1
|
/**
|
2
|
* '$RCSfile$'
|
3
|
* Purpose: A Class that handles the SAX XML events as they
|
4
|
* are generated from XML documents
|
5
|
* Copyright: 2000 Regents of the University of California and the
|
6
|
* National Center for Ecological Analysis and Synthesis
|
7
|
* Authors: Matt Jones, Jivka Bojilova
|
8
|
* Release: @release@
|
9
|
*
|
10
|
* '$Author: jones $'
|
11
|
* '$Date: 2004-03-30 13:35:19 -0800 (Tue, 30 Mar 2004) $'
|
12
|
* '$Revision: 2076 $'
|
13
|
*
|
14
|
* This program is free software; you can redistribute it and/or modify
|
15
|
* it under the terms of the GNU General Public License as published by
|
16
|
* the Free Software Foundation; either version 2 of the License, or
|
17
|
* (at your option) any later version.
|
18
|
*
|
19
|
* This program is distributed in the hope that it will be useful,
|
20
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
21
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
22
|
* GNU General Public License for more details.
|
23
|
*
|
24
|
* You should have received a copy of the GNU General Public License
|
25
|
* along with this program; if not, write to the Free Software
|
26
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
27
|
*/
|
28
|
|
29
|
package edu.ucsb.nceas.metacat;
|
30
|
|
31
|
import java.sql.PreparedStatement;
|
32
|
import java.sql.ResultSet;
|
33
|
import java.sql.SQLException;
|
34
|
import java.sql.Statement;
|
35
|
import java.util.EmptyStackException;
|
36
|
import java.util.Enumeration;
|
37
|
import java.util.Hashtable;
|
38
|
import java.util.Stack;
|
39
|
import java.util.Vector;
|
40
|
|
41
|
import edu.ucsb.nceas.morpho.datapackage.Triple;
|
42
|
import edu.ucsb.nceas.morpho.datapackage.TripleCollection;
|
43
|
|
44
|
import org.xml.sax.Attributes;
|
45
|
import org.xml.sax.SAXException;
|
46
|
import org.xml.sax.SAXParseException;
|
47
|
import org.xml.sax.ext.DeclHandler;
|
48
|
import org.xml.sax.ext.LexicalHandler;
|
49
|
import org.xml.sax.helpers.DefaultHandler;
|
50
|
|
51
|
/**
|
52
|
* A database aware Class implementing callback bethods for the SAX parser to
|
53
|
* call when processing the XML stream and generating events.
|
54
|
*/
|
55
|
public class DBSAXHandler extends DefaultHandler implements LexicalHandler,
|
56
|
DeclHandler, Runnable
|
57
|
{
|
58
|
|
59
|
protected boolean atFirstElement;
|
60
|
|
61
|
protected boolean processingDTD;
|
62
|
|
63
|
protected String docname = null;
|
64
|
|
65
|
protected String doctype;
|
66
|
|
67
|
protected String systemid;
|
68
|
|
69
|
private boolean stackCreated = false;
|
70
|
|
71
|
protected Stack nodeStack;
|
72
|
|
73
|
protected Vector nodeIndex;
|
74
|
|
75
|
protected DBConnection connection = null;
|
76
|
|
77
|
protected DocumentImpl currentDocument;
|
78
|
|
79
|
protected DBSAXNode rootNode;
|
80
|
|
81
|
protected String action = null;
|
82
|
|
83
|
protected String docid = null;
|
84
|
|
85
|
protected String revision = null;
|
86
|
|
87
|
protected String user = null;
|
88
|
|
89
|
protected String[] groups = null;
|
90
|
|
91
|
protected String pub = null;
|
92
|
|
93
|
protected Thread xmlIndex;
|
94
|
|
95
|
private boolean endDocument = false;
|
96
|
|
97
|
protected int serverCode = 1;
|
98
|
|
99
|
protected Hashtable namespaces = new Hashtable();
|
100
|
|
101
|
protected boolean hitTextNode = false; // a flag to hit text node
|
102
|
|
103
|
// a buffer to keep all text nodes for same element
|
104
|
// it is for element was splited
|
105
|
protected StringBuffer textBuffer = new StringBuffer();
|
106
|
|
107
|
protected Stack textBufferStack = new Stack();
|
108
|
|
109
|
protected static final int MAXDATACHARS = 4000;
|
110
|
|
111
|
//protected static final int MAXDATACHARS = 50;
|
112
|
protected static final long INDEXDELAY = 10000;
|
113
|
|
114
|
// methods writeChildNodeToDB, setAttribute, setNamespace,
|
115
|
// writeTextForDBSAXNode will increase endNodeId.
