001/*
002 * This library is part of OpenCms -
003 * the Open Source Content Management System
004 *
005 * Copyright (c) Alkacon Software GmbH & Co. KG (https://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: https://www.alkacon.com
019 *
020 * For further information about OpenCms, please see the
021 * project website: https://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.extractors;
029
030import java.io.InputStream;
031
032import org.apache.tika.parser.microsoft.OfficeParser;
033
034/**
035 * Extracts text data from a VFS resource that is an OLE 2 MS Office document.<p>
036 *
037 * Supported formats are MS Word (.doc), MS PowerPoint (.ppt) and MS Excel (.xls).<p>
038 *
039 * The OLE 2 format was introduced in Microsoft Office version 97 and was the default format until Office version 2007
040 * and the new XML-based OOXML format.<p>
041 *
042 * @since 8.0.1
043 */
044public final class CmsExtractorMsOfficeOLE2 extends A_CmsTextExtractor {
045
046    /** Static member instance of the extractor. */
047    private static final CmsExtractorMsOfficeOLE2 INSTANCE_METADATA = new CmsExtractorMsOfficeOLE2(true);
048
049    /** Static member instance of the extractor. */
050    private static final CmsExtractorMsOfficeOLE2 INSTANCE_NO_METADATA = new CmsExtractorMsOfficeOLE2(false);
051
052    /** Flag, indicating if metadata should be extracted. */
053    private boolean m_extractMetaData;
054
055    /**
056     * Hide the public constructor.<p>
057     *
058     * @param extractMetaData flag, indicating if metadata should be extracted.
059     */
060    private CmsExtractorMsOfficeOLE2(boolean extractMetaData) {
061
062        m_extractMetaData = extractMetaData;
063    }
064
065    /**
066     * Returns an instance of this text extractor.<p>
067     *
068     * @return an instance of this text extractor
069     */
070    public static I_CmsTextExtractor getExtractor() {
071
072        return INSTANCE_METADATA;
073    }
074
075    /**
076     * Returns an instance of this text extractor.<p>
077     *
078     * @param extractMetaData flag, indicating if metadata should be extracted.
079     *
080     * @return an instance of this text extractor
081     */
082    public static I_CmsTextExtractor getExtractor(boolean extractMetaData) {
083
084        return extractMetaData ? INSTANCE_METADATA : INSTANCE_NO_METADATA;
085    }
086
087    /**
088     * @see org.opencms.search.extractors.I_CmsTextExtractor#extractText(java.io.InputStream, java.lang.String)
089     */
090    @Override
091    public I_CmsExtractionResult extractText(InputStream in) throws Exception {
092
093        return extractText(in, new OfficeParser(), m_extractMetaData);
094    }
095}