1 |
5025
|
daigle
|
/**
|
2 |
|
|
* '$RCSfile$'
|
3 |
|
|
* Purpose: A Class that implements utility methods for a metadata catalog
|
4 |
|
|
* Copyright: 2009 Regents of the University of California and the
|
5 |
|
|
* National Center for Ecological Analysis and Synthesis
|
6 |
|
|
* Authors: Michael Daigle
|
7 |
|
|
*
|
8 |
|
|
* '$Author: daigle $'
|
9 |
|
|
* '$Date: 2009-08-04 14:32:58 -0700 (Tue, 04 Aug 2009) $'
|
10 |
|
|
* '$Revision: 5015 $'
|
11 |
|
|
*
|
12 |
|
|
* This program is free software; you can redistribute it and/or modify
|
13 |
|
|
* it under the terms of the GNU General Public License as published by
|
14 |
|
|
* the Free Software Foundation; either version 2 of the License, or
|
15 |
|
|
* (at your option) any later version.
|
16 |
|
|
*
|
17 |
|
|
* This program is distributed in the hope that it will be useful,
|
18 |
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
19 |
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
20 |
|
|
* GNU General Public License for more details.
|
21 |
|
|
*
|
22 |
|
|
* You should have received a copy of the GNU General Public License
|
23 |
|
|
* along with this program; if not, write to the Free Software
|
24 |
|
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
25 |
|
|
*/
|
26 |
|
|
|
27 |
|
|
package edu.ucsb.nceas.metacat.util;
|
28 |
|
|
|
29 |
|
|
import java.util.Stack;
|
30 |
|
|
import java.util.Vector;
|
31 |
|
|
|
32 |
|
|
import org.apache.log4j.Logger;
|
33 |
|
|
|
34 |
|
|
import edu.ucsb.nceas.dbadapter.AbstractDatabase;
|
35 |
|
|
import edu.ucsb.nceas.metacat.DBSAXHandler;
|
36 |
|
|
import edu.ucsb.nceas.metacat.NodeRecord;
|
37 |
5030
|
daigle
|
import edu.ucsb.nceas.metacat.properties.PropertyService;
|
38 |
5025
|
daigle
|
import edu.ucsb.nceas.utilities.PropertyNotFoundException;
|
39 |
|
|
|
40 |
|
|
/**
|
41 |
|
|
* A suite of utility classes for the metadata catalog server
|
42 |
|
|
*/
|
43 |
|
|
public class DocumentUtil
|
44 |
|
|
{
|
45 |
|
|
|
46 |
|
|
public static AbstractDatabase dbAdapter;
|
47 |
|
|
|
48 |
|
|
private static Logger logMetacat = Logger.getLogger(DocumentUtil.class);
|
49 |
|
|
private static char separator = '.';
|
50 |
|
|
|
51 |
|
|
static {
|
52 |
|
|
try {
|
53 |
|
|
separator = PropertyService.getProperty("document.accNumSeparator").charAt(0);
|
54 |
|
|
} catch (PropertyNotFoundException pnfe) {
|
55 |
|
|
logMetacat.error("Could not retrieve account number separator. "
|
56 |
|
|
+ "Separator set to '.' : " + pnfe.getMessage());
|
57 |
|
|
}
|
58 |
|
|
}
|
59 |
|
|
|
60 |
|
|
/**
|
61 |
|
|
* Get docid from online/url string
|
62 |
|
|
*/
|
63 |
|
|
public static String getDocIdWithRevFromOnlineURL(String url)
|
64 |
|
|
{
|
65 |
|
|
String docid = null;
|
66 |
|
|
String DOCID = "docid";
|
67 |
|
|
boolean find = false;
|
68 |
|
|
char limited = '&';
|
69 |
|
|
int count = 0; //keep track how many & was found
|
70 |
|
|
Vector list = new Vector();// keep index number for &
|
71 |
|
|
if (url == null) {
|
72 |
|
|
logMetacat.info("url is null and null will be returned");
|
73 |
|
|
return docid;
|
74 |
|
|
}
|
75 |
|
|
// the first element in list is 0
|
76 |
|
|
list.add(new Integer(0));
|
77 |
|
|
for (int i = 0; i < url.length(); i++) {
|
78 |
|
|
if (url.charAt(i) == limited) {
|
79 |
|
|
// count plus 1
|
80 |
|
|
count++;
|
81 |
|
|
list.add(new Integer(i));
|
82 |
|
|
// get substring beween two &
|
83 |
|
|
String str = url.substring(
|
84 |
|
|
((Integer) list.elementAt(count - 1)).intValue(), i);
|
85 |
|
|
logMetacat.info("substring between two & is: " + str);
|
86 |
|
|
//if the subString contains docid, we got it
|
87 |
|
|
