cmssearchmanager.java

一个cms内容管理平台

        // the current list of referred indexsources of the iterated index
        List refsources;
        while (itIndexes.hasNext()) {
            idx = (CmsSearchIndex)itIndexes.next();
            refsources = idx.getSources();
            if (refsources != null) {
                if (refsources.contains(indexsource)) {
                    referrers.add(idx);
                }
            }
        }
        if (referrers.size() > 0) {
            throw new CmsIllegalStateException(Messages.get().container(
                Messages.ERR_INDEX_SOURCE_DELETE_2,
                indexsource.getName(),
                referrers.toString()));
        }

        // remove operation (no exception) 
        return m_indexSources.remove(indexsource.getName()) != null;

    }

    /**
     * Sets the name of the directory below WEB-INF/ where the search indexes are stored.<p>
     * 
     * @param value the name of the directory below WEB-INF/ where the search indexes are stored
     */
    public void setDirectory(String value) {

        m_path = value;
    }

    /**
     * Sets the highlighter.<p>
     *
     * A highlighter is a class implementing org.opencms.search.documents.I_TermHighlighter.<p>
     *
     * @param highlighter the package/class name of the highlighter
     */
    public void setHighlighter(String highlighter) {

        try {
            m_highlighter = (I_CmsTermHighlighter)Class.forName(highlighter).newInstance();
        } catch (Exception exc) {
            m_highlighter = null;
        }
    }

    /**
     * Sets the seconds to wait for an index lock during an update operation.<p>
     * 
     * @param value the seconds to wait for an index lock during an update operation
     */
    public void setIndexLockMaxWaitSeconds(int value) {

        m_indexLockMaxWaitSeconds = value;
    }

    /**
     * Sets the max. excerpt length.<p>
     *
     * @param maxExcerptLength the max. excerpt length to set
     */
    public void setMaxExcerptLength(String maxExcerptLength) {

        try {
            m_maxExcerptLength = Integer.parseInt(maxExcerptLength);
        } catch (Exception e) {
            LOG.error(Messages.get().getBundle().key(Messages.LOG_PARSE_EXCERPT_LENGTH_FAILED_1, maxExcerptLength), e);
            m_maxExcerptLength = 1024;
        }
    }

    /**
     * Sets the result cache size.<p>
     * 
     * @param value the result cache size
     */
    public void setResultCacheSize(String value) {

        m_resultCacheSize = value;
    }

    /**
     * Sets the timeout to abandon threads indexing a resource.<p>
     * 
     * @param value the timeout in milliseconds
     */
    public void setTimeout(String value) {

        m_timeout = value;
    }

    /**
     * Checks is a given index is locked, if so waits for a numer of seconds and checks again,
     * until either the index is unlocked or a limit of seconds set by <code>{@link #setIndexLockMaxWaitSeconds(int)}</code>
     * is reached.<p>
     * 
     * @param index the index to check the lock for
     * @param report the report to write error messages on
     * 
     * @return <code>true</code> if the index is locked
     */
    protected boolean checkIndexLock(CmsSearchIndex index, I_CmsReport report) {

        File indexPath = new File(index.getPath());
        // check if the target index path already exists
        if (!indexPath.exists()) {
            // if the folder does not yet exist it is also not locked
            return false;
        }

        // check if the index is locked
        boolean indexLocked = true;
        try {
            int lockSecs = 0;
            while (indexLocked && (lockSecs < m_indexLockMaxWaitSeconds)) {
                indexLocked = IndexReader.isLocked(index.getPath());
                if (indexLocked) {
                    // index is still locked, wait one second
                    report.println(Messages.get().container(
                        Messages.RPT_SEARCH_INDEXING_LOCK_WAIT_2,
                        index.getName(),
                        new Integer(m_indexLockMaxWaitSeconds - lockSecs)), I_CmsReport.FORMAT_ERROR);
                    // sleep one second
                    Thread.sleep(1000);
                    lockSecs++;
                }
            }
        } catch (Exception e) {
            LOG.error(Messages.get().getBundle().key(
                Messages.LOG_IO_INDEX_READER_OPEN_2,
                index.getPath(),
                index.getName()), e);
        }

        return indexLocked;
    }

    /**
     * Returns an analyzer for the given language.<p>
     * The analyzer is selected according to the analyzer configuration.
     * 
     * @param locale a language id, i.e. de, en, it
     * @return the appropriate lucene analyzer
     * @throws CmsIndexException if something goes wrong
     */
    protected Analyzer getAnalyzer(String locale) throws CmsIndexException {

        Analyzer analyzer = null;
        String className = null;

        CmsSearchAnalyzer analyzerConf = (CmsSearchAnalyzer)m_analyzers.get(locale);
        if (analyzerConf == null) {
            throw new CmsIndexException(Messages.get().container(Messages.ERR_ANALYZER_NOT_FOUND_1, locale));
        }

        try {
            className = analyzerConf.getClassName();
            Class analyzerClass = Class.forName(className);

            // added param for snowball analyzer
            String stemmerAlgorithm = analyzerConf.getStemmerAlgorithm();
            if (stemmerAlgorithm != null) {
                analyzer = (Analyzer)analyzerClass.getDeclaredConstructor(new Class[] {String.class}).newInstance(
                    new Object[] {stemmerAlgorithm});
            } else {
                analyzer = (Analyzer)analyzerClass.newInstance();
            }

        } catch (Exception e) {
            throw new CmsIndexException(Messages.get().container(Messages.ERR_LOAD_ANALYZER_1, className), e);
        }

        return analyzer;
    }

    /**
     * Returns a lucene document factory for given resource.<p>
     * The type of the document factory is selected by the type of the resource
     * and the mimetype of the resource content according to the documenttype configuration.
     * 
     * @param resource a cms resource
     * @return a lucene document factory or null
     */
    protected I_CmsDocumentFactory getDocumentFactory(A_CmsIndexResource resource) {

