Search in sources :

Example 1 with Tokenizer

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

the class AbstractTermVectorsTestCase method indexDocsWithLucene.

protected DirectoryReader indexDocsWithLucene(TestDoc[] testDocs) throws IOException {
    Map<String, Analyzer> mapping = new HashMap<>();
    for (TestFieldSetting field : testDocs[0].fieldSettings) {
        if (field.storedPayloads) {
            mapping.put(field.name, new Analyzer() {

                @Override
                protected TokenStreamComponents createComponents(String fieldName) {
                    Tokenizer tokenizer = new StandardTokenizer();
                    TokenFilter filter = new LowerCaseFilter(tokenizer);
                    filter = new TypeAsPayloadTokenFilter(filter);
                    return new TokenStreamComponents(tokenizer, filter);
                }
            });
        }
    }
    PerFieldAnalyzerWrapper wrapper = new PerFieldAnalyzerWrapper(new StandardAnalyzer(CharArraySet.EMPTY_SET), mapping);
    Directory dir = new RAMDirectory();
    IndexWriterConfig conf = new IndexWriterConfig(wrapper);
    conf.setOpenMode(IndexWriterConfig.OpenMode.CREATE);
    IndexWriter writer = new IndexWriter(dir, conf);
    for (TestDoc doc : testDocs) {
        Document d = new Document();
        d.add(new Field("id", doc.id, StringField.TYPE_STORED));
        for (int i = 0; i < doc.fieldContent.length; i++) {
            FieldType type = new FieldType(TextField.TYPE_STORED);
            TestFieldSetting fieldSetting = doc.fieldSettings[i];
            type.setStoreTermVectorOffsets(fieldSetting.storedOffset);
            type.setStoreTermVectorPayloads(fieldSetting.storedPayloads);
            type.setStoreTermVectorPositions(fieldSetting.storedPositions || fieldSetting.storedPayloads || fieldSetting.storedOffset);
            type.setStoreTermVectors(true);
            type.freeze();
            d.add(new Field(fieldSetting.name, doc.fieldContent[i], type));
        }
        writer.updateDocument(new Term("id", doc.id), d);
        writer.commit();
    }
    writer.close();
    return DirectoryReader.open(dir);
}
Also used : HashMap(java.util.HashMap) TypeAsPayloadTokenFilter(org.apache.lucene.analysis.payloads.TypeAsPayloadTokenFilter) Term(org.apache.lucene.index.Term) Analyzer(org.apache.lucene.analysis.Analyzer) StandardAnalyzer(org.apache.lucene.analysis.standard.StandardAnalyzer) Document(org.apache.lucene.document.Document) RAMDirectory(org.apache.lucene.store.RAMDirectory) PerFieldAnalyzerWrapper(org.apache.lucene.analysis.miscellaneous.PerFieldAnalyzerWrapper) FieldType(org.apache.lucene.document.FieldType) StringField(org.apache.lucene.document.StringField) Field(org.apache.lucene.document.Field) TextField(org.apache.lucene.document.TextField) IndexWriter(org.apache.lucene.index.IndexWriter) StandardTokenizer(org.apache.lucene.analysis.standard.StandardTokenizer) StandardAnalyzer(org.apache.lucene.analysis.standard.StandardAnalyzer) Tokenizer(org.apache.lucene.analysis.Tokenizer) StandardTokenizer(org.apache.lucene.analysis.standard.StandardTokenizer) LowerCaseFilter(org.apache.lucene.analysis.LowerCaseFilter) TypeAsPayloadTokenFilter(org.apache.lucene.analysis.payloads.TypeAsPayloadTokenFilter) TokenFilter(org.apache.lucene.analysis.TokenFilter) RAMDirectory(org.apache.lucene.store.RAMDirectory) Directory(org.apache.lucene.store.Directory) IndexWriterConfig(org.apache.lucene.index.IndexWriterConfig)

Example 2 with Tokenizer

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

the class TransportAnalyzeAction method createStackedTokenStream.

