A_CmsVfsDocument


1   /*
2    * File   : $Source: /usr/local/cvs/opencms/src/org/opencms/search/documents/A_CmsVfsDocument.java,v $
3    * Date   : $Date: 2006/03/27 14:53:05 $
4    * Version: $Revision: 1.14 $
5    *
6    * This library is part of OpenCms -
7    * the Open Source Content Mananagement System
8    *
9    * Copyright (c) 2005 Alkacon Software GmbH (http://www.alkacon.com)
10   *
11   * This library is free software; you can redistribute it and/or
12   * modify it under the terms of the GNU Lesser General Public
13   * License as published by the Free Software Foundation; either
14   * version 2.1 of the License, or (at your option) any later version.
15   *
16   * This library is distributed in the hope that it will be useful,
17   * but WITHOUT ANY WARRANTY; without even the implied warranty of
18   * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19   * Lesser General Public License for more details.
20   *
21   * For further information about Alkacon Software GmbH, please see the
22   * company website: http://www.alkacon.com
23   *
24   * For further information about OpenCms, please see the
25   * project website: http://www.opencms.org
26   * 
27   * You should have received a copy of the GNU Lesser General Public
28   * License along with this library; if not, write to the Free Software
29   * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
30   */
31  
32  package org.opencms.search.documents;
33  
34  import org.opencms.file.CmsFile;
35  import org.opencms.file.CmsObject;
36  import org.opencms.file.CmsPropertyDefinition;
37  import org.opencms.file.CmsResource;
38  import org.opencms.file.types.I_CmsResourceType;
39  import org.opencms.main.CmsException;
40  import org.opencms.main.CmsLog;
41  import org.opencms.main.OpenCms;
42  import org.opencms.search.A_CmsIndexResource;
43  import org.opencms.search.CmsIndexException;
44  import org.opencms.search.CmsSearchCategoryCollector;
45  import org.opencms.search.CmsSearchIndex;
46  import org.opencms.search.extractors.I_CmsExtractionResult;
47  import org.opencms.util.CmsStringUtil;
48  
49  import java.util.ArrayList  ;
50  import java.util.Date  ;
51  import java.util.Iterator  ;
52  import java.util.List  ;
53  import java.util.Map  ;
54  
55  import org.apache.commons.logging.Log;
56  import org.apache.lucene.document.DateTools;
57  import org.apache.lucene.document.Document;
58  import org.apache.lucene.document.Field;
59  
60  /**
61   * Base document factory class for a VFS <code>{@link org.opencms.file.CmsResource}</code>, 
62   * just requires a specialized implementation of 
63   * <code>{@link I_CmsDocumentFactory#extractContent(CmsObject, A_CmsIndexResource, String)}</code>
64   * for text extraction from the binary document content.<p>
65   * 
66   * @author Carsten Weinholz 
67   * @author Alexander Kandzior 
68   * 
69   * @version $Revision: 1.14 $ 
70   * 
71   * @since 6.0.0 
72   */
73  public abstract class A_CmsVfsDocument implements I_CmsDocumentFactory {
74  
75      /** The vfs prefix for document keys. */
76      public static final String   VFS_DOCUMENT_KEY_PREFIX = "VFS";
77  
78      /** The log object for this class. */
79      private static final Log LOG = CmsLog.getLog(A_CmsVfsDocument.class);
80  
81      /**
82       * Name of the documenttype.
