Source code

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.documents;
029
030import org.opencms.configuration.CmsParameterConfiguration;
031import org.opencms.file.CmsFile;
032import org.opencms.file.CmsObject;
033import org.opencms.file.CmsResource;
034import org.opencms.main.CmsException;
035import org.opencms.main.CmsLog;
036import org.opencms.search.I_CmsSearchDocument;
037import org.opencms.search.I_CmsSearchIndex;
038import org.opencms.search.extractors.I_CmsExtractionResult;
039
040import java.util.ArrayList;
041import java.util.Iterator;
042import java.util.List;
043
044import org.apache.commons.logging.Log;
045
046/**
047 * Base document factory class for a VFS <code>{@link org.opencms.file.CmsResource}</code>,
048 * just requires a specialized implementation of
049 * <code>{@link I_CmsDocumentFactory#extractContent(CmsObject, CmsResource, I_CmsSearchIndex)}</code>
050 * for text extraction from the binary document content.<p>
051 *
052 * @since 6.0.0
053 */
054public abstract class A_CmsVfsDocument implements I_CmsDocumentFactory {
055
056    /**
057     * Generic type name used as default for all types that are globally unconfigured.
058     * Note that any special xml content is already configured if xmlcontent is configured.
059     */
060    public static final String DEFAULT_ALL_UNCONFIGURED_TYPES = "__unconfigured__";
061    /** Generic type name used as default for all types. */
062    public static final String DEFAULT_ALL_TYPES = "__all__";
063    /** The log object for this class. */
064    private static final Log LOG = CmsLog.getLog(A_CmsVfsDocument.class);
065
066    /** Name of the document type. */
067    protected String m_name;
068
069    /** The cache used for storing extracted documents. */
070    private CmsExtractionResultCache m_cache;
071
072    /**
073     * Creates a new instance of this lucene document factory.<p>
074     *
075     * @param name name of the documenttype
076     */
077    public A_CmsVfsDocument(String name) {
078
079        m_name = name;
080    }
081
082    /**
083     * Creates a document factory lookup key for the given resource type name / MIME type configuration.<p>
084     *
085     * If the given <code>mimeType</code> is <code>null</code>, this indicates that the key should
086     * match all VFS resource of the given resource type regardless of the MIME type.<p>
087     *
088     * @param type the resource type name to use
089     * @param mimeType the MIME type to use
090     *
091     * @return a document factory lookup key for the given resource id / MIME type configuration
092     */
093    public static String getDocumentKey(String type, String mimeType) {
094
095        StringBuffer result = new StringBuffer(16);
096        result.append(I_CmsSearchDocument.VFS_DOCUMENT_KEY_PREFIX);
097        result.append('_');
098        result.append(type);
099        if (mimeType != null) {
100            result.append(':');
101            result.append(mimeType);
102        }
103        return result.toString();
104    }
105
106    /**
107     * @see org.opencms.search.I_CmsSearchIndex#addConfigurationParameter(java.lang.String, java.lang.String)
108     */
109    public void addConfigurationParameter(String key, String value) {
110
111        // by default no parameters are excepted
112
113    }
114
115    /**
116     * {@inheritDoc}
117     *
118     * @see org.opencms.search.documents.I_CmsDocumentFactory#createDocument(CmsObject, CmsResource, I_CmsSearchIndex)
119     */
120    public I_CmsSearchDocument createDocument(CmsObject cms, CmsResource resource, I_CmsSearchIndex index)
121    throws CmsException {
122
123        // extract the content from the resource
124        I_CmsExtractionResult content = null;
125
126        if (index.isExtractingContent()) {
127            // do full text content extraction only if required
128
129            // check if caching is enabled for this document type
130            CmsExtractionResultCache cache = getCache();
131            String cacheName = null;
132            if ((cache != null) && (resource.getSiblingCount() > 1)) {
133                // hard drive based caching only makes sense for resources that have siblings,
134                // because the index will also store the content as a blob
135                cacheName = cache.getCacheName(
136                    resource,
137                    isLocaleDependend() ? index.getLocaleForResource(cms, resource, null) : null,
138                    getName());
139                content = cache.getCacheObject(cacheName);
140                if ((content != null) && LOG.isDebugEnabled()) {
141                    LOG.debug("Not re-extracting. Using cached content for '" + resource.getRootPath() + "'.");
142                }
143            }
144
145            if ((content == null) && isOnlyDependentOnContent()) {
146                // extraction result has not been found in the cache
147                // use the currently indexed content, if it is still up to date.