if (str.indexOf(DOCID) != -1) {
|
88 |
|
|
//get index of '="
|
89 |
|
|
int start = getIndexForGivenChar(str, '=') + 1;
|
90 |
|
|
int end = str.length();
|
91 |
|
|
docid = str.substring(start, end);
|
92 |
|
|
find = true;
|
93 |
|
|
}//if
|
94 |
|
|
}//if
|
95 |
|
|
}//for
|
96 |
|
|
//if not find, we need check the subtring between the index of last &
|
97 |
|
|
// and
|
98 |
|
|
// the end of string
|
99 |
|
|
if (!find) {
|
100 |
|
|
logMetacat.info("Checking the last substring");
|
101 |
|
|
String str = url.substring(((Integer) list.elementAt(count))
|
102 |
|
|
.intValue() + 1, url.length());
|
103 |
|
|
logMetacat.info("Last substring is: " + str);
|
104 |
|
|
if (str.indexOf(DOCID) != -1) {
|
105 |
|
|
//get index of '="
|
106 |
|
|
int start = getIndexForGivenChar(str, '=') + 1;
|
107 |
|
|
int end = str.length();
|
108 |
|
|
docid = str.substring(start, end);
|
109 |
|
|
find = true;
|
110 |
|
|
}//if
|
111 |
|
|
}//if
|
112 |
|
|
logMetacat.info("The docid from online url is:" + docid);
|
113 |
|
|
return docid.trim();
|
114 |
|
|
}
|
115 |
|
|
|
116 |
|
|
|
117 |
|
|
/**
|
118 |
|
|
* Eocgorid identifier will look like: ecogrid://knb/tao.1.1
|
119 |
|
|
* The AccessionNumber tao.1.1 will be returned. If the given doesn't
|
120 |
|
|
* contains ecogrid, null will be returned.
|
121 |
|
|
* @param identifier String
|
122 |
|
|
* @return String
|
123 |
|
|
*/
|
124 |
|
|
public static String getAccessionNumberFromEcogridIdentifier(String identifier)
|
125 |
|
|
{
|
126 |
|
|
String accessionNumber = null;
|
127 |
|
|
if (identifier != null && identifier.startsWith(DBSAXHandler.ECOGRID))
|
128 |
|
|
{
|
129 |
|
|
// find the last "/" in identifier
|
130 |
|
|
int indexOfLastSlash = identifier.lastIndexOf("/");
|
131 |
|
|
int start = indexOfLastSlash+1;
|
132 |
|
|
int end = identifier.length();
|
133 |
|
|
accessionNumber = identifier.substring(start, end);
|
134 |
|
|
}
|
135 |
|
|
logMetacat.warn("The accession number from url is " +
|
136 |
|
|
accessionNumber);
|
137 |
|
|
return accessionNumber;
|
138 |
|
|
}
|
139 |
|
|
|
140 |
|
|
private static int getIndexForGivenChar(String str, char character)
|
141 |
|
|
{
|
142 |
|
|
int index = -1;
|
143 |
|
|
// make sure str is not null
|
144 |
|
|
if (str == null) {
|
145 |
|
|
logMetacat.info(
|
146 |
|
|
"The given str is null and -1 will be returned");
|
147 |
|
|
return index;
|
148 |
|
|
}
|
149 |
|
|
// got though the string
|
150 |
|
|
for (int i = 0; i < str.length(); i++) {
|
151 |
|
|
// find the first one then break the loop
|
152 |
|
|
if (str.charAt(i) == character) {
|
153 |
|
|
index = i;
|
154 |
|
|
break;
|
155 |
|
|
}//if
|
156 |
|
|
}//for
|
157 |
|
|
logMetacat.info("the index for char " + character + " is: "
|
158 |
|
|
+ index);
|
159 |
|
|
return index;
|
160 |
|
|
}
|
161 |
|
|
|
162 |
|
|
/**
|
163 |
|
|
* Utility method to get docid from a given string
|
164 |
|
|
*
|
165 |
|
|
* @param string, the given string should be these two format: 1) str1.str2
|
166 |
|
|
* in this case docid= str1.str2 2) str1.str2.str3, in this case
|
167 |
|
|
* docid =str1.str2
|
168 |
|
|
* @param the sperator char
|
169 |
|
|
*/
|
170 |
|
|
public static String getDocIdFromString(String str)
|
171 |
|
|
{
|
172 |
|
|
String docId = null;
|
173 |
|
|
if (str == null) {
|
174 |
|
|
logMetacat.info(
|
175 |
|
|
"The given str is null and null will be returned"
|
176 |
|
|
+ " in getDocIdfromString");
|
177 |
|
|
return docId;
|
178 |
|
|
} //make sure docid is not null
|
179 |
|
|
int dotNumber = 0;//count how many dots in given string
|
180 |
|
|
int indexOfLastDot = 0;
|
181 |
|
|
|
182 |
|
|
for (int i = 0; i < str.length(); i++) {