83       */
84      protected String   m_name;
85  
86      /**
87       * Creates a new instance of this lucene document factory.<p>
88       * 
89       * @param name name of the documenttype
90       */
91      public A_CmsVfsDocument(String   name) {
92  
93          m_name = name;
94      }
95  
96      /**
97       * @see org.opencms.search.documents.I_CmsDocumentFactory#getDocumentKey(java.lang.String)
98       */
99      public String   getDocumentKey(String   resourceType) throws CmsIndexException {
100 
101         try {
102             return VFS_DOCUMENT_KEY_PREFIX + ((I_CmsResourceType)Class.forName(resourceType).newInstance()).getTypeId();
103         } catch (Exception   exc) {
104             throw new CmsIndexException(Messages.get().container(
105                 Messages.ERR_RESOURCE_TYPE_INSTANTIATION_1,
106                 resourceType), exc);
107         }
108     }
109 
110     /**
111      * @see org.opencms.search.documents.I_CmsDocumentFactory#getDocumentKeys(java.util.List, java.util.List)
112      */
113     public List   getDocumentKeys(List   resourceTypes, List   mimeTypes) throws CmsException {
114 
115         ArrayList   keys = new ArrayList  ();
116 
117         if (resourceTypes.contains("*")) {
118             ArrayList   allTypes = new ArrayList  ();
119             for (Iterator   i = OpenCms.getResourceManager().getResourceTypes().iterator(); i.hasNext();) {
120                 I_CmsResourceType resourceType = (I_CmsResourceType)i.next();
121                 allTypes.add(resourceType.getTypeName());
122             }
123             resourceTypes = allTypes;
124         }
125 
126         try {
127             for (Iterator   i = resourceTypes.iterator(); i.hasNext();) {
128 
129                 int id = OpenCms.getResourceManager().getResourceType((String  )i.next()).getTypeId();
130                 for (Iterator   j = mimeTypes.iterator(); j.hasNext();) {
131                     keys.add(VFS_DOCUMENT_KEY_PREFIX + id + ":" + (String  )j.next());
132                 }
133                 if (mimeTypes.isEmpty()) {
134                     keys.add(VFS_DOCUMENT_KEY_PREFIX + id);
135                 }
136             }
137         } catch (Exception   exc) {
138             throw new CmsException(Messages.get().container(Messages.ERR_CREATE_DOC_KEY_0), exc);
139         }
140 
141         return keys;
142     }
143 
144     /**
145      * @see org.opencms.search.documents.I_CmsDocumentFactory#getName()
146      */
147     public String   getName() {
148 
149         return m_name;
150     }
151 
152     /**
153      * Generates a new lucene document instance from contents of the given resource.<p>
154      * 
155      * @see org.opencms.search.documents.I_CmsDocumentFactory#newInstance(org.opencms.file.CmsObject, org.opencms.search.A_CmsIndexResource, java.lang.String)
156      */
157     public Document newInstance(CmsObject cms, A_CmsIndexResource resource, String   language) throws CmsException {
158 
159         Document document = new Document();
160         CmsResource res = (CmsResource)resource.getData();
161         String   path = cms.getRequestContext().removeSiteRoot(resource.getRootPath());
162 
163         // extract the content from the resource
164         String   text = null;
165         try {
166             I_CmsExtractionResult content = extractContent(cms, resource, language);
167             text = mergeMetaInfo(content);
168             content.release();
169         } catch (Exception   e) {
170             // text extraction failed for document - continue indexing meta information only
171             LOG.error(Messages.get().getBundle().key(Messages.ERR_TEXT_EXTRACTION_1, resource.getRootPath()), e);
172         }
173         if (text != null) {
174             document.add(new Field(I_CmsDocumentFactory.DOC_CONTENT, text, Field.Store.YES, Field.Index.TOKENIZED));
175         }
176 
177         StringBuffer   meta = new StringBuffer  (512);
178         String   value;
179         Field field;
180 
181         // add the title from the property
182         value = cms.readPropertyObject(path, CmsPropertyDefinition.PROPERTY_TITLE, false).getValue();
183         if (CmsStringUtil.isNotEmpty(value)) {
184             value = value.trim();
185             if (value.length() > 0) {
186                 // add title as keyword, required for sorting
