source: trunk/gli/src/org/greenstone/gatherer/metadata/MetadataXMLFile.java@ 9525

Last change on this file since 9525 was 9525, checked in by mdewsnip, 19 years ago

Fixed a bug where you couldn't assign or remove metadata from files whose names contained '+' characters (this appears to be valid on Unix... not sure about Windows).

  • Property svn:keywords set to Author Date Id Revision
File size: 19.8 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 // Form a regular expression that specifies the scope of the metadata
71 String file_path_regexp = file_relative_path;
72 if (file_path_regexp.equals("")) {
73 file_path_regexp = DIRECTORY_FILENAME;
74 }
75 else {
76 // Convert the file path into a regular expression that will match it
77 file_path_regexp = file_path_regexp.replaceAll("\\.", "\\\\.");
78 file_path_regexp = file_path_regexp.replaceAll("\\(", "\\\\(");
79 file_path_regexp = file_path_regexp.replaceAll("\\)", "\\\\)");
80 file_path_regexp = file_path_regexp.replaceAll("\\[", "\\\\[");
81 file_path_regexp = file_path_regexp.replaceAll("\\]", "\\\\]");
82 file_path_regexp = file_path_regexp.replaceAll("\\{", "\\\\{");
83 file_path_regexp = file_path_regexp.replaceAll("\\}", "\\\\}");
84 file_path_regexp = file_path_regexp.replaceAll("\\+", "\\\\+");
85 }
86
87 // Square brackets need to be escaped because they are a special character in Greenstone
88 String metadata_value_string = metadata_value.getFullValue();
89 metadata_value_string = metadata_value_string.replaceAll("\\[", "[");
90 metadata_value_string = metadata_value_string.replaceAll("\\]", "]");
91
92 // Create a new Metadata element to record this metadata
93 Element new_metadata_element = document.createElement(METADATA_ELEMENT);
94 new_metadata_element.setAttribute("name", metadata_value.getMetadataElement().getFullName());
95 new_metadata_element.setAttribute("mode", (metadata_value.isAccumulatingMetadata() ? "accumulate" : "override"));
96 new_metadata_element.appendChild(document.createTextNode(metadata_value_string));
97 boolean must_add_new_fileset_for_metadata = true;
98
99 // Read all the FileSet elements in the file
100 String metadata_element_name_full = metadata_value.getMetadataElement().getFullName();
101 NodeList fileset_elements_nodelist = document.getElementsByTagName(FILESET_ELEMENT);
102 for (int i = 0; i < fileset_elements_nodelist.getLength(); i++) {
103 Element current_fileset_element = (Element) fileset_elements_nodelist.item(i);
104
105 // Check the FileName elements of the FileSet to see if we have a match
106 NodeList filename_elements_nodelist = current_fileset_element.getElementsByTagName(FILENAME_ELEMENT);
107 for (int j = 0; j < filename_elements_nodelist.getLength(); j++) {
108 Element current_filename_element = (Element) filename_elements_nodelist.item(j);
109 String current_filename_element_value = XMLTools.getElementTextValue(current_filename_element);
110
111 // Only exact matches can be extended with new metadata
112 if (current_filename_element_value.equals(file_path_regexp)) {
113 Element description_element = (Element) current_fileset_element.getElementsByTagName(DESCRIPTION_ELEMENT).item(0);
114 boolean metadata_already_assigned = false;
115
116 // Check if this piece of metadata has already been assigned to this FileSet element
117 NodeList metadata_elements_nodelist = description_element.getElementsByTagName(METADATA_ELEMENT);
118 for (int k = 0; k < metadata_elements_nodelist.getLength(); k++) {
119 Element current_metadata_element = (Element) metadata_elements_nodelist.item(k);
120
121 // Check the metadata element name matches
122 String current_metadata_element_name_full = current_metadata_element.getAttribute("name");
123 if (!current_metadata_element_name_full.equals(metadata_element_name_full)) {
124 continue;
125 }
126
127 // Check the metadata element value matches
128 String current_metadata_value_string = XMLTools.getElementTextValue(current_metadata_element);
129 if (!current_metadata_value_string.equals(metadata_value_string)) {
130 continue;
131 }
132
133 // Metadata already assigned
134 metadata_already_assigned = true;
135 break;
136 }
137
138 // If the metadata has already been assigned to this FileSet, move onto the next
139 if (metadata_already_assigned) {
140 must_add_new_fileset_for_metadata = false;
141 break;
142 }
143
144 // Accumulating metadata: add at the end
145 if (metadata_value.isAccumulatingMetadata()) {
146 description_element.appendChild(new_metadata_element);
147 }
148 // Override metadata: add at the start (so it overrides inherited metadata without affecting other assigned metadata)
