1   /**
2    * Distribution License:
3    * JSword is free software; you can redistribute it and/or modify it under
4    * the terms of the GNU Lesser General Public License, version 2.1 as published by
5    * the Free Software Foundation. This program is distributed in the hope
6    * that it will be useful, but WITHOUT ANY WARRANTY; without even the
7    * implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
8    * See the GNU Lesser General Public License for more details.
9    *
10   * The License is available on the internet at:
11   *       http://www.gnu.org/copyleft/lgpl.html
12   * or by writing to:
13   *      Free Software Foundation, Inc.
14   *      59 Temple Place - Suite 330
15   *      Boston, MA 02111-1307, USA
16   *
17   * Copyright: 2009
18   *     The copyright to this program is held by it's authors.
19   *
20   * ID: $Id:  $
21   */
22  package org.crosswire.jsword.index.lucene.analysis;
23  
24  import java.io.IOException;
25  import java.io.Reader;
26  
27  import org.apache.lucene.analysis.TokenStream;
28  import org.apache.lucene.analysis.cn.smart.SmartChineseAnalyzer;
29  import org.apache.lucene.util.Version;
30  
31  /**
32   * A simple wrapper for {@link SmartChineseAnalyzer}, which takes overlapping
33   * two character tokenization approach which leads to larger index size, like
34   * {@link org.apache.lucene.analyzer.cjk.CJKAnalyzer}. This analyzer's stop list
35   * is merely of punctuation. It does stemming of English.
36   * 
37   * @see gnu.lgpl.License for license details.<br>
38   *      The copyright to this program is held by it's authors.
39   * @author DM Smith [dmsmith555 at yahoo dot com]
40   */
41  public class SmartChineseLuceneAnalyzer extends AbstractBookAnalyzer {
42      public SmartChineseLuceneAnalyzer() {
43          myAnalyzer = new SmartChineseAnalyzer(Version.LUCENE_29);
44      }
45  
46      /* (non-Javadoc)
47       * @see org.apache.lucene.analysis.Analyzer#tokenStream(java.lang.String, java.io.Reader)
48       */
49      @Override
50      public final TokenStream tokenStream(String fieldName, Reader reader) {
51          return myAnalyzer.tokenStream(fieldName, reader);
52      }
53  
54      /* (non-Javadoc)
55       * @see org.apache.lucene.analysis.Analyzer#reusableTokenStream(java.lang.String, java.io.Reader)
56       */
57      @Override
58      public final TokenStream reusableTokenStream(String fieldName, Reader reader) throws IOException {
59          return myAnalyzer.reusableTokenStream(fieldName, reader);
60      }
61  
62      private SmartChineseAnalyzer myAnalyzer;
63  }
64