187                 field = new Field(I_CmsDocumentFactory.DOC_TITLE_KEY, value, Field.Store.YES, Field.Index.UN_TOKENIZED);
188                 // title keyword field should not affect the boost factor
189                 field.setBoost(0);
190                 document.add(field);
191                 // add title again as indexed field for searching
192                 document.add(new Field(
193                     I_CmsDocumentFactory.DOC_TITLE_INDEXED,
194                     value,
195                     Field.Store.NO,
196                     Field.Index.TOKENIZED));
197                 meta.append(value);
198                 meta.append(" ");
199             }
200         }
201         // add the keywords from the property
202         value = cms.readPropertyObject(path, CmsPropertyDefinition.PROPERTY_KEYWORDS, false).getValue();
203         if (CmsStringUtil.isNotEmptyOrWhitespaceOnly(value)) {
204             document.add(new Field(I_CmsDocumentFactory.DOC_KEYWORDS, value, Field.Store.YES, Field.Index.TOKENIZED));
205             meta.append(value);
206             meta.append(" ");
207         }
208         // add the description from the property
209         value = cms.readPropertyObject(path, CmsPropertyDefinition.PROPERTY_DESCRIPTION, false).getValue();
210         if (CmsStringUtil.isNotEmptyOrWhitespaceOnly(value)) {
211             document.add(new Field(I_CmsDocumentFactory.DOC_DESCRIPTION, value, Field.Store.YES, Field.Index.TOKENIZED));
212             meta.append(value);
213             meta.append(" ");
214         }
215         // add the collected meta information
216         String   metaInf = meta.toString();
217         if (CmsStringUtil.isNotEmptyOrWhitespaceOnly(metaInf)) {
218             document.add(new Field(I_CmsDocumentFactory.DOC_META, metaInf, Field.Store.NO, Field.Index.TOKENIZED));
219         }
220 
221         // add the category of the file (this is searched so the value can also be attached on a folder)
222         value = cms.readPropertyObject(path, CmsPropertyDefinition.PROPERTY_SEARCH_CATEGORY, true).getValue();
223         if (CmsStringUtil.isNotEmpty(value)) {
224             // all categorys are internally stored lower case
225             value = value.trim().toLowerCase();
226             if (value.length() > 0) {
227                 field = new Field(I_CmsDocumentFactory.DOC_CATEGORY, value, Field.Store.YES, Field.Index.UN_TOKENIZED);
228                 field.setBoost(0);
229                 document.add(field);
230             }
231         } else {
232             // synthetic "unknown" category if no category property defined for resource
233             field = new Field(
234                 I_CmsDocumentFactory.DOC_CATEGORY,
235                 CmsSearchCategoryCollector.UNKNOWN_CATEGORY,
236                 Field.Store.YES,
237                 Field.Index.UN_TOKENIZED);
238             document.add(field);
239         }
240 
241         // add the document root path, optimized for use with a phrase query
242         String   rootPath = CmsSearchIndex.rootPathRewrite(resource.getRootPath());
243         field = new Field(I_CmsDocumentFactory.DOC_ROOT, rootPath, Field.Store.YES, Field.Index.TOKENIZED);
244         // set boost of 0 to root path field, since root path should have no effect on search result score 
245         field.setBoost(0);
246         document.add(field);
247         // root path is stored again in "plain" format, but not for indexing since I_CmsDocumentFactory.DOC_ROOT is used for that
248         // must be indexed as a keyword ONLY to be able to use this when deleting a resource from the index
249         document.add(new Field(
250             I_CmsDocumentFactory.DOC_PATH,
251             resource.getRootPath(),
252             Field.Store.YES,
253             Field.Index.UN_TOKENIZED));
254 
255         // add date of creation and last modification as keywords (for sorting)
256         field = new Field(I_CmsDocumentFactory.DOC_DATE_CREATED, DateTools.dateToString(
257             new Date  (res.getDateCreated()),
258             DateTools.Resolution.MILLISECOND), Field.Store.YES, Field.Index.UN_TOKENIZED);
259         field.setBoost(0);
260         document.add(field);
261         field = new Field(I_CmsDocumentFactory.DOC_DATE_LASTMODIFIED, DateTools.dateToString(new Date  (
