1 |
7512
|
leinfelder
|
/**
|
2 |
|
|
* '$RCSfile$'
|
3 |
|
|
* Copyright: 2000 Regents of the University of California and the
|
4 |
|
|
* National Center for Ecological Analysis and Synthesis
|
5 |
|
|
*
|
6 |
|
|
* '$Author$'
|
7 |
|
|
* '$Date$'
|
8 |
|
|
*
|
9 |
|
|
* This program is free software; you can redistribute it and/or modify
|
10 |
|
|
* it under the terms of the GNU General Public License as published by
|
11 |
|
|
* the Free Software Foundation; either version 2 of the License, or
|
12 |
|
|
* (at your option) any later version.
|
13 |
|
|
*
|
14 |
|
|
* This program is distributed in the hope that it will be useful,
|
15 |
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
16 |
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
17 |
|
|
* GNU General Public License for more details.
|
18 |
|
|
*
|
19 |
|
|
* You should have received a copy of the GNU General Public License
|
20 |
|
|
* along with this program; if not, write to the Free Software
|
21 |
|
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
22 |
|
|
*/
|
23 |
|
|
package edu.ucsb.nceas.metacat.dataone;
|
24 |
|
|
|
25 |
7516
|
leinfelder
|
import java.io.InputStream;
|
26 |
7512
|
leinfelder
|
import java.text.SimpleDateFormat;
|
27 |
7516
|
leinfelder
|
import java.util.Arrays;
|
28 |
7512
|
leinfelder
|
import java.util.HashMap;
|
29 |
|
|
|
30 |
|
|
import org.apache.log4j.Logger;
|
31 |
7622
|
leinfelder
|
import org.apache.wicket.protocol.http.mock.MockHttpServletRequest;
|
32 |
7516
|
leinfelder
|
import org.dataone.client.D1Client;
|
33 |
7512
|
leinfelder
|
import org.dataone.service.exceptions.InvalidRequest;
|
34 |
7516
|
leinfelder
|
import org.dataone.service.exceptions.InvalidToken;
|
35 |
|
|
import org.dataone.service.exceptions.NotAuthorized;
|
36 |
|
|
import org.dataone.service.exceptions.NotFound;
|
37 |
7512
|
leinfelder
|
import org.dataone.service.exceptions.NotImplemented;
|
38 |
|
|
import org.dataone.service.exceptions.ServiceFailure;
|
39 |
|
|
import org.dataone.service.types.v1.Identifier;
|
40 |
|
|
import org.dataone.service.types.v1.Node;
|
41 |
7517
|
leinfelder
|
import org.dataone.service.types.v1.ObjectFormat;
|
42 |
7516
|
leinfelder
|
import org.dataone.service.types.v1.Permission;
|
43 |
|
|
import org.dataone.service.types.v1.Person;
|
44 |
|
|
import org.dataone.service.types.v1.Session;
|
45 |
|
|
import org.dataone.service.types.v1.Subject;
|
46 |
|
|
import org.dataone.service.types.v1.SubjectInfo;
|
47 |
7512
|
leinfelder
|
import org.dataone.service.types.v1.SystemMetadata;
|
48 |
7516
|
leinfelder
|
import org.dataone.service.types.v1.util.AuthUtils;
|
49 |
|
|
import org.dataone.service.util.Constants;
|
50 |
|
|
import org.ecoinformatics.datamanager.parser.DataPackage;
|
51 |
|
|
import org.ecoinformatics.datamanager.parser.generic.DataPackageParserInterface;
|
52 |
|
|
import org.ecoinformatics.datamanager.parser.generic.Eml200DataPackageParser;
|
53 |
7512
|
leinfelder
|
|
54 |
|
|
import edu.ucsb.nceas.ezid.EZIDException;
|
55 |
|
|
import edu.ucsb.nceas.ezid.EZIDService;
|
56 |
|
|
import edu.ucsb.nceas.ezid.profile.DataCiteProfile;
|
57 |
|
|
import edu.ucsb.nceas.ezid.profile.DataCiteProfileResourceTypeValues;
|
58 |
|
|
import edu.ucsb.nceas.ezid.profile.ErcMissingValueCode;
|
59 |
|
|
import edu.ucsb.nceas.ezid.profile.InternalProfile;
|
60 |
|
|
import edu.ucsb.nceas.ezid.profile.InternalProfileValues;
|
61 |
|
|
import edu.ucsb.nceas.metacat.properties.PropertyService;
|
62 |
|
|
import edu.ucsb.nceas.utilities.PropertyNotFoundException;
|
63 |
|
|
|
64 |
|
|
/**
|
65 |
|
|
*
|
66 |
|
|
* Singleton for interacting with the EZID DOI library.
