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, 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.solr.updateprocessors;
029
030import java.io.IOException;
031import java.util.regex.Pattern;
032
033import javax.annotation.Nonnull;
034
035import org.apache.solr.common.SolrInputDocument;
036import org.apache.solr.update.AddUpdateCommand;
037import org.apache.solr.update.processor.UpdateRequestProcessor;
038
039/**
040 * Implementation of an {@link UpdateRequestProcessor} that
041 * <ul>
042 *  <li>reads a field's (source) value,</li>
043 *  <li>applies a regex replacement,</li>
044 *  <li>adds a field (target) with the modified value.</li>
045 * </ul>
046 *
047 *  @see CmsSolrCopyModifiedUpateProcessorFactory
048 */
049public class CmsSolrCopyModifiedUpateProcessor extends UpdateRequestProcessor {
050
051    /** The source field where the original value is read from. */
052    private final String m_source;
053    /** The target field where the modified value is written to. */
054    private final String m_target;
055    /** The regular expression that the original value is matched to. */
056    private final Pattern m_regex;
057    /** The replacement for the matches of the regex on the original value. */
058    private final String m_replacement;
059
060    /**
061     * Default constructor.
062     * @param source the name of the source field (where the original value is read from).
063     * @param target the name of the target field (where the modified value is written to).
064     * @param regex the regex applied to the original value.
065     * @param replacement the replacement for the matched parts in the original value.
066     * @param nextProcessor the {@link UpdateRequestProcessor} to process next.
067     */
068    public CmsSolrCopyModifiedUpateProcessor(
069        @Nonnull String source,
070        @Nonnull String target,
071        @Nonnull Pattern regex,
072        @Nonnull String replacement,
073        UpdateRequestProcessor nextProcessor) {
074        super(nextProcessor);
075        m_source = source;
076        m_target = target;
077        m_regex = regex;
078        m_replacement = replacement;
079    }
080
081    /**
082     * @see org.apache.solr.update.processor.UpdateRequestProcessor#processAdd(org.apache.solr.update.AddUpdateCommand)
083     */
084    @Override
085    public void processAdd(AddUpdateCommand cmd) throws IOException {
086
087        SolrInputDocument doc = cmd.getSolrInputDocument();
088
089        String v = (String)doc.getFieldValue(m_source);
090        if (v != null) {
091            String v1 = m_regex.matcher(v).replaceAll(m_replacement);
092            doc.addField(m_target, v1);
093        }
094
095        // pass it up the chain
096        super.processAdd(cmd);
097    }
098}