1
|
/**
|
2
|
* '$RCSfile$'
|
3
|
* Copyright: 2000-2005 Regents of the University of California and the
|
4
|
* National Center for Ecological Analysis and Synthesis
|
5
|
*
|
6
|
* '$Author: leinfelder $'
|
7
|
* '$Date: 2010-12-27 14:47:01 -0800 (Mon, 27 Dec 2010) $'
|
8
|
* '$Revision: 5760 $'
|
9
|
*
|
10
|
* This program is free software; you can redistribute it and/or modify
|
11
|
* it under the terms of the GNU General Public License as published by
|
12
|
* the Free Software Foundation; either version 2 of the License, or
|
13
|
* (at your option) any later version.
|
14
|
*
|
15
|
* This program is distributed in the hope that it will be useful,
|
16
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
17
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
18
|
* GNU General Public License for more details.
|
19
|
*
|
20
|
* You should have received a copy of the GNU General Public License
|
21
|
* along with this program; if not, write to the Free Software
|
22
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
23
|
*/
|
24
|
|
25
|
package edu.ucsb.nceas.metacat.lsid;
|
26
|
|
27
|
import java.io.BufferedReader;
|
28
|
import java.io.ByteArrayInputStream;
|
29
|
import java.io.ByteArrayOutputStream;
|
30
|
import java.io.IOException;
|
31
|
import java.io.InputStream;
|
32
|
import java.io.InputStreamReader;
|
33
|
import java.io.Reader;
|
34
|
import java.io.StringReader;
|
35
|
import java.io.UnsupportedEncodingException;
|
36
|
import java.util.Hashtable;
|
37
|
import java.util.ResourceBundle;
|
38
|
import java.util.Enumeration;
|
39
|
|
40
|
import javax.xml.transform.Transformer;
|
41
|
import javax.xml.transform.TransformerFactory;
|
42
|
import javax.xml.transform.stream.StreamResult;
|
43
|
import javax.xml.transform.stream.StreamSource;
|
44
|
|
45
|
import org.apache.commons.logging.Log;
|
46
|
import org.apache.commons.logging.LogFactory;
|
47
|
|
48
|
import com.ibm.lsid.LSID;
|
49
|
import com.ibm.lsid.MalformedLSIDException;
|
50
|
import com.ibm.lsid.MetadataResponse;
|
51
|
import com.ibm.lsid.server.LSIDMetadataService;
|
52
|
import com.ibm.lsid.server.LSIDRequestContext;
|
53
|
import com.ibm.lsid.server.LSIDServerException;
|
54
|
import com.ibm.lsid.server.LSIDServiceConfig;
|
55
|
|
56
|
import edu.ucsb.nceas.metacat.MetaCatServlet;
|
57
|
import edu.ucsb.nceas.metacat.client.Metacat;
|
58
|
import edu.ucsb.nceas.metacat.client.MetacatFactory;
|
59
|
import edu.ucsb.nceas.metacat.client.MetacatInaccessibleException;
|
60
|
|
61
|
public class LSIDAuthorityMetaData implements LSIDMetadataService
|
62
|
{
|
63
|
private LSIDDataLookup lookup = null;
|
64
|
private static Hashtable currentLSIDs = new Hashtable();
|
65
|
private static Log logger = LogFactory
|
66
|
.getLog("edu.ucsb.nceas.metacat.lsid");
|
67
|
|
68
|
public void initService(LSIDServiceConfig cf) throws LSIDServerException
|
69
|
{
|
70
|
logger.info("Starting LSIDAuthorityMetadata.");
|
71
|
lookup = new LSIDDataLookup();
|
72
|
}
|
73
|
|
74
|
private static final String RDF_NS = "http://www.w3.org/1999/02/22-rdf-syntax-ns#";
|
75
|
private static final String DC_NS = "http://purl.org/dc/elements/1.1/";
|
76
|
private static final String I3CP_NS = "urn:lsid:i3c.org:predicates:";
|
77
|
private static final String I3C_CONTENT = "urn:lsid:i3c.org:types:content";
|
78
|
private static final String DEFAULT_STYLESHEET = "default.xslt";
|
79
|
|
80
|
public MetadataResponse getMetadata(LSIDRequestContext req, String[] formats)
|
81
|
throws LSIDServerException
|
82
|
{
|
83
|
LSID lsid = req.getLsid();
|
84
|
ByteArrayInputStream theMetadata = doMetadataRequest(lsid);
|
85
|
return new MetadataResponse(theMetadata, null,
|
86
|
MetadataResponse.RDF_FORMAT);
|
87
|
}
|
88
|
|
89
|
private ByteArrayInputStream doMetadataRequest(LSID lsid) throws
|
90
|
LSIDServerException {
|
91
|
logger.debug("getting metadata for lsid " + lsid.getLsid());
|
92
|
|
93
|
try {
|
94
|
|
95
|
LSIDDataLookup myLookup = new LSIDDataLookup();
|
96
|
InputStream content = myLookup.lsidData(lsid);
|
97
|
InputStream content2 = myLookup.lsidData(lsid);
|
98
|
if (!isEML(content2)) {
|
99
|
content = getEML(lsid);
|
100
|
}
|
101
|
content2.close();
|
102
|
setCurrentLSID(lsid);
|
103
|
|
104
|
// based on the doctype choose the appropriate stylesheet
|
105
|
String styleSheetName = null;
|
106
|
String _docType = myLookup.getDocType(lsid);
|
107
|
|
108
|
try {
|
109
|
ResourceBundle rb = ResourceBundle.getBundle("metacat-lsid");
|
110
|
styleSheetName = rb.getString(_docType.replaceAll(":", ""));
|
111
|
}
|
112
|
catch (java.util.MissingResourceException mre) {
|
113
|
logger.warn("there is no style corresponding to: '" + _docType +
|
114
|
"' -- using default");
|
115
|
styleSheetName = this.DEFAULT_STYLESHEET;
|
116
|
mre.getMessage();
|
117
|
}
|
118
|
InputStream styleSheet = getClass()
|
119
|
.getResourceAsStream(styleSheetName);
|
120
|
TransformerFactory factory = TransformerFactory.newInstance();
|
121
|
Transformer transformer = factory
|
122
|
.newTransformer(new StreamSource(styleSheet));
|
123
|
ByteArrayOutputStream out = new ByteArrayOutputStream();
|
124
|
transformer.transform(new StreamSource(content),
|
125
|
new StreamResult(out));
|
126
|
content.close();
|
127
|
clearState();
|
128
|
return new ByteArrayInputStream(out.toByteArray());
|
129
|
}
|
130
|
catch (Exception e) {
|
131
|
throw new LSIDServerException(e, "Error transforming XML for: "
|
132
|
+ lsid);
|
133
|
}
|
134
|
}
|
135
|
|
136
|
public static String getStringFromInputStream(InputStream input)
|
137
|
{
|
138
|
StringBuffer result = new StringBuffer();
|
139
|
BufferedReader in = new BufferedReader(new InputStreamReader(input));
|
140
|
String line;
|
141
|
try {
|
142
|
while ((line = in.readLine()) != null) {
|
143
|
result.append(line);
|
144
|
}
|
145
|
} catch (IOException e) {
|
146
|
logger.error("IOexception " + e);
|
147
|
}
|
148
|
return result.toString();
|
149
|
}
|
150
|
|
151
|
/**
|
152
|
* figure out is this inputstream is an eml document or not
|
153
|
* TODO: need a better way to figure out if this is an eml document
|
154
|
*/
|
155
|
private boolean isEML(InputStream input)
|
156
|
{
|
157
|
|
158
|
if (input == null) { return false; }
|
159
|
|
160
|
int loop = 0;
|
161
|
boolean itIsEML = false;
|
162
|
String line = "";
|
163
|
try {
|
164
|
BufferedReader in = new BufferedReader(new InputStreamReader(input));
|
165
|
while ((loop < 20) && (line != null) && (!itIsEML)) {
|
166
|
line = in.readLine();
|
167
|
line = line.toLowerCase();
|
168
|
if (line.indexOf("eml:eml") != -1) {
|
169
|
itIsEML = true;
|
170
|
}
|
171
|
loop++;
|
172
|
}
|
173
|
} catch (IOException e) {
|
174
|
logger.error("ioerror in LSIDAuthorityMetadata: " + e);
|
175
|
}
|
176
|
return itIsEML;
|
177
|
}
|
178
|
|
179
|
/**
|
180
|
* this lsid points to a data object - get the metadata objects which refer
|
181
|
* to this document
|
182
|
*/
|
183
|
private InputStream getEML(LSID theLSID)
|
184
|
{
|
185
|
|
186
|
InputStream response = null;
|
187
|
|
188
|
// need to find things with this object in any of the elements
|
189
|
// using the metacat api
|
190
|
// get back dataset/docid and dataset/title
|
191
|
|
192
|
// create the query
|
193
|
String theQuery = getMetaCatQuery(theLSID);
|
194
|
|
195
|
// get the metacat record
|
196
|
Reader metaCatResponse = getMetaCatResponse(theQuery);
|
197
|
|
198
|
// parse the metadata to get the applicable rdf information
|
199
|
response = parseMetaCatResponse(metaCatResponse, theLSID);
|
200
|
|
201
|
return response;
|
202
|
|
203
|
}
|
204
|
|
205
|
/**
|
206
|
* given an LSID return a metacat query which will return docs mentioning
|
207
|
* this LSID
|
208
|
*/
|
209
|
private String getMetaCatQuery(LSID lsid)
|
210
|
{
|
211
|
logger.debug("getting Metacat Query for: " + lsid.toString());
|
212
|
String ns = lsid.getNamespace();
|
213
|
String id = lsid.getObject();
|
214
|
String ver = lsid.getRevision();
|
215
|
String theName = ns + "." + id + "." + ver;
|
216
|
|
217
|
String theQuery = null;
|
218
|
theQuery = "<?xml version=\"1.0\"?>\n"
|
219
|
+ "<pathquery version=\"1.2\">\n"
|
220
|
+ " <querytitle>"
|
221
|
+ theName
|
222
|
+ " search</querytitle>\n"
|
223
|
+ " <returnfield>dataset/docid</returnfield>\n"
|
224
|
+ " <returnfield>dataset/title</returnfield>\n"
|
225
|
+ " <querygroup operator=\"UNION\">\n"
|
226
|
+ " <queryterm searchmode=\"contains\" casesensitive=\"false\">\n"
|
227
|
+ " <value>" + theName + "</value>\n"
|
228
|
+ " <pathexpr>anyfield</pathexpr>\n"
|
229
|
+ " </queryterm>\n" + " </querygroup>\n"
|
230
|
+ "<pathquery>\n";
|
231
|
|
232
|
return theQuery;
|
233
|
|
234
|
}
|
235
|
|
236
|
/**
|
237
|
* given a query string, query MetaCat and return the response
|
238
|
*/
|
239
|
private Reader getMetaCatResponse(String query)
|
240
|
{
|
241
|
logger.debug("Querying the metacat server.");
|
242
|
// get the metacat server from the configuration file
|
243
|
//
|
244
|
ResourceBundle rb = ResourceBundle.getBundle("metacat-lsid");
|
245
|
String url = rb.getString("metacatserver");
|
246
|
Reader r = null;
|
247
|
try {
|
248
|
|
249
|
Metacat m = MetacatFactory.createMetacatConnection(url);
|
250
|
r = m.query(new StringReader(query));
|
251
|
|
252
|
} catch (MetacatInaccessibleException mie) {
|
253
|
logger.error("Metacat Inaccessible:\n" + mie.getMessage());
|
254
|
} catch (Exception e) {
|
255
|
logger.error("General exception:\n" + e.getMessage());
|
256
|
}
|
257
|
return r;
|
258
|
}
|
259
|
|
260
|
/**
|
261
|
* Given a reader which is a metacat response, parse it and return the
|
262
|
* appropriate rdf
|
263
|
*/
|
264
|
private InputStream parseMetaCatResponse(Reader reader, LSID theLSID)
|
265
|
{
|
266
|
InputStream response = null;
|
267
|
logger.debug("Parsing the metacat response.");
|
268
|
// if there's more than one document, then return rdf listing the
|
269
|
// documents
|
270
|
// otherwise get the document and return rdf based on it
|
271
|
|
272
|
String contents = getStringFromReader(reader);
|
273
|
if (numberDocuments(contents) < 1) {
|
274
|
response = noMetaDataResponse(theLSID);
|
275
|
} else if (numberDocuments(contents) > 1) {
|
276
|
response = metaDataList(contents, theLSID);
|
277
|
} else {
|
278
|
response = getMetaData(contents, theLSID);
|
279
|
}
|
280
|
return response;
|
281
|
}
|
282
|
|
283
|
/**
|
284
|
* There's no metadata for this document
|
285
|
*/
|
286
|
private ByteArrayInputStream noMetaDataResponse(LSID theLSID)
|
287
|
{
|
288
|
ResourceBundle rb = ResourceBundle.getBundle("metacat-lsid");
|
289
|
String metadataLabels = rb.getString("metadatalabels");
|
290
|
|
291
|
String result = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n"
|
292
|
+ "<rdf:RDF xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\" \n"
|
293
|
+ " xmlns:dc=\"http://purl.org/dc/elements/1.1/\" \n"
|
294
|
+ " xmlns:pred=\"urn:lsid:i3c.org:predicates:\" xmlns=\"urn:lsid:"
|
295
|
+ metadataLabels
|
296
|
+ ":predicates:\"> \n"
|
297
|
+ "<rdf:Description rdf:about=\""
|
298
|
+ theLSID.getLsid()
|
299
|
+ "\"> \n"
|
300
|
+ " <pred:title xmlns:pred=\"http://purl.org/dc/elements/1.1/\">There is no metadata for this LSID.</pred:title>\n"
|
301
|
+ "</rdf:Description>\n" + "</rdf:RDF>\n";
|
302
|
|
303
|
ByteArrayInputStream resultBytes = null;
|
304
|
try {
|
305
|
resultBytes = new ByteArrayInputStream(result.getBytes(MetaCatServlet.DEFAULT_ENCODING));
|
306
|
} catch (UnsupportedEncodingException e) {
|
307
|
// TODO Auto-generated catch block
|
308
|
e.printStackTrace();
|
309
|
}
|
310
|
return resultBytes;
|
311
|
}
|
312
|
|
313
|
/**
|
314
|
* There's more than one metdata document
|
315
|
*/
|
316
|
private ByteArrayInputStream metaDataList(String contents, LSID theLSID)
|
317
|
{
|
318
|
ResourceBundle rb = ResourceBundle.getBundle("metacat");
|
319
|
String metadataLabels = rb.getString("metadatalabels");
|
320
|
|
321
|
String result = "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n"
|
322
|
+ "<rdf:RDF xmlns:rdf=\"http://www.w3.org/1999/02/22-rdf-syntax-ns#\" \n"
|
323
|
+ " xmlns:dc=\"http://purl.org/dc/elements/1.1/\" \n"
|
324
|
+ " xmlns:pred=\"urn:lsid:i3c.org:predicates:\" xmlns=\"urn:lsid:"
|
325
|
+ metadataLabels
|
326
|
+ ":predicates:\"> \n"
|
327
|
+ "<rdf:Description rdf:about=\""
|
328
|
+ theLSID.getLsid()
|
329
|
+ "\"> \n"
|
330
|
+ " <pred:title xmlns:pred=\"http://purl.org/dc/elements/1.1/\">There is more than one metadata document for this LSID - which confuses me right now. Try again soon and I'll be less confused.</pred:title>\n"
|
331
|
+ "</rdf:Description>\n" + "</rdf:RDF>\n";
|
332
|
|
333
|
ByteArrayInputStream resultBytes = null;
|
334
|
try {
|
335
|
resultBytes = new ByteArrayInputStream(result.getBytes(MetaCatServlet.DEFAULT_ENCODING));
|
336
|
} catch (UnsupportedEncodingException e) {
|
337
|
// TODO Auto-generated catch block
|
338
|
e.printStackTrace();
|
339
|
}
|
340
|
return resultBytes;
|
341
|
}
|
342
|
|
343
|
/**
|
344
|
* There's just one metadata document
|
345
|
*/
|
346
|
private ByteArrayInputStream getMetaData(String contents, LSID theLSID)
|
347
|
{
|
348
|
String paramString = "<param name=\"dataset/title\">";
|
349
|
ByteArrayInputStream result = null;
|
350
|
|
351
|
if (contents.indexOf(paramString) == -1) {
|
352
|
return noMetaDataResponse(theLSID);
|
353
|
} else {
|
354
|
String parts[] = contents.split(paramString);
|
355
|
String parts2[] = parts[1].split("</param>");
|
356
|
try {
|
357
|
LSID newLSID = new LSID(parts2[0]);
|
358
|
result = doMetadataRequest(newLSID);
|
359
|
|
360
|
} catch (MalformedLSIDException e) {
|
361
|
logger.error("problem generating LSID: " + e);
|
362
|
e.printStackTrace();
|
363
|
} catch (LSIDServerException e) {
|
364
|
logger.error("problem generating LSID: " + e);
|
365
|
e.printStackTrace();
|
366
|
}
|
367
|
}
|
368
|
return result;
|
369
|
}
|
370
|
|
371
|
/**
|
372
|
* Find out how many contents are in this metacat response I'm just using
|
373
|
* string stuff for this - sort of lame, but does the trick more cool would
|
374
|
* be to use xml stuff
|
375
|
*/
|
376
|
private int numberDocuments(String contents)
|
377
|
{
|
378
|
|
379
|
String[] docSplit = contents.split("<document>");
|
380
|
return (docSplit.length - 1);
|
381
|
}
|
382
|
|
383
|
/**
|
384
|
* Given a reader, return a string
|
385
|
*/
|
386
|
private String getStringFromReader(Reader reader)
|
387
|
{
|
388
|
StringBuffer response = new StringBuffer();
|
389
|
|
390
|
try {
|
391
|
BufferedReader bufReader = new BufferedReader(reader);
|
392
|
|
393
|
String line = null;
|
394
|
while ((line = bufReader.readLine()) != null) {
|
395
|
response.append(line);
|
396
|
}
|
397
|
bufReader.close();
|
398
|
|
399
|
} catch (IOException e) {
|
400
|
logger.error("error getting string from reader " + e);
|
401
|
}
|
402
|
return response.toString();
|
403
|
}
|
404
|
|
405
|
/**
|
406
|
* set the LSID for the current thread
|
407
|
*/
|
408
|
static void setCurrentLSID(LSID lsid)
|
409
|
{
|
410
|
currentLSIDs.put(Thread.currentThread(), lsid);
|
411
|
}
|
412
|
|
413
|
static void clearState()
|
414
|
{
|
415
|
currentLSIDs.remove(Thread.currentThread());
|
416
|
}
|
417
|
|
418
|
/**
|
419
|
* get the current LSID for the given thread, for use in XSLT so return a
|
420
|
* string
|
421
|
*/
|
422
|
public static String getLSID(
|
423
|
org.apache.xalan.extensions.XSLProcessorContext foo,
|
424
|
org.apache.xalan.templates.ElemExtensionCall bar)
|
425
|
throws MalformedLSIDException
|
426
|
{
|
427
|
return ((LSID) currentLSIDs.get(Thread.currentThread())).toString();
|
428
|
}
|
429
|
|
430
|
public static String getLSID() throws MalformedLSIDException
|
431
|
{
|
432
|
return ((LSID) currentLSIDs.get(Thread.currentThread())).toString();
|
433
|
}
|
434
|
}
|