149 else {
150 description_element.insertBefore(new_metadata_element, description_element.getFirstChild());
151 }
152
153 must_add_new_fileset_for_metadata = false;
154 break;
155 }
156 }
157 }
158
159 // If the metadata didn't apply to an existing FileSet, add a new FileSet element for this metadata
160 if (must_add_new_fileset_for_metadata) {
161 Element new_fileset_element = document.createElement(FILESET_ELEMENT);
162
163 Element new_filename_element = document.createElement(FILENAME_ELEMENT);
164 new_filename_element.appendChild(document.createTextNode(file_path_regexp));
165 new_fileset_element.appendChild(new_filename_element);
166
167 // Append the new Metadata element to the Description element of this FileSet
168 Element new_description_element = document.createElement(DESCRIPTION_ELEMENT);
169 new_description_element.appendChild(new_metadata_element);
170 new_fileset_element.appendChild(new_description_element);
171
172 document.getDocumentElement().appendChild(new_fileset_element);
173 }
174
175 // Rewrite the metadata.xml file
176 XMLTools.writeXMLFile(this, document);
177 }
178
179
180 public ArrayList getMetadataAssignedToFile(File file)
181 {
182 // Parse the metadata.xml file
183 Document document = XMLTools.parseXMLFile(this);
184 if (document == null) {
185 System.err.println("Error: Could not parse metadata.xml file " + getAbsolutePath());
186 return new ArrayList();
187 }
188
189 // Determine the file's path relative to the location of the metadata.xml file
190 File metadata_xml_file_directory = getParentFile();
191 String file_relative_path = file.getAbsolutePath().substring(metadata_xml_file_directory.getAbsolutePath().length());
192 if (file_relative_path.startsWith(File.separator)) {
193 file_relative_path = file_relative_path.substring(File.separator.length());
194 }
195
196 // Build up a list of metadata assigned to this file
197 ArrayList metadata_values = new ArrayList();
198
199 // Read all the FileSet elements in the file
200 NodeList fileset_elements_nodelist = document.getElementsByTagName(FILESET_ELEMENT);
201 for (int i = 0; i < fileset_elements_nodelist.getLength(); i++) {
202 Element current_fileset_element = (Element) fileset_elements_nodelist.item(i);
203 boolean current_fileset_matches = false;
204 File folder_metadata_inherited_from = null;
205
206 // Check the FileName elements of the FileSet to see if we have a match
207 NodeList filename_elements_nodelist = current_fileset_element.getElementsByTagName(FILENAME_ELEMENT);
208 for (int j = 0; j < filename_elements_nodelist.getLength(); j++) {
209 Element current_filename_element = (Element) filename_elements_nodelist.item(j);
210 String current_filename_element_value = XMLTools.getElementTextValue(current_filename_element);
211
212 // This fileset specifies metadata for the file
213 if (file_relative_path.matches(current_filename_element_value)) {
214 current_fileset_matches = true;
215 if (!file_relative_path.equals("") && current_filename_element_value.equals(DIRECTORY_FILENAME)) {
216 folder_metadata_inherited_from = metadata_xml_file_directory;
217 }
218 break;
219 }
220
221 // This fileset specifies metadata for the folder the file is in
222 if (file_relative_path.startsWith(current_filename_element_value + File.separator)) {
223 current_fileset_matches = true;
224 folder_metadata_inherited_from = new File(metadata_xml_file_directory, current_filename_element_value);
225 break;
226 }
227 }
228
229 // The FileSet doesn't apply, so move onto the next one
230 if (current_fileset_matches == false) {
231 continue;
232 }
233
234 // Read all the Metadata elements in the fileset
235 NodeList metadata_elements_nodelist = current_fileset_element.getElementsByTagName(METADATA_ELEMENT);
236 for (int k = 0; k < metadata_elements_nodelist.getLength(); k++) {
237 Element current_metadata_element = (Element) metadata_elements_nodelist.item(k);
238 String metadata_element_name_full = current_metadata_element.getAttribute("name");
239 String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
240
241 // Ignore legacy crap
242 if (metadata_set_namespace.equals("hidden")) {
243 continue;
244 }
245
246 MetadataSet metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
247 if (metadata_set == null) {
248 // The metadata set isn't loaded, so give the option of mapping the element into a loaded set
249 String target_metadata_element_name_full = MetadataSetManager.mapUnloadedMetadataElement(metadata_element_name_full);
250 if (target_metadata_element_name_full == null || target_metadata_element_name_full.equals("")) {