148                content = index.getContentIfUnchanged(resource);
149                if ((content != null) && LOG.isDebugEnabled()) {
150                    LOG.debug(
151                        "Not re-extracting. Using previously indexed content for '" + resource.getRootPath() + "'.");
152                }
153            }
154
155            if (content == null) {
156                // extraction result has not been attached to the resource
157                try {
158                    content = extractContent(cms, resource, index);
159                    if (LOG.isDebugEnabled()) {
160                        LOG.debug("Extracting content for '" + resource.getRootPath() + "' successful.");
161                    }
162                    if ((cache != null) && (resource.getSiblingCount() > 1)) {
163                        // save extracted content to the cache
164                        cache.saveCacheObject(cacheName, content);
165                    }
166                } catch (CmsIndexNoContentException e) {
167                    // there was no content found for the resource
168                    LOG.info(
169                        Messages.get().getBundle().key(Messages.ERR_TEXT_EXTRACTION_1, resource.getRootPath())
170                            + " "
171                            + e.getMessage());
172                } catch (Throwable e) {
173                    // text extraction failed for document - continue indexing meta information only
174                    LOG.error(
175                        Messages.get().getBundle().key(Messages.ERR_TEXT_EXTRACTION_1, resource.getRootPath()),
176                        e);
177                }
178            }
179        }
180
181        // create the Lucene document according to the index field configuration
182        return index.getFieldConfiguration().createDocument(cms, resource, index, content);
183    }
184
185    /**
186     * @see org.opencms.search.documents.I_CmsDocumentFactory#getCache()
187     */
188    public CmsExtractionResultCache getCache() {
189
190        return m_cache;
191    }
192
193    /**
194     * Returns the empty configuration.
195     * Override the method if your document factory is configurable.
196     *
197     * @see org.opencms.search.I_CmsSearchIndex#getConfiguration()
198     */
199    public CmsParameterConfiguration getConfiguration() {
200
201        return new CmsParameterConfiguration();
202    }
203
204    /**
205     * @see org.opencms.search.documents.I_CmsDocumentFactory#getDocumentKeys(java.util.List, java.util.List)
206     */
207    public List<String> getDocumentKeys(List<String> resourceTypes, List<String> mimeTypes) throws CmsException {
208
209        List<String> keys = new ArrayList<String>();
210
211        try {
212            for (Iterator<String> i = resourceTypes.iterator(); i.hasNext();) {
213
214                String typeName = i.next();
215                if (typeName.equals("*")) {
216                    typeName = DEFAULT_ALL_UNCONFIGURED_TYPES;
217                }
218                if (typeName.equals("**")) {
219                    typeName = DEFAULT_ALL_TYPES;
220                }
221                for (Iterator<String> j = mimeTypes.iterator(); j.hasNext();) {
222                    keys.add(getDocumentKey(typeName, j.next()));
223                }
224                if (mimeTypes.isEmpty()) {
225                    keys.add(getDocumentKey(typeName, null));
226                }
227            }
228        } catch (Exception exc) {
229            throw new CmsException(Messages.get().container(Messages.ERR_CREATE_DOC_KEY_0), exc);
230        }
231
232        return keys;
233    }
234
235    /**
236     * @see org.opencms.search.documents.I_CmsDocumentFactory#getName()
237     */
238    public String getName() {
239
240        return m_name;
241    }
242
243    /**
244     * @see org.opencms.configuration.I_CmsConfigurationParameterHandler#initConfiguration()
245     */
246    public void initConfiguration() {
247
248        // Do nothing by default
249
250    }
251
252    /**
253     * @see org.opencms.search.documents.I_CmsDocumentFactory#setCache(org.opencms.search.documents.CmsExtractionResultCache)
254     */
255    public void setCache(CmsExtractionResultCache cache) {
256
257        m_cache = cache;
258    }
259
260    /**
261     * Logs content extraction for the specified resource and index.<p>
262     *
263     * @param resource the resource to log content extraction for
264     * @param index the search index to log content extraction for
265     */
266    protected void logContentExtraction(CmsResource resource, I_CmsSearchIndex index) {
267
268        if (LOG.isDebugEnabled()) {
269            LOG.debug(
270                Messages.get().getBundle().key(
271                    Messages.LOG_EXTRACT_CONTENT_2,
272                    resource.getRootPath(),
273                    index.getName()));
274        }
275    }
276
277    /**
278     * Upgrades the given resource to a {@link CmsFile} with content.<p>
279     *
280     * @param cms the current users OpenCms context
281     * @param resource the resource to upgrade
282     *
283     * @return the given resource upgraded to a {@link CmsFile} with content
284     *
285     * @throws CmsException if the resource could not be read
286     * @throws CmsIndexNoContentException if the resource has no content
287     */
288    protected CmsFile readFile(CmsObject cms, CmsResource resource) throws CmsException, CmsIndexNoContentException {
289
290        CmsFile file = cms.readFile(resource);
291        if (file.getLength() <= 0) {
292            throw new CmsIndexNoContentException(
293                Messages.get().container(Messages.ERR_NO_CONTENT_1, resource.getRootPath()));
294        }
295        return file;
296    }
297}