor the god jewel"));
    assertThat(corrections[2].join(space).utf8ToString(), equalTo("xorn the god jewel"));
    assertThat(corrections[3].join(space).utf8ToString(), equalTo("xorr the got jewel"));
    assertThat(corrections[0].join(space, preTag, postTag).utf8ToString(), equalTo("<em>xorr</em> the <em>god</em> jewel"));
    assertThat(corrections[1].join(space, preTag, postTag).utf8ToString(), equalTo("xor the <em>god</em> jewel"));
    assertThat(corrections[2].join(space, preTag, postTag).utf8ToString(), equalTo("<em>xorn</em> the <em>god</em> jewel"));
    assertThat(corrections[3].join(space, preTag, postTag).utf8ToString(), equalTo("<em>xorr</em> the got jewel"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor the Got-Jewel"), generator, 0.5f, 4, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections.length, equalTo(4));
    assertThat(corrections[0].join(space).utf8ToString(), equalTo("xorr the god jewel"));
    assertThat(corrections[1].join(space).utf8ToString(), equalTo("xor the god jewel"));
    assertThat(corrections[2].join(space).utf8ToString(), equalTo("xorn the god jewel"));
    assertThat(corrections[3].join(space).utf8ToString(), equalTo("xorr the got jewel"));
    // Test some of the highlighting corner cases
    suggester = new NoisyChannelSpellChecker(0.85);
    wordScorer = new LaplaceScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.85d, new BytesRef(" "), 0.5f);
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor teh Got-Jewel"), generator, 4f, 4, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections.length, equalTo(4));
    assertThat(corrections[0].join(space).utf8ToString(), equalTo("xorr the god jewel"));
    assertThat(corrections[1].join(space).utf8ToString(), equalTo("xor the god jewel"));
    assertThat(corrections[2].join(space).utf8ToString(), equalTo("xorn the god jewel"));
    assertThat(corrections[3].join(space).utf8ToString(), equalTo("xor teh god jewel"));
    assertThat(corrections[0].join(space, preTag, postTag).utf8ToString(), equalTo("<em>xorr the god</em> jewel"));
    assertThat(corrections[1].join(space, preTag, postTag).utf8ToString(), equalTo("xor <em>the god</em> jewel"));
    assertThat(corrections[2].join(space, preTag, postTag).utf8ToString(), equalTo("<em>xorn the god</em> jewel"));
    assertThat(corrections[3].join(space, preTag, postTag).utf8ToString(), equalTo("xor teh <em>god</em> jewel"));
    // test synonyms
    Analyzer analyzer = new Analyzer() {

        @Override
        protected TokenStreamComponents createComponents(String fieldName) {
            Tokenizer t = new StandardTokenizer();
            TokenFilter filter = new LowerCaseFilter(t);
            try {
                SolrSynonymParser parser = new SolrSynonymParser(true, false, new WhitespaceAnalyzer());
                parser.parse(new StringReader("usa => usa, america, american"));
                filter = new SynonymFilter(filter, parser.build(), true);
            } catch (Exception e) {
                throw new RuntimeException(e);
            }
            return new TokenStreamComponents(t, filter);
        }
    };
    spellchecker.setAccuracy(0.0f);
    spellchecker.setMinPrefix(1);
    spellchecker.setMinQueryLength(1);
    suggester = new NoisyChannelSpellChecker(0.85);
    wordScorer = new LaplaceScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.85d, new BytesRef(" "), 0.5f);
    corrections = suggester.getCorrections(analyzer, new BytesRef("captian usa"), generator, 2, 4, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections[0].join(space).utf8ToString(), equalTo("captain america"));
    assertThat(corrections[0].join(space, preTag, postTag).utf8ToString(), equalTo("<em>captain america</em>"));
    generator = new DirectCandidateGenerator(spellchecker, "body", SuggestMode.SUGGEST_MORE_POPULAR, ir, 0.85, 10, null, analyzer, MultiFields.getTerms(ir, "body"));
    corrections = suggester.getCorrections(analyzer, new BytesRef("captian usw"), generator, 2, 4, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("captain america"));
    assertThat(corrections[0].join(space, preTag, postTag).utf8ToString(), equalTo("<em>captain america</em>"));
    // Make sure that user supplied text is not marked as highlighted in the presence of a synonym filter
    generator = new DirectCandidateGenerator(spellchecker, "body", SuggestMode.SUGGEST_MORE_POPULAR, ir, 0.85, 10, null, analyzer, MultiFields.getTerms(ir, "body"));
    corrections = suggester.getCorrections(analyzer, new BytesRef("captain usw"), generator, 2, 4, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("captain america"));
    assertThat(corrections[0].join(space, preTag, postTag).utf8ToString(), equalTo("captain <em>america</em>"));
}

Example 45 with StandardTokenizer

use of org.apache.lucene.analysis.standard.StandardTokenizer in project elasticsearch by elastic.

the class NoisyChannelSpellCheckerTests method testTrigram.

public void testTrigram() throws IOException {
    RAMDirectory dir = new RAMDirectory();
    Map<String, Analyzer> mapping = new HashMap<>();
    mapping.put("body_ngram", new Analyzer() {

        @Override
        protected TokenStreamComponents createComponents(String fieldName) {
            Tokenizer t = new StandardTokenizer();
            ShingleFilter tf = new ShingleFilter(t, 2, 3);
            tf.setOutputUnigrams(false);
            return new TokenStreamComponents(t, new LowerCaseFilter(tf));
        }
    });
    mapping.put("body", new Analyzer() {

