cmssearchindex.java

来自「找了很久才找到到源代码」· Java 代码 · 共 1,124 行 · 第 1/3 页

/*
 * File   : $Source: /usr/local/cvs/opencms/src/org/opencms/search/CmsSearchIndex.java,v $
 * Date   : $Date: 2007-08-27 11:28:14 $
 * Version: $Revision: 1.65 $
 *
 * This library is part of OpenCms -
 * the Open Source Content Management System
 *
 * Copyright (c) 2002 - 2007 Alkacon Software GmbH (http://www.alkacon.com)
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
 * Lesser General Public License for more details.
 *
 * For further information about Alkacon Software GmbH, please see the
 * company website: http://www.alkacon.com
 *
 * For further information about OpenCms, please see the
 * project website: http://www.opencms.org
 * 
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

package org.opencms.search;

import org.opencms.configuration.I_CmsConfigurationParameterHandler;
import org.opencms.file.CmsObject;
import org.opencms.file.CmsProject;
import org.opencms.file.CmsRequestContext;
import org.opencms.file.CmsResource;
import org.opencms.i18n.CmsLocaleManager;
import org.opencms.main.CmsException;
import org.opencms.main.CmsIllegalArgumentException;
import org.opencms.main.CmsLog;
import org.opencms.main.OpenCms;
import org.opencms.search.documents.A_CmsVfsDocument;
import org.opencms.search.documents.I_CmsDocumentFactory;
import org.opencms.search.documents.I_CmsTermHighlighter;
import org.opencms.search.fields.CmsSearchField;
import org.opencms.search.fields.CmsSearchFieldConfiguration;
import org.opencms.util.CmsStringUtil;

import java.io.File;
import java.io.IOException;
import java.text.ParseException;
import java.util.ArrayList;
import java.util.Date;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Locale;
import java.util.Map;
import java.util.TreeMap;

import org.apache.commons.logging.Log;
import org.apache.lucene.analysis.Analyzer;
import org.apache.lucene.document.DateTools;
import org.apache.lucene.document.Document;
import org.apache.lucene.document.Field;
import org.apache.lucene.index.IndexWriter;
import org.apache.lucene.index.Term;
import org.apache.lucene.queryParser.QueryParser;
import org.apache.lucene.search.BooleanClause;
import org.apache.lucene.search.BooleanQuery;
import org.apache.lucene.search.Hits;
import org.apache.lucene.search.IndexSearcher;
import org.apache.lucene.search.PhraseQuery;
import org.apache.lucene.search.Query;
import org.apache.lucene.search.TermQuery;

/**
 * Implements the search within an index and the management of the index configuration.<p>
 * 
 * @author Carsten Weinholz 
 * @author Thomas Weckert  
 * @author Alexander Kandzior 
 * 
 * @version $Revision: 1.65 $ 
 * 
 * @since 6.0.0 
 */
public class CmsSearchIndex implements I_CmsConfigurationParameterHandler {

    /** Constant for additional param to enable excerpt creation (default: true). */
    public static final String EXCERPT = CmsSearchIndex.class.getName() + ".createExcerpt";

    /** Constant for additional param to enable permission checks (default: true). */
    public static final String PERMISSIONS = CmsSearchIndex.class.getName() + ".checkPermissions";

    /** Constant for additional param to set the thread priority during search. */
    public static final String PRIORITY = CmsSearchIndex.class.getName() + ".priority";

    /** Automatic ("auto") index rebuild mode. */
    public static final String REBUILD_MODE_AUTO = "auto";

    /** Manual ("manual") index rebuild mode. */
    public static final String REBUILD_MODE_MANUAL = "manual";

    /** 
     * Special root path append token for optimized path queries.<p>
     * 
     * @deprecated This is not longer requires since OpenCms version 7.0.2, since the implementation 
     * of {@link CmsSearchManager#getAnalyzer(Locale)} was modified to use always 
     * use for the {@link CmsSearchField#FIELD_ROOT} filed.
     * 
     * @see #rootPathRewrite(String)
     */
    public static final String ROOT_PATH_SUFFIX = "";

    /** Special root path start token for optimized path queries. */
    public static final String ROOT_PATH_TOKEN = "root";

