Example usage for org.apache.lucene.analysis.charfilter HTMLStripCharFilterFactory HTMLStripCharFilterFactory

Introduction

In this page you can find the example usage for org.apache.lucene.analysis.charfilter HTMLStripCharFilterFactory HTMLStripCharFilterFactory.

Prototype

public HTMLStripCharFilterFactory(Map<String, String> args) 

Source Link

Document

Creates a new HTMLStripCharFilterFactory

Usage

From source file:com.basistech.IndexFiles.java

License:Open Source License

static Analyzer setupAnalyzer(final boolean filters) {
    Map<String, String> tokenizerFactoryArgs = Maps.newHashMap();
    tokenizerFactoryArgs.put("rlpContext", "rlp-context.xml");
    tokenizerFactoryArgs.put("lang", "kor");
    tokenizerFactoryArgs.put("postLemma", "true");
    tokenizerFactoryArgs.put("postCompoundComponents", "true");
    tokenizerFactoryArgs.put("postPartOfSpeech", "true");

    final RLPTokenizerFactory tokenizerFactory = new RLPTokenizerFactory(tokenizerFactoryArgs);
    Map<String, String> emptyOptions = Maps.newHashMap();
    final HTMLStripCharFilterFactory charFilterFactory = new HTMLStripCharFilterFactory(emptyOptions);
    final ICUFoldingFilterFactory foldingFilterFactory = new ICUFoldingFilterFactory(emptyOptions);
    final ReversedWildcardFilterFactory reversedWildcardFilterFactory = new ReversedWildcardFilterFactory(
            emptyOptions);//from w  w w.  j a v a  2  s  .  co m
    final RemoveDuplicatesTokenFilterFactory removeDuplicatesTokenFilterFactory = new RemoveDuplicatesTokenFilterFactory(
            emptyOptions);

    return new Analyzer() {
        @Override
        protected TokenStreamComponents createComponents(String fieldName, Reader reader) {
            final Tokenizer source = tokenizerFactory.create(reader);
            if (filters) {
                TokenStream filter = foldingFilterFactory.create(source);
                filter = reversedWildcardFilterFactory.create(filter);
                filter = removeDuplicatesTokenFilterFactory.create(filter);
                return new TokenStreamComponents(source, filter);
            } else {
                return new TokenStreamComponents(source);
            }
        }

        @Override
        protected Reader initReader(String fieldName, Reader reader) {
            if (filters) {
                return charFilterFactory.create(reader);
            } else {
                return reader;
            }
        }
    };
}

---

From source file:de.walware.statet.r.internal.core.rhelp.index.WriteAnalyzer.java

License:Open Source License

public WriteAnalyzer() {
    super(PER_FIELD_REUSE_STRATEGY);
    this.defaultAnalyzer = new DefaultAnalyzer();
    this.htmlAnalyzers = new DefaultAnalyzer(
            new HTMLStripCharFilterFactory(Collections.<String, String>emptyMap()));
}

---