|
183 |
|
|
if (str.charAt(i) == separator) {
|
184 |
|
|
dotNumber++;//count how many dots
|
185 |
|
|
indexOfLastDot = i;//keep the last dot postion
|
186 |
|
|
}
|
187 |
|
|
}//for
|
188 |
|
|
|
189 |
|
|
//The string formatt is wrong, because it has more than two or less
|
190 |
|
|
// than
|
191 |
|
|
//one seperator
|
192 |
|
|
if (dotNumber > 2 || dotNumber < 1) {
|
193 |
|
|
docId = null;
|
194 |
|
|
} else if (dotNumber == 2) //the case for str1.str2.str3
|
195 |
|
|
{
|
196 |
|
|
docId = str.substring(0, indexOfLastDot);
|
197 |
|
|
} else if (dotNumber == 1) //the case for str1.str2
|
198 |
|
|
{
|
199 |
|
|
docId = str;
|
200 |
|
|
}
|
201 |
|
|
|
202 |
|
|
return docId;
|
203 |
|
|
}//getDocIdFromString
|
204 |
|
|
|
205 |
|
|
/**
|
206 |
|
|
* Utility method to get version number from a given string
|
207 |
|
|
*
|
208 |
|
|
* @param string, the given string should be these two format: 1)
|
209 |
|
|
* str1.str2(no version) version =-1; 2) str1.str2.str3, in this
|
210 |
|
|
* case version = str3; 3) other, vresion =-2
|
211 |
|
|
*/
|
212 |
|
|
public static int getVersionFromString(String str)
|
213 |
|
|
throws NumberFormatException
|
214 |
|
|
{
|
215 |
|
|
int version = -1;
|
216 |
|
|
String versionString = null;
|
217 |
|
|
int dotNumber = 0;//count how many dots in given string
|
218 |
|
|
int indexOfLastDot = 0;
|
219 |
|
|
|
220 |
|
|
for (int i = 0; i < str.length(); i++) {
|
221 |
|
|
if (str.charAt(i) == separator) {
|
222 |
|
|
dotNumber++;//count how many dots
|
223 |
|
|
indexOfLastDot = i;//keep the last dot postion
|
224 |
|
|
}
|
225 |
|
|
}//for
|
226 |
|
|
|
227 |
|
|
//The string formatt is wrong, because it has more than two or less
|
228 |
|
|
// than
|
229 |
|
|
//one seperator
|
230 |
|
|
if (dotNumber > 2 || dotNumber < 1) {
|
231 |
|
|
version = -2;
|
232 |
|
|
} else if (dotNumber == 2 && (indexOfLastDot != (str.length() - 1)))
|
233 |
|
|
//the case for str1.str2.str3
|
234 |
|
|
{
|
235 |
|
|
versionString = str.substring((indexOfLastDot + 1), str.length());
|
236 |
|
|
version = Integer.parseInt(versionString);
|
237 |
|
|
} else if (dotNumber == 1) //the case for str1.str2
|
238 |
|
|
{
|
239 |
|
|
version = -1;
|
240 |
|
|
}
|
241 |
|
|
|
242 |
|
|
return version;
|
243 |
|
|
}//getVersionFromString
|
244 |
|
|
|
245 |
|
|
/**
|
246 |
|
|
* Utility method to get version string from a given string
|
247 |
|
|
*
|
248 |
|
|
* @param string, the given string should be these two format: 1)
|
249 |
|
|
* str1.str2(no version) version=null; 2) str1.str2.str3, in
|
250 |
|
|
* this case version = str3; 3) other, vresion =null;
|
251 |
|
|
*/
|
252 |
|
|
public static String getRevisionStringFromString(String str)
|
253 |
|
|
throws NumberFormatException
|
254 |
|
|
{
|
255 |
|
|
// String to store the version
|
256 |
|
|
String versionString = null;
|
257 |
|
|
int dotNumber = 0;//count how many dots in given string
|
258 |
|
|
int indexOfLastDot = 0;
|
259 |
|
|
|
260 |
|
|
for (int i = 0; i < str.length(); i++) {
|
261 |
|
|
if (str.charAt(i) == separator) {
|
262 |
|
|
dotNumber++;//count how many dots
|
263 |
|
|
indexOfLastDot = i;//keep the last dot postion
|
264 |
|
|
}
|
265 |
|
|
}//for
|
266 |
|
|
|
267 |
|
|
//The string formatt is wrong, because it has more than two or less
|
268 |
|
|
// than
|
269 |
|
|
//one seperator
|
270 |
|
|
if (dotNumber > 2 || dotNumber < 1) {
|
271 |
|
|
versionString = null;
|
272 |
|
|
} else if (dotNumber == 2 && (indexOfLastDot != (str.length() - 1))) {
|
273 |
|
|
//the case for str1.str2.str3
|
274 |
|
|
// indexOfLastDot != (str.length() -1) means get rid of str1.str2.