    /** Constant for a field list that contains the "meta" field as well as the "content" field. */
    static final String[] DOC_META_FIELDS = new String[] {CmsSearchField.FIELD_META, CmsSearchField.FIELD_CONTENT};

    /** The log object for this class. */
    private static final Log LOG = CmsLog.getLog(CmsSearchIndex.class);

    /** The list of configured index sources. */
    List m_sources;

    /** The excerpt mode for this index. */
    private boolean m_createExcerpt;

    /** Documenttypes of folders/channels. */
    private Map m_documenttypes;

    /** The permission check mode for this index. */
    private boolean m_dontCheckPermissions;

    /** An internal enabled flag, used to disable the index if for instance the configured project does not exist. */
    private boolean m_enabled;

    /** The search field configuration of this index. */
    private CmsSearchFieldConfiguration m_fieldConfiguration;

    /** The name of the search field configuration used by this index. */
    private String m_fieldConfigurationName;

    /** The locale of this index. */
    private Locale m_locale;

    /** The name of this index. */
    private String m_name;

    /** The path where this index stores it's data in the "real" file system. */
    private String m_path;

    /** The thread priority for a search. */
    private int m_priority;

    /** The project of this index. */
    private String m_project;

    /** The rebuild mode for this index. */
    private String m_rebuild;

    /** The configured sources for this index. */
    private List m_sourceNames;

    /**
     * Default constructor only intended to be used by the xml configuration. <p>
     * 
     * It is recommended to use the constructor <code>{@link #CmsSearchIndex(String)}</code> 
     * as it enforces the mandatory name argument. <p>
     * 
     */
    public CmsSearchIndex() {

        m_sourceNames = new ArrayList();
        m_documenttypes = new HashMap();
        m_createExcerpt = true;
        m_enabled = true;
        m_priority = -1;
    }

    /**
     * Creates a new CmsSearchIndex with the given name.<p>
     * 
     * @param name the system-wide unique name for the search index 
     * 
     * @throws org.opencms.main.CmsIllegalArgumentException 
     *   if the given name is null, empty or already taken 
     *   by another search index. 
     * 
     */
    public CmsSearchIndex(String name)
    throws CmsIllegalArgumentException {

        this();
        setName(name);
    }

    /**
     * Rewrites the a resource path for use in the {@link CmsSearchField#FIELD_ROOT} field.<p>
     * 
     * This is required in order to use a Lucene "phrase query" on the resource path.
     * Using a phrase query is much, much better for the search performance then using a straightforward 
     * "prefix query". With a "prefix query", Lucene would interally generate a huge list of boolean sub-queries,
     * exactly one for every document in the VFS subtree of the query. So if you query on "/sites/default/*" on 
     * a large OpenCms installation, this means thousands of sub-queries.
     * Using the "phrase query", only one (or very few) queries are internally generated, and the result 
     * is just the same.<p>  
     * 
     * Since OpenCms version 7.0.2, the {@link CmsSearchField#FIELD_ROOT} field always uses a whitespace analyzer.
     * This is ensured by the {@link CmsSearchManager#getAnalyzer(Locale)} implementation. 
     * The Lucene whitespace analyzer uses all words as tokens, no lower case transformation or word stemming is done. 
     * So the root path is now just split along the '/' chars, which are replaced by simple space chars.<p>
     * 
     * <i>Historical implementation sidenote:</i>
     * Before 7.0.2, the {@link CmsSearchField#FIELD_ROOT} used the analyzer configured by the language. 
     * This introduced a number of issues as the language analyzer might modify the directory names, leading to potential
     * duplicates (e.g. <code>members/</code> and <code>member/</code> may both be trimmed to <code>member</code>),
     * so that the prefix search returns more or different results then expected. 
     * This was avoided by a workaround where this method basically replaced the "/" of a path with "@o.c ". 
     * Using this trick most Lucene analyzers left the directory names untouched, 
     * and treated them like literal email addresses. However, this trick did not work with all analyzers,
     * for example the Russian analyzer does not work as expected.
     * An additional workaround was required to avoid problems with folders that that are different
     * only by the upper / lower chars. Since 7.0.2, these workarounds are not longer required, since the 
     * {@link CmsSearchField#FIELD_ROOT} field always uses a whitespace analyzer, which is a much better solution.<p>
     * 
     * @param path the path to rewrite
     * 
     * @return the re-written path
     */
    public static String rootPathRewrite(String path) {

