KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > opencms > search > documents > CmsDocumentXmlPage


1 /*
2  * File : $Source: /usr/local/cvs/opencms/src/org/opencms/search/documents/CmsDocumentXmlPage.java,v $
3  * Date : $Date: 2005/07/29 12:13:00 $
4  * Version: $Revision: 1.11 $
5  *
6  * This library is part of OpenCms -
7  * the Open Source Content Mananagement System
8  *
9  * Copyright (c) 2005 Alkacon Software GmbH (http://www.alkacon.com)
10  *
11  * This library is free software; you can redistribute it and/or
12  * modify it under the terms of the GNU Lesser General Public
13  * License as published by the Free Software Foundation; either
14  * version 2.1 of the License, or (at your option) any later version.
15  *
16  * This library is distributed in the hope that it will be useful,
17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19  * Lesser General Public License for more details.
20  *
21  * For further information about Alkacon Software GmbH, please see the
22  * company website: http://www.alkacon.com
23  *
24  * For further information about OpenCms, please see the
25  * project website: http://www.opencms.org
26  *
27  * You should have received a copy of the GNU Lesser General Public
28  * License along with this library; if not, write to the Free Software
29  * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
30  */
31
32 package org.opencms.search.documents;
33
34 import org.opencms.file.CmsFile;
35 import org.opencms.file.CmsObject;
36 import org.opencms.file.CmsProperty;
37 import org.opencms.file.CmsPropertyDefinition;
38 import org.opencms.file.CmsResource;
39 import org.opencms.i18n.CmsLocaleManager;
40 import org.opencms.main.CmsException;
41 import org.opencms.main.OpenCms;
42 import org.opencms.search.A_CmsIndexResource;
43 import org.opencms.search.CmsIndexException;
44 import org.opencms.search.extractors.CmsExtractionResult;
45 import org.opencms.search.extractors.I_CmsExtractionResult;
46 import org.opencms.util.CmsHtmlExtractor;
47 import org.opencms.xml.page.CmsXmlPage;
48 import org.opencms.xml.page.CmsXmlPageFactory;
49
50 import java.util.Iterator JavaDoc;
51 import java.util.List JavaDoc;
52 import java.util.Locale JavaDoc;
53
54 /**
55  * Lucene document factory class to extract index data from a cms resource
56  * of type <code>CmsResourceTypeXmlPage</code>.<p>
57  *
58  * @author Carsten Weinholz
59  *
60  * @version $Revision: 1.11 $
61  *
62  * @since 6.0.0
63  */
64 public class CmsDocumentXmlPage extends A_CmsVfsDocument {
65
66     /**
67      * Creates a new instance of this lucene document factory.<p>
68      *
69      * @param name name of the documenttype
70      */
71     public CmsDocumentXmlPage(String JavaDoc name) {
72
73         super(name);
74     }
75
76     /**
77      * Returns the raw text content of a given vfs resource of type <code>CmsResourceTypeXmlPage</code>.<p>
78      *
79      * @see org.opencms.search.documents.A_CmsVfsDocument#extractContent(org.opencms.file.CmsObject, org.opencms.search.A_CmsIndexResource, java.lang.String)
80      */
81     public I_CmsExtractionResult extractContent(CmsObject cms, A_CmsIndexResource indexResource, String JavaDoc language)
82     throws CmsException {
83
84         CmsResource resource = (CmsResource)indexResource.getData();
85         String JavaDoc result = null;
86
87         try {
88             String JavaDoc path = cms.getRequestContext().removeSiteRoot(resource.getRootPath());
89
90             CmsFile file = CmsFile.upgrade(resource, cms);
91             String JavaDoc absolutePath = cms.getSitePath(file);
92             CmsXmlPage page = CmsXmlPageFactory.unmarshal(cms, file);
93
94             List JavaDoc pageLocales = page.getLocales();
95             if (pageLocales.size() == 0) {
96                 pageLocales = OpenCms.getLocaleManager().getDefaultLocales(cms, absolutePath);
97             }
98             Locale JavaDoc locale = OpenCms.getLocaleManager().getBestMatchingLocale(
99                 CmsLocaleManager.getLocale(language),
100                 OpenCms.getLocaleManager().getDefaultLocales(cms, absolutePath),
101                 pageLocales);
102
103             List JavaDoc elements = page.getNames(locale);
104             StringBuffer JavaDoc content = new StringBuffer JavaDoc();
105             for (Iterator JavaDoc i = elements.iterator(); i.hasNext();) {
106                 String JavaDoc value = page.getStringValue(cms, (String JavaDoc)i.next(), locale);
107                 if (value != null) {
108                     content.append(value);
109                 }
110             }
111
112             result = CmsHtmlExtractor.extractText(content.toString(), page.getEncoding());
113
114             CmsProperty extractionClass = cms.readPropertyObject(
115                 path,
116                 CmsPropertyDefinition.PROPERTY_SEARCH_EXTRACTIONCLASS,
117                 true);
118             if (extractionClass != CmsProperty.getNullProperty()) {
119                 Object JavaDoc ext = Class.forName(extractionClass.getValue()).newInstance();
120
121                 if (ext instanceof I_CmsSearchExtractor) {
122                     I_CmsSearchExtractor extra = (I_CmsSearchExtractor)ext;
123                     I_CmsExtractionResult extract = extra.extractContent(cms, indexResource, language);
124                     result = result + "\n" + extract.getContent();
125                     extract.release();
126                 } else {
127                     throw new CmsIndexException(Messages.get().container(
128                         Messages.ERR_EXTRACTION_CLASS_2,
129                         resource.getRootPath(),
130                         ext.getClass().getName()));
131                 }
132             }
133
134             return new CmsExtractionResult(result);
135
136         } catch (Exception JavaDoc e) {
137             throw new CmsIndexException(
138                 Messages.get().container(Messages.ERR_TEXT_EXTRACTION_1, resource.getRootPath()),
139                 e);
140         }
141     }
142 }
Popular Tags