|
67 |
|
|
* Allows DOI minting/initial registration, creating and updating
|
68 |
|
|
* existing DOI registrations.
|
69 |
|
|
*
|
70 |
|
|
* @author leinfelder
|
71 |
|
|
*/
|
72 |
|
|
public class DOIService {
|
73 |
|
|
|
74 |
|
|
private Logger logMetacat = Logger.getLogger(DOIService.class);
|
75 |
|
|
|
76 |
|
|
private static DOIService instance = null;
|
77 |
|
|
|
78 |
|
|
public static DOIService getInstance() {
|
79 |
|
|
if (instance == null) {
|
80 |
|
|
instance = new DOIService();
|
81 |
|
|
}
|
82 |
|
|
return instance;
|
83 |
|
|
}
|
84 |
|
|
|
85 |
|
|
/**
|
86 |
|
|
* Constructor, private for singleton access
|
87 |
|
|
*/
|
88 |
|
|
private DOIService() {
|
89 |
|
|
|
90 |
|
|
}
|
91 |
|
|
|
92 |
|
|
/**
|
93 |
|
|
* submits DOI metadata information about the object to EZID
|
94 |
|
|
* @param sysMeta
|
95 |
|
|
* @return
|
96 |
|
|
* @throws EZIDException
|
97 |
|
|
* @throws ServiceFailure
|
98 |
|
|
* @throws NotImplemented
|
99 |
|
|
*/
|
100 |
|
|
public boolean registerDOI(SystemMetadata sysMeta) throws EZIDException, NotImplemented, ServiceFailure {
|
101 |
|
|
|
102 |
|
|
// for DOIs
|
103 |
|
|
String ezidUsername = null;
|
104 |
|
|
String ezidPassword = null;
|
105 |
|
|
String shoulder = null;
|
106 |
|
|
boolean doiEnabled = false;
|
107 |
|
|
try {
|
108 |
|
|
doiEnabled = new Boolean(PropertyService.getProperty("guid.ezid.enabled")).booleanValue();
|
109 |
|
|
shoulder = PropertyService.getProperty("guid.ezid.doishoulder.1");
|
110 |
|
|
ezidUsername = PropertyService.getProperty("guid.ezid.username");
|
111 |
|
|
ezidPassword = PropertyService.getProperty("guid.ezid.password");
|
112 |
|
|
} catch (PropertyNotFoundException e) {
|
113 |
|
|
logMetacat.warn("DOI support is not configured at this node.", e);
|
114 |
|
|
return false;
|
115 |
|
|
}
|
116 |
|
|
|
117 |
|
|
// only continue if we have the feature turned on
|
118 |
|
|
if (doiEnabled) {
|
119 |
|
|
|
120 |
|
|
String identifier = sysMeta.getIdentifier().getValue();
|
121 |
|
|
|
122 |
|
|
// only continue if this DOI is in our configured shoulder
|
123 |
|
|
if (identifier.startsWith(shoulder)) {
|
124 |
|
|
|
125 |
|
|
// enter metadata about this identifier
|
126 |
|
|
HashMap<String, String> metadata = null;
|
127 |
|
|
|
128 |
|
|
// login to EZID service
|
129 |
|
|
String ezidServiceBaseUrl = null;
|
130 |
|
|
try {
|
131 |
|
|
ezidServiceBaseUrl = PropertyService.getProperty("guid.ezid.baseurl");
|
132 |
|
|
} catch (PropertyNotFoundException e) {
|
133 |
|
|
logMetacat.warn("Using default EZID baseUrl");
|
134 |
|
|
}
|
135 |
|
|
EZIDService ezid = new EZIDService(ezidServiceBaseUrl);
|
136 |
|
|
ezid.login(ezidUsername, ezidPassword);
|
137 |
|
|
|
138 |
|
|
// check for existing metadata
|
139 |
|
|
boolean create = false;
|
140 |
|
|
try {
|
141 |
|
|
metadata = ezid.getMetadata(identifier);
|
142 |
|
|
} catch (EZIDException e) {
|
143 |
|
|
// expected much of the time
|
144 |
|
|
logMetacat.warn("No metadata found for given identifier: " + e.getMessage());
|
145 |
|
|
}
|
146 |
|
|
if (metadata == null) {
|
147 |
|
|
create = true;
|
148 |
|
|
}
|
149 |
|
|