251 // Skip this element if we still don't have a loaded element for it
252 continue;
253 }
254
255 metadata_element_name_full = target_metadata_element_name_full;
256 metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
257 metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
258 }
259
260 MetadataElement metadata_element = MetadataTools.getMetadataElementWithName(metadata_element_name_full);
261
262 // If the element doesn't exist in the metadata set, we're not interested
263 if (metadata_element == null) {
264 continue;
265 }
266
267 // Square brackets need to be escaped because they are a special character in Greenstone
268 String metadata_value_string = XMLTools.getElementTextValue(current_metadata_element);
269 metadata_value_string = metadata_value_string.replaceAll("&#091;", "[");
270 metadata_value_string = metadata_value_string.replaceAll("&#093;", "]");
271
272 MetadataValueTreeNode metadata_value_tree_node = metadata_element.getMetadataValueTreeNode(metadata_value_string);
273
274 // If there is no metadata value tree node for this value, create it
275 if (metadata_value_tree_node == null) {
276 DebugStream.println("Note: No value tree node for metadata value \"" + metadata_value_string + "\"");
277 metadata_element.addMetadataValue(metadata_value_string);
278 metadata_value_tree_node = metadata_element.getMetadataValueTreeNode(metadata_value_string);
279 }
280
281 MetadataValue metadata_value = new MetadataValue(metadata_element, metadata_value_tree_node);
282 metadata_value.inheritsMetadataFromFolder(folder_metadata_inherited_from);
283
284 // Is this accumulating metadata?
285 if (current_metadata_element.getAttribute("mode").equals("accumulate")) {
286 metadata_value.setIsAccumulatingMetadata(true);
287 }
288
289 // Add the new metadata value to the list
290 metadata_values.add(metadata_value);
291 }
292 }
293
294 return metadata_values;
295 }
296
297
298 public void removeMetadata(File file, MetadataValue metadata_value)
299 {
300 // Parse the metadata.xml file
301 Document document = XMLTools.parseXMLFile(this);
302 if (document == null) {
303 System.err.println("Error: Could not parse metadata.xml file " + getAbsolutePath());
304 return;
305 }
306
307 // Determine the file's path relative to the location of the metadata.xml file
308 File metadata_xml_file_directory = getParentFile();
309 String file_relative_path = file.getAbsolutePath().substring(metadata_xml_file_directory.getAbsolutePath().length());
310 if (file_relative_path.startsWith(File.separator)) {
311 file_relative_path = file_relative_path.substring(File.separator.length());
312 }
313
314 // Form a regular expression that specifies the scope of the metadata
315 String file_path_regexp = file_relative_path;
316 if (file_path_regexp.equals("")) {
317 file_path_regexp = DIRECTORY_FILENAME;
318 }
319 else {
320 // Convert the file path into a regular expression that will match it
321 file_path_regexp = file_path_regexp.replaceAll("\\.", "\\\\.");
322 file_path_regexp = file_path_regexp.replaceAll("\\(", "\\\\(");
323 file_path_regexp = file_path_regexp.replaceAll("\\)", "\\\\)");
324 file_path_regexp = file_path_regexp.replaceAll("\\[", "\\\\[");
325 file_path_regexp = file_path_regexp.replaceAll("\\]", "\\\\]");
326 file_path_regexp = file_path_regexp.replaceAll("\\{", "\\\\{");
327 file_path_regexp = file_path_regexp.replaceAll("\\}", "\\\\}");
328 file_path_regexp = file_path_regexp.replaceAll("\\+", "\\\\+");
329 }
330
331 // Square brackets need to be escaped because they are a special character in Greenstone
332 String metadata_value_string = metadata_value.getFullValue();
333 metadata_value_string = metadata_value_string.replaceAll("\\[", "&#091;");
334 metadata_value_string = metadata_value_string.replaceAll("\\]", "&#093;");
335
336 // Read all the FileSet elements in the file
337 NodeList fileset_elements_nodelist = document.getElementsByTagName(FILESET_ELEMENT);
338 for (int i = 0; i < fileset_elements_nodelist.getLength(); i++) {
339 Element current_fileset_element = (Element) fileset_elements_nodelist.item(i);
340 boolean current_fileset_matches = false;
341
342 // Check the FileName elements of the FileSet to see if we have a match
343 NodeList filename_elements_nodelist = current_fileset_element.getElementsByTagName(FILENAME_ELEMENT);
344 for (int j = 0; j < filename_elements_nodelist.getLength(); j++) {
345 Element current_filename_element = (Element) filename_elements_nodelist.item(j);
346 String current_filename_element_value = XMLTools.getElementTextValue(current_filename_element);
347