        StringBuffer result = new StringBuffer(256);
        String[] elements = rootPathSplit(path);
        for (int i = 0; i < elements.length; i++) {
            result.append(elements[i]);
            if ((i + 1) < elements.length) {
                result.append(' ');
            }
        }
        return result.toString();
    }

    /**
     * Spits the a resource path into tokens for use in the <code>{@link CmsSearchField#FIELD_ROOT}</code> field
     * and with the <code>{@link #rootPathRewrite(String)}</code> method.<p>
     * 
     * @param path the path to split
     * 
     * @return the split path
     * 
     * @see #rootPathRewrite(String)
     */
    public static String[] rootPathSplit(String path) {

        if (CmsStringUtil.isEmpty(path)) {
            return new String[] {ROOT_PATH_TOKEN};
        }

        // split the path
        String[] elements = CmsStringUtil.splitAsArray(path, '/');
        String[] result = new String[elements.length + 1];
        result[0] = ROOT_PATH_TOKEN;
        System.arraycopy(elements, 0, result, 1, elements.length);
        return result;
    }

    /**
     * Adds a parameter.<p>
     * 
     * @param key the key/name of the parameter
     * @param value the value of the parameter
     */
    public void addConfigurationParameter(String key, String value) {

        if (PERMISSIONS.equals(key)) {
            m_dontCheckPermissions = !Boolean.valueOf(value).booleanValue();
        } else if (EXCERPT.equals(key)) {
            m_createExcerpt = Boolean.valueOf(value).booleanValue();
        } else if (PRIORITY.equals(key)) {
            m_priority = Integer.parseInt(value);
            if (m_priority < Thread.MIN_PRIORITY) {
                m_priority = Thread.MIN_PRIORITY;
                LOG.error(Messages.get().getBundle().key(
                    Messages.LOG_SEARCH_PRIORITY_TOO_LOW_2,
                    value,
                    new Integer(Thread.MIN_PRIORITY)));

            } else if (m_priority > Thread.MAX_PRIORITY) {
                m_priority = Thread.MAX_PRIORITY;
                LOG.debug(Messages.get().getBundle().key(
                    Messages.LOG_SEARCH_PRIORITY_TOO_HIGH_2,
                    value,
                    new Integer(Thread.MAX_PRIORITY)));

            }
        }
    }

    /**
     * Adds am index source to this search index.<p>
     * 
     * @param sourceName the index source name to add
     */
    public void addSourceName(String sourceName) {

        m_sourceNames.add(sourceName);
    }

    /**
     * Checks is this index has been configured correctly.<p>
     * 
     * In case the check fails, the <code>enabled</code> property
     * is set to <code>false</code>
     * 
     * @param cms a OpenCms user context to perform the checks with (should have "Administrator" permissions)
     *
     * @return <code>true</code> in case the index is correctly configured and enabled after the check
     * 
     * @see #isEnabled()
     */
    public boolean checkConfiguration(CmsObject cms) {

        if (isEnabled()) {
            // check if the project for the index exists        
            try {
                cms.readProject(getProject());
                setEnabled(true);
            } catch (CmsException e) {
                // the project does not exist, disable the index
                setEnabled(false);
                if (LOG.isErrorEnabled()) {
                    LOG.error(Messages.get().getBundle().key(
                        Messages.LOG_SEARCHINDEX_CREATE_BAD_PROJECT_2,
                        getProject(),
                        getName()));
                }
            }
        } else {
            if (LOG.isInfoEnabled()) {
                LOG.info(Messages.get().getBundle().key(Messages.LOG_SEARCHINDEX_DISABLED_1, getName()));
            }
        }

        return isEnabled();
    }

    /**
     * @see java.lang.Object#equals(java.lang.Object)
     */
    public boolean equals(Object obj) {

        if (obj == this) {
            return true;
        }
        if (obj instanceof CmsSearchIndex) {
            return ((CmsSearchIndex)obj).m_name.equals(m_name);
        }
        return false;
    }

    /**
     * @see org.opencms.configuration.I_CmsConfigurationParameterHandler#getConfiguration()
     */
    public Map getConfiguration() {

        Map result = new TreeMap();
        if (m_priority > 0) {
            result.put(PRIORITY, new Integer(m_priority));
        }
        if (!m_createExcerpt) {