// confuses the API if we send the original metadata that it gave us, so start from scratch
|
150 |
|
|
metadata = new HashMap<String, String>();
|
151 |
|
|
|
152 |
7516
|
leinfelder
|
// title
|
153 |
|
|
String title = ErcMissingValueCode.UNKNOWN.toString();
|
154 |
|
|
try {
|
155 |
|
|
title = lookupTitle(sysMeta);
|
156 |
|
|
} catch (Exception e) {
|
157 |
|
|
e.printStackTrace();
|
158 |
|
|
// ignore
|
159 |
|
|
}
|
160 |
|
|
|
161 |
|
|
// creator
|
162 |
|
|
String creator = sysMeta.getRightsHolder().getValue();
|
163 |
|
|
try {
|
164 |
|
|
creator = lookupCreator(sysMeta.getRightsHolder());
|
165 |
|
|
} catch (Exception e) {
|
166 |
|
|
// ignore and use default
|
167 |
|
|
}
|
168 |
|
|
|
169 |
|
|
// publisher
|
170 |
|
|
String publisher = ErcMissingValueCode.UNKNOWN.toString();
|
171 |
7512
|
leinfelder
|
Node node = MNodeService.getInstance(null).getCapabilities();
|
172 |
7516
|
leinfelder
|
publisher = node.getName();
|
173 |
|
|
|
174 |
|
|
// publication year
|
175 |
7512
|
leinfelder
|
SimpleDateFormat sdf = new SimpleDateFormat("yyyy");
|
176 |
7517
|
leinfelder
|
String year = sdf.format(sysMeta.getDateUploaded());
|
177 |
7516
|
leinfelder
|
|
178 |
|
|
// type
|
179 |
7517
|
leinfelder
|
String resourceType = lookupResourceType(sysMeta);
|
180 |
7516
|
leinfelder
|
|
181 |
7517
|
leinfelder
|
// format
|
182 |
|
|
String format = sysMeta.getFormatId().getValue();
|
183 |
|
|
|
184 |
|
|
//size
|
185 |
|
|
String size = sysMeta.getSize().toString();
|
186 |
|
|
|
187 |
7516
|
leinfelder
|
// target (URL)
|
188 |
|
|
String target = node.getBaseURL() + "/v1/object/" + identifier;
|
189 |
|
|
|
190 |
|
|
// status and export fields for public/protected data
|
191 |
|
|
String status = InternalProfileValues.UNAVAILABLE.toString();
|
192 |
|
|
String export = InternalProfileValues.NO.toString();
|
193 |
|
|
Subject publicSubject = new Subject();
|
194 |
|
|
publicSubject.setValue(Constants.SUBJECT_PUBLIC);
|
195 |
|
|
if (AuthUtils.isAuthorized(Arrays.asList(new Subject[] {publicSubject}), Permission.READ, sysMeta)) {
|
196 |
|
|
status = InternalProfileValues.PUBLIC.toString();
|
197 |
|
|
export = InternalProfileValues.YES.toString();
|
198 |
|
|
}
|
199 |
|
|
|
200 |
7517
|
leinfelder
|
// set the datacite metadata fields
|
201 |
7516
|
leinfelder
|
metadata.put(DataCiteProfile.TITLE.toString(), title);
|
202 |
|
|
metadata.put(DataCiteProfile.CREATOR.toString(), creator);
|
203 |
|
|
metadata.put(DataCiteProfile.PUBLISHER.toString(), publisher);
|
204 |
7512
|
leinfelder
|
metadata.put(DataCiteProfile.PUBLICATION_YEAR.toString(), year);
|
205 |
7516
|
leinfelder
|
metadata.put(DataCiteProfile.RESOURCE_TYPE.toString(), resourceType);
|
206 |
7517
|
leinfelder
|
metadata.put(DataCiteProfile.FORMAT.toString(), format);
|
207 |
|
|
metadata.put(DataCiteProfile.SIZE.toString(), size);
|
208 |
7516
|
leinfelder
|
metadata.put(InternalProfile.TARGET.toString(), target);
|
209 |
|
|
metadata.put(InternalProfile.STATUS.toString(), status);
|
210 |
|
|
metadata.put(InternalProfile.EXPORT.toString(), export);
|
211 |
7512
|
leinfelder
|
|
212 |
|
|
// set using the API
|
213 |
|
|
if (create) {
|
214 |
|
|
ezid.createIdentifier(identifier, metadata);
|
215 |
|
|
} else {
|
216 |
|
|
ezid.setMetadata(identifier, metadata);
|
217 |
|
|
}
|
218 |
|
|
|
219 |
|
|
ezid.logout();
|
220 |
|
|
}
|
221 |
|
|
|
222 |
|
|
}
|
223 |
|
|
|
224 |
|
|
return true;
|
225 |
|
|
}
|
226 |
|
|
|
227 |
|
|
/**
|
228 |
|
|
* Generate a DOI using the EZID service as configured
|
229 |
|
|
* @return
|
230 |
|
|
* @throws EZIDException
|
231 |
|
|
* @throws InvalidRequest
|
232 |
|
|
*/
|
233 |
|
|
public Identifier generateDOI() throws EZIDException, InvalidRequest {
|
234 |
|
|
|
235 |
|
|
Identifier identifier = new Identifier();
|
236 |
|
|
|
237 |
|
|
// look up configuration values
|
238 |
|
|
String shoulder = null;
|
239 |
|
|
String ezidUsername = null;
|
240 |
|
|
String ezidPassword = null;
|
241 |
|
|
boolean doiEnabled = false;
|
242 |
|
|
try {
|
243 |
|
|
doiEnabled = new Boolean(PropertyService.getProperty("guid.ezid.enabled")).booleanValue();
|
244 |
|
|
shoulder = PropertyService.getProperty("guid.ezid.doishoulder.1");
|
245 |
|
|
ezidUsername = PropertyService.getProperty("guid.ezid.username");
|
246 |
|
|
ezidPassword = PropertyService.getProperty("guid.ezid.password");
|
247 |
|
|
} catch (PropertyNotFoundException e1) {
|
248 |
|
|
throw new InvalidRequest("2193", "DOI shoulder is not configured at this node.");
|
249 |
|
|
}
|
250 |
|
|
|
251 |
|
|
// only continue if we have the feature turned on
|
252 |
|
|
if (!doiEnabled) {
|
253 |
|
|
throw new InvalidRequest("2193", "DOI scheme is not enabled at this node.");
|
254 |
|
|
}
|
255 |
|
|
|
256 |
|
|
// add only the minimal metadata required for this DOI
|
257 |
|
|
HashMap<String, String> metadata = new HashMap<String, String>();
|
258 |
|
|
metadata.put(DataCiteProfile.TITLE.toString(), ErcMissingValueCode.UNKNOWN.toString());
|
259 |
|
|
metadata.put(DataCiteProfile.CREATOR.toString(), ErcMissingValueCode.UNKNOWN.toString());
|
260 |
|
|
metadata.put(DataCiteProfile.PUBLISHER.toString(), ErcMissingValueCode.UNKNOWN.toString());
|
261 |
|
|
metadata.put(DataCiteProfile.PUBLICATION_YEAR.toString(), ErcMissingValueCode.UNKNOWN.toString());
|
262 |
|
|
metadata.put(InternalProfile.STATUS.toString(), InternalProfileValues.RESERVED.toString());
|
263 |
|
|
metadata.put(InternalProfile.EXPORT.toString(), InternalProfileValues.NO.toString());
|
264 |
|
|
|
265 |
|
|
// call the EZID service
|
266 |
|
|
String ezidServiceBaseUrl = null;
|
267 |
|
|
try {
|
268 |
|
|
ezidServiceBaseUrl = PropertyService.getProperty("guid.ezid.baseurl");
|
269 |
|
|
} catch (PropertyNotFoundException e) {
|
270 |
|
|
logMetacat.warn("Using default EZID baseUrl");
|
271 |
|
|
}
|
272 |
|
|
EZIDService ezid = new EZIDService(ezidServiceBaseUrl);
|
273 |
|
|
ezid.login(ezidUsername, ezidPassword);
|
274 |
|
|
String doi = ezid.mintIdentifier(shoulder, metadata);
|
275 |
|
|
identifier.setValue(doi);
|
276 |
|
|
ezid.logout();
|
277 |
|
|
|
278 |
|
|
return identifier;
|
279 |
|
|
}
|
280 |
7516
|
leinfelder
|
|
281 |
|
|
/**
|
282 |
|
|
* Locates an appropriate title for the object identified by the given SystemMetadata.
|
283 |
|
|
* Different types of objects will be handled differently for titles:
|
284 |
|
|
* 1. EML formats - parsed by the Datamanager library to find dataset title
|
285 |
|
|
* 2. Data objects - TODO: use title from EML file that describes that data
|
286 |
|
|
* 3. ORE objects - TODO: use title from EML file contained in that package
|
287 |
|
|
* @param sysMeta
|
288 |
|
|
* @return appropriate title if known, or the missing value code
|
289 |
|
|
* @throws Exception
|
290 |
|
|
*/
|
291 |
|
|
private String lookupTitle(SystemMetadata sysMeta) throws Exception {
|
292 |
|
|
String title = ErcMissingValueCode.UNKNOWN.toString();
|
293 |
|
|
if (sysMeta.getFormatId().getValue().startsWith("eml://")) {
|
294 |
|
|
DataPackageParserInterface parser = new Eml200DataPackageParser();
|
295 |
|
|
// for using the MN API as the MN itself
|
296 |
|
|
MockHttpServletRequest request = new MockHttpServletRequest(null, null, null);
|
297 |
|
|
Session session = new Session();
|
298 |
|
|
Subject subject = MNodeService.getInstance(request).getCapabilities().getSubject(0);
|
299 |
|
|
session.setSubject(subject);
|
300 |
|
|
InputStream emlStream = MNodeService.getInstance(request).get(session, sysMeta.getIdentifier());
|
301 |
|
|
parser.parse(emlStream);
|
302 |
|
|
DataPackage dataPackage = parser.getDataPackage();
|
303 |
|
|
title = dataPackage.getTitle();
|
304 |
|
|
}
|
305 |
|
|
return title;
|
306 |
|
|
}
|
307 |
7517
|
leinfelder
|
|
308 |
|
|
private String lookupResourceType(SystemMetadata sysMeta) {
|
309 |
|
|
String resourceType = DataCiteProfileResourceTypeValues.DATASET.toString();
|
310 |
|
|
try {
|
311 |
|
|
ObjectFormat objectFormat = D1Client.getCN().getFormat(sysMeta.getFormatId());
|
312 |
|
|
resourceType += "/" + objectFormat.getFormatType().toLowerCase();
|
313 |
|
|
} catch (Exception e) {
|
314 |
|
|
// ignore
|
315 |
|
|
logMetacat.warn("Could not lookup resource type for formatId" + e.getMessage());
|
316 |
|
|
}
|
317 |
|
|
|
318 |
|
|
return resourceType;
|
319 |
|
|
}
|
320 |
7512
|
leinfelder
|
|
321 |
7516
|
leinfelder
|
/**
|
322 |
|
|
* Lookup the citable name for the given Subject
|
323 |
|
|
* Calls the configured CN to determine this information.
|
324 |
|
|
* If the person is not registered with the CN identity service,
|
325 |
|
|
* a NotFound exception will be raised as expected from the service.
|
326 |
|
|
* @param subject
|
327 |
|
|
* @return fullName if found
|
328 |
|
|
* @throws ServiceFailure
|
329 |
|
|
* @throws NotAuthorized
|
330 |
|
|
* @throws NotImplemented
|
331 |
|
|
* @throws NotFound
|
332 |
|
|
* @throws InvalidToken
|
333 |
|
|
*/
|
334 |
|
|
private String lookupCreator(Subject subject) throws ServiceFailure, NotAuthorized, NotImplemented, NotFound, InvalidToken {
|
335 |
|
|
// default to given DN
|
336 |
|
|
String fullName = subject.getValue();
|
337 |
|
|
|
338 |
|
|
SubjectInfo subjectInfo = D1Client.getCN().getSubjectInfo(subject);
|
339 |
|
|
if (subjectInfo != null && subjectInfo.getPersonList() != null) {
|
340 |
|
|
for (Person p: subjectInfo.getPersonList()) {
|
341 |
|
|
if (p.getSubject().equals(subject)) {
|
342 |
|
|
fullName = p.getFamilyName();
|
343 |
|
|
if (p.getGivenNameList() != null && p.getGivenNameList().size() > 0) {
|
344 |
|
|
fullName = fullName + ", " + p.getGivenName(0);
|
345 |
|
|
}
|
346 |
|
|
break;
|
347 |
|
|
}
|
348 |
|
|
}
|
349 |
|
|
}
|
350 |
|
|
|
351 |
|
|
return fullName;
|
352 |
|
|
|
353 |
|
|
}
|
354 |
7512
|
leinfelder
|
|
355 |
|
|
}
|