|
275 |
|
|
versionString = str.substring((indexOfLastDot + 1), str.length());
|
276 |
|
|
} else if (dotNumber == 1) //the case for str1.str2 or str1.str2.
|
277 |
|
|
{
|
278 |
|
|
versionString = null;
|
279 |
|
|
}
|
280 |
|
|
|
281 |
|
|
return versionString;
|
282 |
|
|
}//getVersionFromString
|
283 |
|
|
|
284 |
|
|
/**
|
285 |
|
|
* This method will get docid from an AccessionNumber. There is no
|
286 |
|
|
* assumption the accessnumber will be str1.str2.str3. It can be more. So
|
287 |
|
|
* we think the docid will be get rid of last part
|
288 |
|
|
*/
|
289 |
|
|
public static String getDocIdFromAccessionNumber(String accessionNumber)
|
290 |
|
|
{
|
291 |
|
|
String docid = null;
|
292 |
|
|
if (accessionNumber == null) { return docid; }
|
293 |
|
|
int indexOfLastSeperator = accessionNumber.lastIndexOf(separator);
|
294 |
|
|
docid = accessionNumber.substring(0, indexOfLastSeperator);
|
295 |
|
|
logMetacat.info("after parsing accessionnumber, docid is "
|
296 |
|
|
+ docid);
|
297 |
|
|
return docid;
|
298 |
|
|
}
|
299 |
|
|
|
300 |
|
|
/**
|
301 |
|
|
* This method will get inline data id without the revision number.
|
302 |
|
|
* So if inlineData.1.2 is passed as input, inlineData.2 is returned.
|
303 |
|
|
*/
|
304 |
|
|
public static String getInlineDataIdWithoutRev(String accessionNumber)
|
305 |
|
|
{
|
306 |
|
|
String docid = null;
|
307 |
|
|
if (accessionNumber == null) { return docid; }
|
308 |
|
|
int indexOfLastSeperator = accessionNumber.lastIndexOf(separator);
|
309 |
|
|
String version = accessionNumber.substring(indexOfLastSeperator,
|
310 |
|
|
accessionNumber.length());
|
311 |
|
|
accessionNumber = accessionNumber.substring(0, indexOfLastSeperator);
|
312 |
|
|
indexOfLastSeperator = accessionNumber.lastIndexOf(separator);
|
313 |
|
|
docid = accessionNumber.substring(0, indexOfLastSeperator) + version;
|
314 |
|
|
logMetacat.info("after parsing accessionnumber, docid is "
|
315 |
|
|
+ docid);
|
316 |
|
|
|
317 |
|
|
return docid;
|
318 |
|
|
}
|
319 |
|
|
|
320 |
|
|
/**
|
321 |
|
|
* This method will call both getDocIdFromString and
|
322 |
|
|
* getDocIdFromAccessionNumber. So first, if the string looks str1.str2,
|
323 |
|
|
* the docid will be str1.str2. If the string is str1.str2.str3, the docid
|
324 |
|
|
* will be str1.str2. If the string is str1.str2.str3.str4 or more, the
|
325 |
|
|
* docid will be str1.str2.str3. If the string look like str1, null will be
|
326 |
|
|
* returned
|
327 |
|
|
*
|
328 |
|
|
*/
|
329 |
|
|
public static String getSmartDocId(String str)
|
330 |
|
|
{
|
331 |
|
|
String docid = null;
|
332 |
|
|
//call geDocIdFromString first.