348 // Only exact matches can be edited
349 if (current_filename_element_value.equals(file_path_regexp)) {
350 current_fileset_matches = true;
351 break;
352 }
353 }
354
355 // The FileSet doesn't apply, so move onto the next one
356 if (current_fileset_matches == false) {
357 continue;
358 }
359
360 // Find the Metadata element to delete from the fileset
361 String metadata_element_name_full = metadata_value.getMetadataElement().getFullName();
362 NodeList metadata_elements_nodelist = current_fileset_element.getElementsByTagName(METADATA_ELEMENT);
363 for (int k = 0; k < metadata_elements_nodelist.getLength(); k++) {
364 Element current_metadata_element = (Element) metadata_elements_nodelist.item(k);
365
366 // Check the metadata element name matches
367 String current_metadata_element_name_full = current_metadata_element.getAttribute("name");
368 if (!current_metadata_element_name_full.equals(metadata_element_name_full)) {
369 continue;
370 }
371
372 // Check the metadata element value matches
373 String current_metadata_value_string = XMLTools.getElementTextValue(current_metadata_element);
374 if (!current_metadata_value_string.equals(metadata_value_string)) {
375 continue;
376 }
377
378 // Remove this Metadata element
379 current_metadata_element.getParentNode().removeChild(current_metadata_element);
380 }
381 }
382
383 // Rewrite the metadata.xml file
384 XMLTools.writeXMLFile(this, document);
385 }
386
387
388 /**
389 * Every metadata.xml file must be skimmed when a collection is opened, for two reasons:
390 * - To build complete and accurate metadata value trees (needed for Enrich pane and hierarchy files)
391 * - To handle any non-namespaced metadata in the file
392 */
393 public void skimFile()
394 {
395 boolean file_changed = false;
396
397 // Parse the metadata.xml file
398 DebugStream.println("Skimming metadata.xml file " + this + "...");
399 Document document = XMLTools.parseXMLFile(this);
400 if (document == null) {
401 System.err.println("Error: Could not parse metadata.xml file " + getAbsolutePath());
402 return;
403 }
404
405 // Read all the Metadata elements in the file
406 NodeList metadata_elements_nodelist = document.getElementsByTagName(METADATA_ELEMENT);
407 for (int i = 0; i < metadata_elements_nodelist.getLength(); i++) {
408 Element current_metadata_element = (Element) metadata_elements_nodelist.item(i);
409 String metadata_element_name_full = current_metadata_element.getAttribute("name");
410 String metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
411
412 // Ignore legacy crap
413 if (metadata_set_namespace.equals("hidden")) {
414 continue;
415 }
416
417 MetadataSet metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
418 if (metadata_set == null) {
419 // The metadata set isn't loaded, so give the option of mapping the element into a loaded set
420 String target_metadata_element_name_full = MetadataSetManager.mapUnloadedMetadataElement(metadata_element_name_full);
421 if (target_metadata_element_name_full == null || target_metadata_element_name_full.equals("")) {
422 // Skip this element if we still don't have a loaded element for it
423 continue;
424 }
425
426 // Update the metadata.xml file to have the new element name
427 current_metadata_element.setAttribute("name", target_metadata_element_name_full);
428 file_changed = true;
429
430 metadata_element_name_full = target_metadata_element_name_full;
431 metadata_set_namespace = MetadataTools.getMetadataSetNamespace(metadata_element_name_full);
432 metadata_set = MetadataSetManager.getMetadataSet(metadata_set_namespace);
433 }
434
435 String metadata_element_name = MetadataTools.getMetadataElementName(metadata_element_name_full);
436 MetadataElement metadata_element = metadata_set.getMetadataElementWithName(metadata_element_name);
437
438 // If the element doesn't exist in the metadata set, add it
439 if (metadata_element == null) {
440 metadata_element = metadata_set.addMetadataElementForThisSession(metadata_element_name);
441 }
442
443 // Square brackets need to be escaped because they are a special character in Greenstone
444 String metadata_value_string = XMLTools.getElementTextValue(current_metadata_element);
445 metadata_value_string = metadata_value_string.replaceAll("&#091;", "[");
446 metadata_value_string = metadata_value_string.replaceAll("&#093;", "]");
447
448 metadata_element.addMetadataValue(metadata_value_string);
449 }
450
451 // Rewrite the metadata.xml file if it has changed
452 if (file_changed) {
453 XMLTools.writeXMLFile(this, document);
454 }
455 }
456}
Note: See TracBrowser for help on using the repository browser.