Classes in this File | Line Coverage | Branch Coverage | Complexity | ||||
LuceneAnalyzer |
|
| 1.5;1.5 |
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 or later | |
5 | * as published by the Free Software Foundation. This program is distributed | |
6 | * in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even | |
7 | * the 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 | * © CrossWire Bible Society, 2005 - 2016 | |
18 | * | |
19 | */ | |
20 | package org.crosswire.jsword.index.lucene.analysis; | |
21 | ||
22 | import java.io.Reader; | |
23 | ||
24 | import org.apache.lucene.analysis.Analyzer; | |
25 | import org.apache.lucene.analysis.PerFieldAnalyzerWrapper; | |
26 | import org.apache.lucene.analysis.SimpleAnalyzer; | |
27 | import org.apache.lucene.analysis.TokenStream; | |
28 | import org.crosswire.jsword.book.Book; | |
29 | import org.crosswire.jsword.index.lucene.IndexMetadata; | |
30 | import org.crosswire.jsword.index.lucene.InstalledIndex; | |
31 | import org.crosswire.jsword.index.lucene.LuceneIndex; | |
32 | import org.slf4j.Logger; | |
33 | import org.slf4j.LoggerFactory; | |
34 | ||
35 | /** | |
36 | * A specialized analyzer for Books that analyzes different fields differently. | |
37 | * This is book specific since it is possible that each book has specialized | |
38 | * search requirements. | |
39 | * | |
40 | * Uses AnalyzerFactory for InstalledIndexVersion > 1.1 | |
41 | * | |
42 | * @see gnu.lgpl.License The GNU Lesser General Public License for details. | |
43 | * @author DM Smith | |
44 | */ | |
45 | public class LuceneAnalyzer extends Analyzer { | |
46 | ||
47 | 0 | public LuceneAnalyzer(Book book) { |
48 | // The default analysis | |
49 | 0 | analyzer = new PerFieldAnalyzerWrapper(new SimpleAnalyzer()); |
50 | ||
51 | 0 | if (InstalledIndex.instance().getInstalledIndexDefaultVersion() > IndexMetadata.INDEX_VERSION_1_1) { |
52 | // Content is analyzed using natural language analyzer | |
53 | // (stemming, stopword etc) | |
54 | 0 | Analyzer myNaturalLanguageAnalyzer = AnalyzerFactory.getInstance().createAnalyzer(book); |
55 | 0 | analyzer.addAnalyzer(LuceneIndex.FIELD_BODY, myNaturalLanguageAnalyzer); |
56 | //analyzer.addAnalyzer(LuceneIndex.FIELD_HEADING, myNaturalLanguageAnalyzer); //heading to use same analyzer as BODY | |
57 | //analyzer.addAnalyzer(LuceneIndex.FIELD_INTRO, myNaturalLanguageAnalyzer); | |
58 | 0 | log.debug("{}: Using languageAnalyzer: {}", book.getBookMetaData().getInitials(), myNaturalLanguageAnalyzer.getClass().getName()); |
59 | } | |
60 | ||
61 | // Keywords are normalized to osisIDs | |
62 | 0 | analyzer.addAnalyzer(LuceneIndex.FIELD_KEY, new KeyAnalyzer()); |
63 | ||
64 | // Strong's Numbers are normalized to a consistent representation | |
65 | 0 | analyzer.addAnalyzer(LuceneIndex.FIELD_STRONG, new StrongsNumberAnalyzer()); |
66 | ||
67 | // Strong's Numbers and Robinson's morphological codes are normalized to a consistent representation | |
68 | 0 | analyzer.addAnalyzer(LuceneIndex.FIELD_MORPHOLOGY, new MorphologyAnalyzer()); |
69 | ||
70 | // XRefs are normalized from ranges into a list of osisIDs | |
71 | 0 | analyzer.addAnalyzer(LuceneIndex.FIELD_XREF, new XRefAnalyzer()); |
72 | ||
73 | ||
74 | 0 | } |
75 | ||
76 | @Override | |
77 | public TokenStream tokenStream(String fieldName, Reader reader) { | |
78 | 0 | return analyzer.tokenStream(fieldName, reader); |
79 | } | |
80 | ||
81 | private PerFieldAnalyzerWrapper analyzer; | |
82 | 0 | private static final Logger log = LoggerFactory.getLogger(LuceneAnalyzer.class); |
83 | ||
84 | } |