private static TokenStream createStackedTokenStream(String source, CharFilterFactory[] charFilterFactories, TokenizerFactory tokenizerFactory, TokenFilterFactory[] tokenFilterFactories, int current) {
    Reader reader = new FastStringReader(source);
    for (CharFilterFactory charFilterFactory : charFilterFactories) {
        reader = charFilterFactory.create(reader);
    }
    Tokenizer tokenizer = tokenizerFactory.create();
    tokenizer.setReader(reader);
    TokenStream tokenStream = tokenizer;
    for (int i = 0; i < current; i++) {
        tokenStream = tokenFilterFactories[i].create(tokenStream);
    }
    return tokenStream;
}
Also used : TokenStream(org.apache.lucene.analysis.TokenStream) FastStringReader(org.elasticsearch.common.io.FastStringReader) CharFilterFactory(org.elasticsearch.index.analysis.CharFilterFactory) FastStringReader(org.elasticsearch.common.io.FastStringReader) Reader(java.io.Reader) Tokenizer(org.apache.lucene.analysis.Tokenizer)

Example 3 with Tokenizer

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

the class CustomAnalyzer method createComponents.

@Override
protected TokenStreamComponents createComponents(String fieldName) {
    Tokenizer tokenizer = tokenizerFactory.create();
    TokenStream tokenStream = tokenizer;
    for (TokenFilterFactory tokenFilter : tokenFilters) {
        tokenStream = tokenFilter.create(tokenStream);
    }
    return new TokenStreamComponents(tokenizer, tokenStream);
}
Also used : TokenStream(org.apache.lucene.analysis.TokenStream) Tokenizer(org.apache.lucene.analysis.Tokenizer)

Example 4 with Tokenizer

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

the class PatternAnalyzer method createComponents.

@Override
protected TokenStreamComponents createComponents(String s) {
    final Tokenizer tokenizer = new PatternTokenizer(pattern, -1);
    TokenStream stream = tokenizer;
    if (lowercase) {
        stream = new LowerCaseFilter(stream);
    }
    if (stopWords != null) {
        stream = new StopFilter(stream, stopWords);
    }
    return new TokenStreamComponents(tokenizer, stream);
}
Also used : TokenStream(org.apache.lucene.analysis.TokenStream) StopFilter(org.apache.lucene.analysis.StopFilter) PatternTokenizer(org.apache.lucene.analysis.pattern.PatternTokenizer) Tokenizer(org.apache.lucene.analysis.Tokenizer) PatternTokenizer(org.apache.lucene.analysis.pattern.PatternTokenizer) LowerCaseFilter(org.apache.lucene.analysis.LowerCaseFilter)

Example 5 with Tokenizer

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

the class NoisyChannelSpellCheckerTests method testMultiGenerator.

public void testMultiGenerator() 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));
        }
    });
    mapping.put("body_reverse", new Analyzer() {

