KickJava   Java API By Example, From Geeks To Geeks.

Java > Open Source Codes > org > opencms > search > extractors > CmsExtractorMsWord


1 /*
2  * File : $Source: /usr/local/cvs/opencms/src/org/opencms/search/extractors/CmsExtractorMsWord.java,v $
3  * Date : $Date: 2005/07/29 10:35:06 $
4  * Version: $Revision: 1.8 $
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.extractors;
33
34 import java.io.InputStream JavaDoc;
35 import java.util.Map JavaDoc;
36
37 import org.apache.poi.poifs.eventfilesystem.POIFSReader;
38
39 import org.textmining.text.extraction.WordExtractor;
40
41 /**
42  * Extracts the text form an MS Word document.<p>
43  *
44  * @author Alexander Kandzior
45  *
46  * @version $Revision: 1.8 $
47  *
48  * @since 6.0.0
49  */
50 public final class CmsExtractorMsWord extends A_CmsTextExtractorMsOfficeBase {
51
52     /** Static member instance of the extractor. */
53     private static final CmsExtractorMsWord INSTANCE = new CmsExtractorMsWord();
54
55     /**
56      * Hide the public constructor.<p>
57      */
58     private CmsExtractorMsWord() {
59
60         // noop
61 }
62
63     /**
64      * Returns an instance of this text extractor.<p>
65      *
66      * @return an instance of this text extractor
67      */
68     public static I_CmsTextExtractor getExtractor() {
69
70         return INSTANCE;
71     }
72
73     /**
74      * @see org.opencms.search.extractors.I_CmsTextExtractor#extractText(java.io.InputStream, java.lang.String)
75      */
76     public I_CmsExtractionResult extractText(InputStream JavaDoc in, String JavaDoc encoding) throws Exception JavaDoc {
77
78         // first extract the text using the text actraction libary
79 WordExtractor wordExtractor = new WordExtractor();
80         String JavaDoc result = wordExtractor.extractText(getStreamCopy(in));
81         result = removeControlChars(result);
82
83         // now extract the meta information using POI
84 POIFSReader reader = new POIFSReader();
85         reader.registerListener(this);
86         reader.read(getStreamCopy(in));
87         Map JavaDoc metaInfo = extractMetaInformation();
88
89         // free some memory
90 cleanup();
91         
92         // return the final result
93 return new CmsExtractionResult(result, metaInfo);
94     }
95 }
Popular Tags