        @Override
        protected TokenStreamComponents createComponents(String fieldName) {
            Tokenizer t = new StandardTokenizer();
            return new TokenStreamComponents(t, new LowerCaseFilter(t));
        }
    });
    PerFieldAnalyzerWrapper wrapper = new PerFieldAnalyzerWrapper(new WhitespaceAnalyzer(), mapping);
    IndexWriterConfig conf = new IndexWriterConfig(wrapper);
    IndexWriter writer = new IndexWriter(dir, conf);
    String[] strings = new String[] { "Xorr the God-Jewel", "Grog the God-Crusher", "Xorn", "Walter Newell", "Wanda Maximoff", "Captain America", "American Ace", "USA Hero", "Wundarr the Aquarian", "Will o' the Wisp", "Xemnu the Titan", "Fantastic Four", "Quasar", "Quasar II" };
    for (String line : strings) {
        Document doc = new Document();
        doc.add(new Field("body", line, TextField.TYPE_NOT_STORED));
        doc.add(new Field("body_ngram", line, TextField.TYPE_NOT_STORED));
        writer.addDocument(doc);
    }
    DirectoryReader ir = DirectoryReader.open(writer);
    WordScorer wordScorer = new LinearInterpolatingScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.85d, new BytesRef(" "), 0.5, 0.4, 0.1);
    NoisyChannelSpellChecker suggester = new NoisyChannelSpellChecker();
    DirectSpellChecker spellchecker = new DirectSpellChecker();
    spellchecker.setMinQueryLength(1);
    DirectCandidateGenerator generator = new DirectCandidateGenerator(spellchecker, "body", SuggestMode.SUGGEST_MORE_POPULAR, ir, 0.95, 5);
    Correction[] corrections = suggester.getCorrections(wrapper, new BytesRef("american ame"), generator, 1, 1, ir, "body", wordScorer, 1, 3).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("american ace"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("american ame"), generator, 1, 1, ir, "body", wordScorer, 1, 1).corrections;
    assertThat(corrections.length, equalTo(0));
    //        assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("american ape"));
    wordScorer = new LinearInterpolatingScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.85d, new BytesRef(" "), 0.5, 0.4, 0.1);
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor the Got-Jewel"), generator, 0.5f, 4, ir, "body", wordScorer, 0, 3).corrections;
    assertThat(corrections.length, equalTo(4));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the god jewel"));
    assertThat(corrections[1].join(new BytesRef(" ")).utf8ToString(), equalTo("xor the god jewel"));
    assertThat(corrections[2].join(new BytesRef(" ")).utf8ToString(), equalTo("xorn the god jewel"));
    assertThat(corrections[3].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the got jewel"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor the Got-Jewel"), generator, 0.5f, 4, ir, "body", wordScorer, 1, 3).corrections;
    assertThat(corrections.length, equalTo(4));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the god jewel"));
    assertThat(corrections[1].join(new BytesRef(" ")).utf8ToString(), equalTo("xor the god jewel"));
    assertThat(corrections[2].join(new BytesRef(" ")).utf8ToString(), equalTo("xorn the god jewel"));
    assertThat(corrections[3].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the got jewel"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor the Got-Jewel"), generator, 0.5f, 1, ir, "body", wordScorer, 100, 3).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the god jewel"));
    // test synonyms
    Analyzer analyzer = new Analyzer() {

        @Override
        protected TokenStreamComponents createComponents(String fieldName) {
            Tokenizer t = new StandardTokenizer();
            TokenFilter filter = new LowerCaseFilter(t);
            try {
                SolrSynonymParser parser = new SolrSynonymParser(true, false, new WhitespaceAnalyzer());
                parser.parse(new StringReader("usa => usa, america, american"));
                filter = new SynonymFilter(filter, parser.build(), true);
            } catch (Exception e) {
                throw new RuntimeException(e);
            }
            return new TokenStreamComponents(t, filter);
        }
    };
    spellchecker.setAccuracy(0.0f);
    spellchecker.setMinPrefix(1);
    spellchecker.setMinQueryLength(1);
    suggester = new NoisyChannelSpellChecker(0.95);
    wordScorer = new LinearInterpolatingScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.95d, new BytesRef(" "), 0.5, 0.4, 0.1);
    corrections = suggester.getCorrections(analyzer, new BytesRef("captian usa"), generator, 2, 4, ir, "body", wordScorer, 1, 3).corrections;
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("captain america"));
    generator = new DirectCandidateGenerator(spellchecker, "body", SuggestMode.SUGGEST_MORE_POPULAR, ir, 0.95, 10, null, analyzer, MultiFields.getTerms(ir, "body"));
    corrections = suggester.getCorrections(analyzer, new BytesRef("captian usw"), generator, 2, 4, ir, "body", wordScorer, 1, 3).corrections;
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("captain america"));
    wordScorer = new StupidBackoffScorer(ir, MultiFields.getTerms(ir, "body_ngram"), "body_ngram", 0.85d, new BytesRef(" "), 0.4);
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor the Got-Jewel"), generator, 0.5f, 2, ir, "body", wordScorer, 0, 3).corrections;
    assertThat(corrections.length, equalTo(2));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the god jewel"));
    assertThat(corrections[1].join(new BytesRef(" ")).utf8ToString(), equalTo("xor the god jewel"));
}