        @Override
        protected TokenStreamComponents createComponents(String fieldName) {
            Tokenizer t = new StandardTokenizer();
            return new TokenStreamComponents(t, new ReverseStringFilter(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", "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_reverse", line, TextField.TYPE_NOT_STORED));
        doc.add(new Field("body_ngram", line, TextField.TYPE_NOT_STORED));
        writer.addDocument(doc);
    }
    DirectoryReader ir = DirectoryReader.open(writer);
    LaplaceScorer 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 forward = new DirectCandidateGenerator(spellchecker, "body", SuggestMode.SUGGEST_ALWAYS, ir, 0.95, 10);
    DirectCandidateGenerator reverse = new DirectCandidateGenerator(spellchecker, "body_reverse", SuggestMode.SUGGEST_ALWAYS, ir, 0.95, 10, wrapper, wrapper, MultiFields.getTerms(ir, "body_reverse"));
    CandidateGenerator generator = new MultiCandidateGeneratorWrapper(10, forward, reverse);
    Correction[] corrections = suggester.getCorrections(wrapper, new BytesRef("american cae"), generator, 1, 1, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("american ace"));
    generator = new MultiCandidateGeneratorWrapper(5, forward, reverse);
    corrections = suggester.getCorrections(wrapper, new BytesRef("american ame"), generator, 1, 1, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("american ace"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("american cae"), forward, 1, 1, ir, "body", wordScorer, 1, 2).corrections;
    // only use forward with constant prefix
    assertThat(corrections.length, equalTo(0));
    corrections = suggester.getCorrections(wrapper, new BytesRef("america cae"), generator, 2, 1, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("american ace"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("Zorr the Got-Jewel"), generator, 0.5f, 4, ir, "body", wordScorer, 0, 2).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("zorr the god jewel"));
    assertThat(corrections[2].join(new BytesRef(" ")).utf8ToString(), equalTo("four the god jewel"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("Zorr the Got-Jewel"), generator, 0.5f, 1, ir, "body", wordScorer, 1.5f, 2).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the god jewel"));
    corrections = suggester.getCorrections(wrapper, new BytesRef("Xor the Got-Jewel"), generator, 0.5f, 1, ir, "body", wordScorer, 1.5f, 2).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("xorr the god jewel"));
    // Test a special case where one of the suggest term is unchanged by the postFilter, 'II' here is unchanged by the reverse analyzer.
    corrections = suggester.getCorrections(wrapper, new BytesRef("Quazar II"), generator, 1, 1, ir, "body", wordScorer, 1, 2).corrections;
    assertThat(corrections.length, equalTo(1));
    assertThat(corrections[0].join(new BytesRef(" ")).utf8ToString(), equalTo("quasar ii"));
}
Also used : HashMap(java.util.HashMap) WhitespaceAnalyzer(org.apache.lucene.analysis.core.WhitespaceAnalyzer) Analyzer(org.apache.lucene.analysis.Analyzer) Document(org.apache.lucene.document.Document) Field(org.apache.lucene.document.Field) TextField(org.apache.lucene.document.TextField) ShingleFilter(org.apache.lucene.analysis.shingle.ShingleFilter) ReverseStringFilter(org.apache.lucene.analysis.reverse.ReverseStringFilter) Tokenizer(org.apache.lucene.analysis.Tokenizer) StandardTokenizer(org.apache.lucene.analysis.standard.StandardTokenizer) DirectSpellChecker(org.apache.lucene.search.spell.DirectSpellChecker) BytesRef(org.apache.lucene.util.BytesRef) WhitespaceAnalyzer(org.apache.lucene.analysis.core.WhitespaceAnalyzer) DirectoryReader(org.apache.lucene.index.DirectoryReader) RAMDirectory(org.apache.lucene.store.RAMDirectory) PerFieldAnalyzerWrapper(org.apache.lucene.analysis.miscellaneous.PerFieldAnalyzerWrapper) IndexWriter(org.apache.lucene.index.IndexWriter) StandardTokenizer(org.apache.lucene.analysis.standard.StandardTokenizer) LowerCaseFilter(org.apache.lucene.analysis.LowerCaseFilter) IndexWriterConfig(org.apache.lucene.index.IndexWriterConfig)

Aggregations

Tokenizer (org.apache.lucene.analysis.Tokenizer)611 MockTokenizer (org.apache.lucene.analysis.MockTokenizer)288 Analyzer (org.apache.lucene.analysis.Analyzer)269 StringReader (java.io.StringReader)264 TokenStream (org.apache.lucene.analysis.TokenStream)245 KeywordTokenizer (org.apache.lucene.analysis.core.KeywordTokenizer)216 Reader (java.io.Reader)91 WhitespaceTokenizer (org.apache.lucene.analysis.core.WhitespaceTokenizer)77 StandardTokenizer (org.apache.lucene.analysis.standard.StandardTokenizer)73 StopFilter (org.apache.lucene.analysis.StopFilter)56 SetKeywordMarkerFilter (org.apache.lucene.analysis.miscellaneous.SetKeywordMarkerFilter)55 LowerCaseFilter (org.apache.lucene.analysis.LowerCaseFilter)51 MockAnalyzer (org.apache.lucene.analysis.MockAnalyzer)47 CharArraySet (org.apache.lucene.analysis.CharArraySet)44 StandardFilter (org.apache.lucene.analysis.standard.StandardFilter)37 CharTermAttribute (org.apache.lucene.analysis.tokenattributes.CharTermAttribute)35 ESTestCase (org.elasticsearch.test.ESTestCase)30 HashMap (java.util.HashMap)24 TokenFilter (org.apache.lucene.analysis.TokenFilter)24 Random (java.util.Random)20