|
333 |
|
|
docid = getDocIdFromString(str);
|
334 |
|
|
// If docid is null, try to call getDocIdFromAccessionNumber
|
335 |
|
|
// it will handle the seperator more than2
|
336 |
|
|
if (docid == null) {
|
337 |
|
|
docid = getDocIdFromAccessionNumber(str);
|
338 |
|
|
}
|
339 |
|
|
logMetacat.info("The docid get from smart docid getor is "
|
340 |
|
|
+ docid);
|
341 |
|
|
return docid;
|
342 |
|
|
}
|
343 |
|
|
|
344 |
|
|
/**
|
345 |
|
|
* This method will get revision from an AccessionNumber. There is no
|
346 |
|
|
* assumption the accessnumber will be str1.str2.str3. It can be more. So
|
347 |
|
|
* we think the docid will be get rid of last part
|
348 |
|
|
*/
|
349 |
|
|
public static int getRevisionFromAccessionNumber(String accessionNumber)
|
350 |
|
|
throws NumberFormatException
|
351 |
|
|
{
|
352 |
|
|
String rev = null;
|
353 |
|
|
int revNumber = -1;
|
354 |
|
|
if (accessionNumber == null) { return revNumber; }
|
355 |
|
|
int indexOfLastSeperator = accessionNumber.lastIndexOf(separator);
|
356 |
|
|
rev = accessionNumber.substring(indexOfLastSeperator + 1,
|
357 |
|
|
accessionNumber.length());
|
358 |
|
|
revNumber = Integer.parseInt(rev);
|
359 |
|
|
logMetacat.info("after parsing accessionnumber, rev is "
|
360 |
|
|
+ revNumber);
|
361 |
|
|
return revNumber;
|
362 |
|
|
}
|
363 |
|
|
|
364 |
|
|
/**
|
365 |
|
|
* Method to get docidwithrev from eml2 inline data id The eml inline data
|
366 |
|
|
* id would look like eml.200.2.3
|
367 |
|
|
*/
|
368 |
|
|
public static String getDocIdWithoutRevFromInlineDataID(String inlineDataID)
|
369 |
|
|
{
|
370 |
|
|
String docidWithoutRev = null;
|
371 |
|
|
if (inlineDataID == null) { return docidWithoutRev; }
|
372 |
|
|
char charSeperator = separator;
|
373 |
|
|
int targetNumberOfSeperator = 2;// we want to know his index
|
374 |
|
|
int numberOfSeperator = 0;
|
375 |
|
|
for (int i = 0; i < inlineDataID.length(); i++) {
|
376 |
|
|
// meet seperator, increase number of seperator
|
377 |
|
|
if (inlineDataID.charAt(i) == charSeperator) {
|
378 |
|
|
numberOfSeperator++;
|
379 |
|
|
}
|
380 |
|
|
// if number of seperator reach the target one, record the index(i)
|
381 |
|
|
// and get substring and terminate the loop
|
382 |
|
|
if (numberOfSeperator == targetNumberOfSeperator) {
|
383 |
|
|
docidWithoutRev = inlineDataID.substring(0, i);
|
384 |
|
|
break;
|
385 |
|
|
}
|
386 |
|
|
}
|
387 |
|
|
|
388 |
|
|
logMetacat.info("Docid without rev from inlinedata id: "
|
389 |
|
|
+ docidWithoutRev);
|
390 |
|
|
return docidWithoutRev;
|
391 |
|
|
|
392 |
|
|
}
|
393 |
|
|
|
394 |
|
|
/**
|
395 |
|
|
* Revise stack change a stack to opposite order
|
396 |
|
|
*/
|
397 |
|
|
public static Stack<NodeRecord> reviseStack(Stack<NodeRecord> stack)
|
398 |
|
|
{
|
399 |
|
|
Stack result = new Stack();
|
400 |
|
|
// make sure the parameter is correct
|
401 |
|
|
if (stack == null || stack.isEmpty()) {
|
402 |
|
|
result = stack;
|
403 |
|
|
return result;
|
404 |
|
|
}
|
405 |
|
|
|
406 |
|
|
while (!stack.isEmpty()) {
|
407 |
|
|
Object obj = stack.pop();
|
408 |
|
|
result.push(obj);
|
409 |
|
|
}
|
410 |
|
|
return result;
|
411 |
|
|
}
|
412 |
|
|
|
413 |
|
|
}
|