262             res.getDateLastModified()), DateTools.Resolution.MILLISECOND), Field.Store.YES, Field.Index.UN_TOKENIZED);
263         field.setBoost(0);
264         document.add(field);
265 
266         // special field for VFS documents - add a marker so that the document can be identified as VFS resource
267         document.add(new Field(I_CmsDocumentFactory.DOC_TYPE, VFS_DOCUMENT_KEY_PREFIX, Field.Store.YES, Field.Index.NO));
268 
269         float boost = 1.0f;
270         // note that the priority property IS searched, so you can easily flag whole folders as "high" or "low"
271         value = cms.readPropertyObject(path, CmsPropertyDefinition.PROPERTY_SEARCH_PRIORITY, true).getValue();
272         if (value != null) {
273             value = value.trim().toLowerCase();
274             if (value.equals(I_CmsDocumentFactory.SEARCH_PRIORITY_MAX_VALUE)) {
275                 boost = 2.0f;
276             } else if (value.equals(I_CmsDocumentFactory.SEARCH_PRIORITY_HIGH_VALUE)) {
277                 boost = 1.5f;
278             } else if (value.equals(I_CmsDocumentFactory.SEARCH_PRIORITY_LOW_VALUE)) {
279                 boost = 0.5f;
280             }
281         }
282         // set document boost factor
283         document.setBoost(boost);
284 
285         return document;
286     }
287 
288     /**
289      * Returns a String created out of the content and the most important meta information in the given 
290      * extraction result.<p>
291      * 
292      * OpenCms uses it's own properties for the text "Title" etc. field, this method ensures
293      * the most important document meta information can still be found as part of the content.<p> 
294      * 
295      * @param extractedContent the extraction result to merge
296      * 
297      * @return a String created out of the most important meta information in the given map and the content
298      */
299     protected String   mergeMetaInfo(I_CmsExtractionResult extractedContent) {
300 
301         Map   metaInfo = extractedContent.getMetaInfo();
302         String   content = extractedContent.getContent();
303 
304         if (((metaInfo == null) || (metaInfo.size() == 0)) && (CmsStringUtil.isEmpty(content))) {
305             return null;
306         }
307 
308         StringBuffer   result = new StringBuffer  (4096);
309         if (metaInfo != null) {
310             String   meta;
311             meta = (String  )metaInfo.get(I_CmsExtractionResult.META_TITLE);
312             if (CmsStringUtil.isNotEmpty(meta)) {
313                 result.append(meta);
314                 result.append('\n');
315             }
316             meta = (String  )metaInfo.get(I_CmsExtractionResult.META_SUBJECT);
317             if (CmsStringUtil.isNotEmpty(meta)) {
318                 result.append(meta);
319                 result.append('\n');
320             }
321             meta = (String  )metaInfo.get(I_CmsExtractionResult.META_KEYWORDS);
322             if (CmsStringUtil.isNotEmpty(meta)) {
323                 result.append(meta);
324                 result.append('\n');
325             }
326             meta = (String  )metaInfo.get(I_CmsExtractionResult.META_COMMENTS);
327             if (CmsStringUtil.isNotEmpty(meta)) {
328                 result.append(meta);
329                 result.append('\n');
330             }
331         }
332 
333         if (content != null) {
334             result.append(content);
335         }
336 
337         return result.toString();
338     }
339 
340     /**
341      * Upgrades the given resource to a {@link CmsFile} with content.<p>
342      * 
343      * @param cms the current users OpenCms context
344      * @param resource the resource to upgrade
345      * 
346      * @return the given resource upgraded to a {@link CmsFile} with content
347      * 
348      * @throws CmsException if the resource could not be read 
349      * @throws CmsIndexException if the resource has no content
350      */
351     protected CmsFile readFile(CmsObject cms, CmsResource resource) throws CmsException, CmsIndexException {
352 
353         CmsFile file = CmsFile.upgrade(resource, cms);
354         if (file.getLength() <= 0) {
355             throw new CmsIndexException(Messages.get().container(Messages.ERR_NO_CONTENT_1, resource.getRootPath()));
356         }
357         return file;
358     }
359 }
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags