source: trunk/gli/src/org/greenstone/gatherer/metadata/MetadataXMLFile.java@ 8371

Last change on this file since 8371 was 8371, checked in by mdewsnip, 20 years ago

Added some code for ignoring legacy "hidden" metadata.

  • Property svn:keywords set to Author Date Id Revision
File size: 17.0 KB
Line 
1/**
2 *############################################################################
3 * A component of the Greenstone Librarian Interface, part of the Greenstone
4 * digital library suite from the New Zealand Digital Library Project at the
5 * University of Waikato, New Zealand.
6 *
7 * Author: Michael Dewsnip, NZDL Project, University of Waikato, NZ
8 *
9 * Copyright (C) 2004 New Zealand Digital Library Project
10 *
11 * This program is free software; you can redistribute it and/or modify
12 * it under the terms of the GNU General Public License as published by
13 * the Free Software Foundation; either version 2 of the License, or
14 * (at your option) any later version.
15 *
16 * This program 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
19 * GNU General Public License for more details.
20 *
21 * You should have received a copy of the GNU General Public License
22 * along with this program; if not, write to the Free Software
23 * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
24 *############################################################################
25 */
26
27package org.greenstone.gatherer.metadata;
28
29
30import java.io.*;
31import java.util.*;
32import org.greenstone.gatherer.DebugStream;
33import org.greenstone.gatherer.util.XMLTools;
34import org.w3c.dom.*;
35
36
37/** This class represents one metadata.xml file */
38public class MetadataXMLFile
39 extends File
40{
41 static final private String DESCRIPTION_ELEMENT = "Description";
42 static final private String DIRECTORY_FILENAME = ".*";
43 static final private String FILENAME_ELEMENT = "FileName";
44 static final private String FILESET_ELEMENT = "FileSet";
45 static final private String METADATA_ELEMENT = "Metadata";
46
47
48 public MetadataXMLFile(String metadata_xml_file_path)
49 {
50 super(metadata_xml_file_path);
51 }
52
53
54 public void addMetadata(File file, MetadataValue metadata_value)
55 {
56 // Parse the metadata.xml file
57 Document document = XMLTools.parseXMLFile(this);
58 if (document == null) {
59 System.err.println("Error: Could not parse metadata.xml file " + getAbsolutePath());
60 return;
61 }
62
63 // Determine the file's path relative to the location of the metadata.xml file
64 File metadata_xml_file_directory = getParentFile();
65 String file_relative_path = file.getAbsolutePath().substring(metadata_xml_file_directory.getAbsolutePath().length());
66 if (file_relative_path.startsWith(File.separator)) {
67 file_relative_path = file_relative_path.substring(File.separator.length());
68 }
69
70 // Convert the file path into a regular expression that will match it
71 String file_path_regexp = file_relative_path.replaceAll("\\.", "\\\\.");
72 if (file_relative_path.equals("")) {
73 file_path_regexp = DIRECTORY_FILENAME;
74 }
75
76 // Square brackets need to be escaped because they are a special character in Greenstone
77 String metadata_value_string = metadata_value.getFullValue();
78 metadata_value_string = metadata_value_string.replaceAll("\\[", "[");
79 metadata_value_string = metadata_value_string.replaceAll("\\]", "]");
80
81 // Create a new Metadata element to record this metadata
82 Element new_metadata_value_element = document.createElement(METADATA_ELEMENT);
83 new_metadata_value_element.setAttribute("name", metadata_value.getMetadataElement().getFullName());
84 new_metadata_value_element.setAttribute("mode", (metadata_value.isAccumulatingMetadata() ? "accumulate" : "override"));
85 new_metadata_value_element.appendChild(document.createTextNode(metadata_value_string));
86 boolean have_added_metadata = false;
87
88 // Read all the FileSet elements in the file
89 NodeList fileset_elements_nodelist = document.getElementsByTagName(FILESET_ELEMENT);
90 for (int i = 0; i < fileset_elements_nodelist.getLength(); i++) {
91 Element current_fileset_element = (Element) fileset_elements_nodelist.item(i);
92
93 // Check the FileName elements of the FileSet to see if we have a match
94 NodeList filename_elements_nodelist = current_fileset_element.getElementsByTagName(FILENAME_ELEMENT);
95 for (int j = 0; j < filename_elements_nodelist.getLength(); j++) {
96 Element current_filename_element = (Element) filename_elements_nodelist.item(j);
97 String current_filename_element_value = XMLTools.getElementTextValue(current_filename_element);
98
99 // Only exact matches can be extended with new metadata
100 if (current_filename_element_value.equals(file_path_regexp)) {
101 // Append the new Metadata element to the Description element of this FileSet
102 Element description_element = (Element) current_fileset_element.getElementsByTagName(DESCRIPTION_ELEMENT).item(0);
103
104 // Accumulating metadata: add at the end
105 if (metadata_value.isAccumulatingMetadata()) {
106 description_element.appendChild(new_metadata_value_element);
107 }
108 // Override metadata: add at the start (so it overrides inherited metadata without affecting other assigned metadata)
109 else {
110 description_element.insertBefore(new_metadata_value_element, description_element.getFirstChild());
111 }
112
113 have_added_metadata = true;
114 break;
115 }
116 }
117 }
118
119 // Check if the metadata was added to an existing FileSet
120 if (!have_added_metadata) {
121 // It wasn't, so create a new FileSet element for it
122 Element new_fileset_element = document.createElement(FILESET_ELEMENT);
123
124 Element new_filename_element = document.createElement(FILENAME_ELEMENT);
125 new_filename_element.appendChild(document.createTextNode(file_path_regexp));
126 new_fileset_element.appendChild(new_filename_element);
127
128 // Append the new Metadata element to the Description element of this FileSet
129 Element new_description_element = document.createElement(DESCRIPTION_ELEMENT);
130 new_description_element.appendChild(new_metadata_value_element);
131 new_fileset_element.appendChild(new_description_element);
132
133 document.getDocumentElement().appendChild(new_fileset_element);
134 }
135
136 // Rewrite the metadata.xml file
137 XMLTools.writeXMLFile(this, document);
138 }
139
140
141 public ArrayList getMetadataAssignedToFile(File file)
142 {
143 // Parse the metadata.xml file
144 Document document = XMLTools.parseXMLFile(this);
145 if (document == null) {
146 System.err.println("Error: Could not parse metadata.xml file " + getAbsolutePath());
147 return null;
148 }
149
150 // Determine the file's path relative to the location of the metadata.xml file
151 File metadata_xml_file_directory = getParentFile();
152 String file_relative_path = file.getAbsolutePath().substring(metadata_xml_file_directory.getAbsolutePath().length());
153 if (file_relative_path.startsWith(File.separator)) {
154 file_relative_path = file_relative_path.substring(File.separator.length());
155 }
156
157 // Build up a list of metadata assigned to this file
158 ArrayList metadata_values = new ArrayList();
159
160 // Read all the FileSet elements in the file
161 NodeList fileset_elements_nodelist = document.getElementsByTagName(FILESET_ELEMENT);
162 for (int i = 0; i < fileset_elements_nodelist.getLength(); i++) {
163 Element current_fileset_element = (Element) fileset_elements_nodelist.item(i);
164 boolean current_fileset_matches = false;
165 File folder_metadata_inherited_from = null;
166
167 // Check the FileName elements of the FileSet to see if we have a match
168 NodeList filename_elements_nodelist = current_fileset_element.getElementsByTagName(FILENAME_ELEMENT);
169 for (int j = 0; j < filename_elements_nodelist.getLength(); j++) {
170 Element current_filename_element = (Element) filename_elements_nodelist.item(j);
171 String current_filename_element_value = XMLTools.getElementTextValue(current_filename_element);
172
173 // This fileset specifies metadata for the file
174 if (file_relative_path.matches(current_filename_element_value)) {
175 current_fileset_matches = true;
176 if (!file_relative_path.equals("") && current_filename_element_value.equals(DIRECTORY_FILENAME)) {
177 folder_metadata_inherited_from = metadata_xml_file_directory;
178 }
179 break;
180 }
181
182 // This fileset specifies metadata for the folder the file is in
183 if (file_relative_path.startsWith(current_filename_element_value + File.separator)) {
184 current_fileset_matches = true;
185 folder_metadata_inherited_from = new File(metadata_xml_file_directory, current_filename_element_value);
186 break;
187 }
188 }
189
190 // The FileSet doesn't apply, so move onto the next one
191 if (current_fileset_matches == false) {
192 continue;
193 }
194
195 // Read all the Metadata elements in the fileset
196 NodeList metadata_elements_nodelist = current_fileset_element.getElementsByTagName(METADATA_ELEMENT);
197 for (int k = 0; k < metadata_elements_nodelist.getLength(); k++) {
198 Element current_metadata_element = (Element) metadata_elements_nodelist.item(k);
199 String metadata_element_name_full = current_metadata_element.getAttribute("name");
200 String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
201
202 // Ignore legacy crap
203 if (metadata_set_namespace.equals("hidden")) {
204 continue;
205 }
206
207 MetadataSet metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
208 if (metadata_set == null) {
209 // The metadata set isn't loaded, so give the option of mapping the element into a loaded set
210 String target_metadata_element_name_full = MetadataSetManager.mapUnloadedMetadataElement(metadata_element_name_full);
211 if (target_metadata_element_name_full == null || target_metadata_element_name_full.equals("")) {
212 // Skip this element if we still don't have a loaded element for it
213 continue;
214 }
215
216 metadata_element_name_full = target_metadata_element_name_full;
217 metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
218 metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
219 }
220
221 MetadataElement metadata_element = MetadataTools.getMetadataElementWithName(metadata_element_name_full);
222
223 // If the element doesn't exist in the metadata set, we're not interested
224 if (metadata_element == null) {
225 continue;
226 }
227
228 // Square brackets need to be escaped because they are a special character in Greenstone
229 String metadata_element_value = XMLTools.getElementTextValue(current_metadata_element);
230 metadata_element_value = metadata_element_value.replaceAll("&#091;", "[");
231 metadata_element_value = metadata_element_value.replaceAll("&#093;", "]");
232
233 MetadataValueTreeNode metadata_value_tree_node = metadata_element.getMetadataValueTreeNode(metadata_element_value);
234
235 // If there is no metadata value tree node for this value, create it
236 if (metadata_value_tree_node == null) {
237 DebugStream.println("Note: No value tree node for metadata value \"" + metadata_element_value + "\"");
238 metadata_element.addMetadataValue(metadata_element_value);
239 metadata_value_tree_node = metadata_element.getMetadataValueTreeNode(metadata_element_value);
240 }
241
242 MetadataValue metadata_value = new MetadataValue(metadata_element, metadata_value_tree_node);
243 metadata_value.inheritsMetadataFromFolder(folder_metadata_inherited_from);
244
245 // Is this accumulating metadata?
246 if (current_metadata_element.getAttribute("mode").equals("accumulate")) {
247 metadata_value.setIsAccumulatingMetadata(true);
248 }
249
250 // Add the new metadata value to the list
251 metadata_values.add(metadata_value);
252 }
253 }
254
255 return metadata_values;
256 }
257
258
259 public void removeMetadata(File file, MetadataValue metadata_value)
260 {
261 // Parse the metadata.xml file
262 Document document = XMLTools.parseXMLFile(this);
263 if (document == null) {
264 System.err.println("Error: Could not parse metadata.xml file " + getAbsolutePath());
265 return;
266 }
267
268 // Determine the file's path relative to the location of the metadata.xml file
269 File metadata_xml_file_directory = getParentFile();
270 String file_relative_path = file.getAbsolutePath().substring(metadata_xml_file_directory.getAbsolutePath().length());
271 if (file_relative_path.startsWith(File.separator)) {
272 file_relative_path = file_relative_path.substring(File.separator.length());
273 }
274
275 // Convert the file path into a regular expression that will match it
276 String file_path_regexp = file_relative_path.replaceAll("\\.", "\\\\.");
277 if (file_relative_path.equals("")) {
278 file_path_regexp = DIRECTORY_FILENAME;
279 }
280
281 // Read all the FileSet elements in the file
282 NodeList fileset_elements_nodelist = document.getElementsByTagName(FILESET_ELEMENT);
283 for (int i = 0; i < fileset_elements_nodelist.getLength(); i++) {
284 Element current_fileset_element = (Element) fileset_elements_nodelist.item(i);
285 boolean current_fileset_matches = false;
286
287 // Check the FileName elements of the FileSet to see if we have a match
288 NodeList filename_elements_nodelist = current_fileset_element.getElementsByTagName(FILENAME_ELEMENT);
289 for (int j = 0; j < filename_elements_nodelist.getLength(); j++) {
290 Element current_filename_element = (Element) filename_elements_nodelist.item(j);
291 String current_filename_element_value = XMLTools.getElementTextValue(current_filename_element);
292
293 // Only exact matches can be edited
294 if (current_filename_element_value.equals(file_path_regexp)) {
295 current_fileset_matches = true;
296 break;
297 }
298 }
299
300 // The FileSet doesn't apply, so move onto the next one
301 if (current_fileset_matches == false) {
302 continue;
303 }
304
305 // Find the Metadata element to delete from the fileset
306 String metadata_element_name_full = metadata_value.getMetadataElement().getFullName();
307 String metadata_element_value = metadata_value.getFullValue();
308 NodeList metadata_elements_nodelist = current_fileset_element.getElementsByTagName(METADATA_ELEMENT);
309 for (int k = 0; k < metadata_elements_nodelist.getLength(); k++) {
310 Element current_metadata_element = (Element) metadata_elements_nodelist.item(k);
311
312 // Check the metadata element name matches
313 String current_metadata_element_name_full = current_metadata_element.getAttribute("name");
314 if (!current_metadata_element_name_full.equals(metadata_element_name_full)) {
315 continue;
316 }
317
318 // Check the metadata element value matches
319 String current_metadata_element_value = XMLTools.getElementTextValue(current_metadata_element);
320 if (!current_metadata_element_value.equals(metadata_element_value)) {
321 continue;
322 }
323
324 // Remove this Metadata element
325 current_metadata_element.getParentNode().removeChild(current_metadata_element);
326 }
327 }
328
329 // Rewrite the metadata.xml file
330 XMLTools.writeXMLFile(this, document);
331 }
332
333
334 /**
335 * Every metadata.xml file must be skimmed when a collection is opened, for two reasons:
336 * - To build complete and accurate metadata value trees (needed for Enrich pane and hierarchy files)
337 * - To handle any non-namespaced metadata in the file
338 */
339 public void skimFile()
340 {
341 boolean file_changed = false;
342
343 // Parse the metadata.xml file
344 DebugStream.println("Skimming metadata.xml file " + this + "...");
345 Document document = XMLTools.parseXMLFile(this);
346 if (document == null) {
347 System.err.println("Error: Could not parse metadata.xml file " + getAbsolutePath());
348 return;
349 }
350
351 // Read all the Metadata elements in the file
352 NodeList metadata_elements_nodelist = document.getElementsByTagName(METADATA_ELEMENT);
353 for (int i = 0; i < metadata_elements_nodelist.getLength(); i++) {
354 Element current_metadata_element = (Element) metadata_elements_nodelist.item(i);
355 String metadata_element_name_full = current_metadata_element.getAttribute("name");
356 String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
357
358 // Ignore legacy crap
359 if (metadata_set_namespace.equals("hidden")) {
360 continue;
361 }
362
363 MetadataSet metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
364 if (metadata_set == null) {
365 // The metadata set isn't loaded, so give the option of mapping the element into a loaded set
366 String target_metadata_element_name_full = MetadataSetManager.mapUnloadedMetadataElement(metadata_element_name_full);
367 if (target_metadata_element_name_full == null || target_metadata_element_name_full.equals("")) {
368 // Skip this element if we still don't have a loaded element for it
369 continue;
370 }
371
372 // Update the metadata.xml file to have the new element name
373 current_metadata_element.setAttribute("name", target_metadata_element_name_full);
374 file_changed = true;
375
376 metadata_element_name_full = target_metadata_element_name_full;
377 metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
378 metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
379 }
380
381 String metadata_element_name = MetadataTools.getMetadataElementName(metadata_element_name_full);
382 MetadataElement metadata_element = metadata_set.getMetadataElementWithName(metadata_element_name);
383
384 // If the element doesn't exist in the metadata set, add it
385 if (metadata_element == null) {
386 metadata_element = metadata_set.addMetadataElementForThisSession(metadata_element_name);
387 }
388
389 String metadata_element_value = XMLTools.getElementTextValue(current_metadata_element);
390 metadata_element.addMetadataValue(metadata_element_value);
391 }
392
393 // Rewrite the metadata.xml file if it has changed
394 if (file_changed) {
395 XMLTools.writeXMLFile(this, document);
396 }
397 }
398}
Note: See TracBrowser for help on using the repository browser.