        String documentTypeKey = resource.getDocumentKey(true);

        I_CmsDocumentFactory factory = (I_CmsDocumentFactory)m_documentTypes.get(documentTypeKey);
        if (factory == null) {
            factory = (I_CmsDocumentFactory)m_documentTypes.get(resource.getDocumentKey(false));
        }

        return factory;
    }

    /**
     * Returns the set of names of all configured documenttypes.<p>
     * 
     * @return the set of names of all configured documenttypes
     */
    protected List getDocumentTypes() {

        List names = new ArrayList();
        for (Iterator i = m_documentTypes.values().iterator(); i.hasNext();) {
            I_CmsDocumentFactory factory = (I_CmsDocumentFactory)i.next();
            names.add(factory.getName());
        }

        return names;
    }

    /**
     * Returns the common cache for buffering search results.<p>
     * 
     * @return the cache
     */
    protected Map getResultCache() {

        return m_resultCache;
    }

    /**
     * Initializes the available Cms resource types to be indexed.<p>
     * 
     * A map stores document factories keyed by a string representing
     * a colon separated list of Cms resource types and/or mimetypes.<p>
     * 
     * The keys of this map are used to trigger a document factory to convert 
     * a Cms resource into a Lucene index document.<p>
     * 
     * A document factory is a class implementing the interface
     * {@link org.opencms.search.documents.I_CmsDocumentFactory}.<p>
     */
    protected void initAvailableDocumentTypes() {

        CmsSearchDocumentType documenttype = null;
        String className = null;
        String name = null;
        I_CmsDocumentFactory documentFactory = null;
        List resourceTypes = null;
        List mimeTypes = null;
        Class c = null;

        m_documentTypes = new HashMap();

        List keys = new ArrayList(m_documentTypeConfigs.keySet());
        for (int i = 0, n = keys.size(); i < n; i++) {

            documenttype = (CmsSearchDocumentType)(m_documentTypeConfigs.get(keys.get(i)));
            name = documenttype.getName();

            try {
                className = documenttype.getClassName();
                resourceTypes = documenttype.getResourceTypes();
                mimeTypes = documenttype.getMimeTypes();

                if (name == null) {
                    throw new CmsIndexException(Messages.get().container(Messages.ERR_DOCTYPE_NO_NAME_0));
                }
                if (className == null) {
                    throw new CmsIndexException(Messages.get().container(Messages.ERR_DOCTYPE_NO_CLASS_DEF_0));
                }
                if (resourceTypes.size() == 0) {
                    throw new CmsIndexException(Messages.get().container(Messages.ERR_DOCTYPE_NO_RESOURCETYPE_DEF_0));
                }

                try {
                    c = Class.forName(className);
                    documentFactory = (I_CmsDocumentFactory)c.getConstructor(new Class[] {String.class}).newInstance(
                        new Object[] {name});
                } catch (ClassNotFoundException exc) {
                    throw new CmsIndexException(
                        Messages.get().container(Messages.ERR_DOCCLASS_NOT_FOUND_1, className),
                        exc);
                } catch (Exception exc) {
                    throw new CmsIndexException(Messages.get().container(Messages.ERR_DOCCLASS_INIT_1, className), exc);
                }

                for (Iterator key = documentFactory.getDocumentKeys(resourceTypes, mimeTypes).iterator(); key.hasNext();) {
                    m_documentTypes.put(key.next(), documentFactory);
                }

            } catch (CmsException e) {
                if (LOG.isWarnEnabled()) {
                    LOG.warn(Messages.get().getBundle().key(Messages.LOG_DOCTYPE_CONFIG_FAILED_1, name), e);
                }
            }
        }
    }

    /**
     * Initializes the configured search indexes.<p>
     * 
     * This initializes also the list of Cms resources types
     * to be indexed by an index source.<p>
     */
    protected void initSearchIndexes() {

        CmsSearchIndex index = null;
        for (int i = 0, n = m_indexes.size(); i < n; i++) {
            index = (CmsSearchIndex)m_indexes.get(i);
            // reset disabled flag
            index.setEnabled(true);
            // check if the index has been configured correctly
            if (index.checkConfiguration(m_adminCms)) {
                // the index is configured correctly
                try {
                    index.initialize();
                } catch (CmsException e) {
                    // in this case the index will be disabled
                    if (CmsLog.INIT.isInfoEnabled()) {
                        CmsLog.INIT.info(Messages.get().getBundle().key(
                            Messages.INIT_SEARCH_INIT_FAILED_1,
                            index.getName()), e);
                    }
                }
            }
            if (CmsLog.INIT.isInfoEnabled()) {
                // output a log message if the index was successfully configured or not
                if (index.isEnabled()) {
                    CmsLog.INIT.info(Messages.get().getBundle().key(
                        Messages.INIT_INDEX_CONFIGURED_2,
                        index.getName(),
                        index.getProject()));
                } else {
                    CmsLog.INIT.info(Messages.get().getBundle().key(
                        Messages.INIT_INDEX_NOT_CONFIGURED_2,
                        index.getName(),
                        index.getProject()));
                }
            }
        }
    }

    /**
     * Incrementally updates all indexes that have their rebuild mode set to <code>"auto"</code>
     * after resources have been published.<p> 
     * 
     * @param adminCms an OpenCms user context with Admin permissions
     * @param publishHistoryId the history ID of the published project 
     * @param report the report to write the output to
     */
    protected synchronized void updateAllIndexes(CmsObject adminCms, CmsUUID publishHistoryId, I_CmsReport report) {

        List publishedResources;
        try {
            // read the list of all published resources