001/*
002 * This library is part of OpenCms -
003 * the Open Source Content Management System
004 *
005 * Copyright (c) Alkacon Software GmbH & Co. KG (http://www.alkacon.com)
006 *
007 * This library is free software; you can redistribute it and/or
008 * modify it under the terms of the GNU Lesser General Public
009 * License as published by the Free Software Foundation; either
010 * version 2.1 of the License, or (at your option) any later version.
011 *
012 * This library is distributed in the hope that it will be useful,
013 * but WITHOUT ANY WARRANTY; without even the implied warranty of
014 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
015 * Lesser General Public License for more details.
016 *
017 * For further information about Alkacon Software GmbH & Co. KG, please see the
018 * company website: http://www.alkacon.com
019 *
020 * For further information about OpenCms, please see the
021 * project website: http://www.opencms.org
022 *
023 * You should have received a copy of the GNU Lesser General Public
024 * License along with this library; if not, write to the Free Software
025 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
026 */
027
028package org.opencms.search.documents;
029
030import org.opencms.file.CmsFile;
031import org.opencms.file.CmsObject;
032import org.opencms.file.CmsResource;
033import org.opencms.main.CmsException;
034import org.opencms.search.CmsIndexException;
035import org.opencms.search.I_CmsSearchIndex;
036import org.opencms.search.extractors.CmsExtractorMsOfficeOOXML;
037import org.opencms.search.extractors.I_CmsExtractionResult;
038
039/**
040 * Lucene document factory class to extract text data from a VFS resource that is an OOXML MS Office document.<p>
041 *
042 * Supported formats are MS Word (.docx), MS PowerPoint (.pptx) and MS Excel (.xlsx).<p>
043 *
044 * The OLE 2 format was introduced in Microsoft Office version 97 and was the default format until Office version 2007
045 * and the new XML-based OOXML format.<p>
046 *
047 * @since 8.0.1
048 */
049public class CmsDocumentMsOfficeOOXML extends A_CmsVfsDocument {
050
051    /**
052     * Creates a new instance of this lucene document factory.<p>
053     *
054     * @param name name of the document type
055     */
056    public CmsDocumentMsOfficeOOXML(String name) {
057
058        super(name);
059    }
060
061    /**
062     * Returns the raw text content of a given vfs resource containing MS Word data.<p>
063     *
064     * @see org.opencms.search.documents.I_CmsSearchExtractor#extractContent(CmsObject, CmsResource, I_CmsSearchIndex)
065     */
066    public I_CmsExtractionResult extractContent(CmsObject cms, CmsResource resource, I_CmsSearchIndex index)
067    throws CmsIndexException, CmsException {
068
069        logContentExtraction(resource, index);
070        CmsFile file = readFile(cms, resource);
071        try {
072            return CmsExtractorMsOfficeOOXML.getExtractor().extractText(file.getContents());
073        } catch (Exception e) {
074            throw new CmsIndexException(
075                Messages.get().container(Messages.ERR_TEXT_EXTRACTION_1, resource.getRootPath()),
076                e);
077        }
078    }
079
080    /**
081     * @see org.opencms.search.documents.I_CmsDocumentFactory#isLocaleDependend()
082     */
083    public boolean isLocaleDependend() {
084
085        return false;
086    }
087
088    /**
089     * @see org.opencms.search.documents.I_CmsDocumentFactory#isUsingCache()
090     */
091    public boolean isUsingCache() {
092
093        return true;
094    }
095}