|
116
|
protected long endNodeId = -1; // The end node id for a substree
|
117
|
// DOCTITLE attr cleared from the db
|
118
|
// private static final int MAXTITLELEN = 1000;
|
119
|
|
120
|
//HandlerTriple stuff
|
121
|
TripleCollection tripleList = new TripleCollection();
|
122
|
|
123
|
Triple currentTriple = new Triple();
|
124
|
|
125
|
boolean startParseTriple = false;
|
126
|
|
127
|
boolean hasTriple = false;
|
128
|
|
129
|
/**
|
130
|
* Construct an instance of the handler class
|
131
|
*
|
132
|
* @param conn the JDBC connection to which information is written
|
133
|
*/
|
134
|
private DBSAXHandler(DBConnection conn)
|
135
|
{
|
136
|
this.connection = conn;
|
137
|
this.atFirstElement = true;
|
138
|
this.processingDTD = false;
|
139
|
|
140
|
// Create the stack for keeping track of node context
|
141
|
// if it doesn't already exist
|
142
|
if (!stackCreated) {
|
143
|
nodeStack = new Stack();
|
144
|
nodeIndex = new Vector();
|
145
|
stackCreated = true;
|
146
|
}
|
147
|
}
|
148
|
|
149
|
/**
|
150
|
* Construct an instance of the handler class
|
151
|
*
|
152
|
* @param conn the JDBC connection to which information is written
|
153
|
* @param action - "INSERT" or "UPDATE"
|
154
|
* @param docid to be inserted or updated into JDBC connection
|
155
|
* @param user the user connected to MetaCat servlet and owns the document
|
156
|
* @param groups the groups to which user belongs
|
157
|
* @param pub flag for public "read" access on document
|
158
|
* @param serverCode the serverid from xml_replication on which this
|
159
|
* document resides.
|
160
|
*
|
161
|
*/
|
162
|
/* TODO excise this constructor because not used anywhere in project
|
163
|
public DBSAXHandler(DBConnection conn, String action, String docid,
|
164
|
String user, String[] groups, String pub, int serverCode)
|
165
|
{
|
166
|
this(conn);
|
167
|
this.action = action;
|
168
|
this.docid = docid;
|
169
|
this.user = user;
|
170
|
this.groups = groups;
|
171
|
this.pub = pub;
|
172
|
this.serverCode = serverCode;
|
173
|
this.xmlIndex = new Thread(this);
|
174
|
}
|
175
|
*/
|
176
|
/**
|
177
|
* Construct an instance of the handler class In this constructor, user can
|
178
|
* specify the version need to upadate
|
179
|
*
|
180
|
* @param conn the JDBC connection to which information is written
|
181
|
* @param action - "INSERT" or "UPDATE"
|
182
|
* @param docid to be inserted or updated into JDBC connection
|
183
|
* @param revision, the user specified the revision need to be update
|
184
|
* @param user the user connected to MetaCat servlet and owns the document
|
185
|
* @param groups the groups to which user belongs
|
186
|
* @param pub flag for public "read" access on document
|
187
|
* @param serverCode the serverid from xml_replication on which this
|
188
|
* document resides.
|
189
|
*
|
190
|
*/
|
191
|
public DBSAXHandler(DBConnection conn, String action, String docid,
|
192
|
String revision, String user, String[] groups, String pub,
|
193
|
int serverCode)
|
194
|
{
|
195
|
this(conn);
|
196
|
this.action = action;
|
197
|
this.docid = docid;
|
198
|
this.revision = revision;
|
199
|
this.user = user;
|
200
|
this.groups = groups;
|
201
|
this.pub = pub;
|
202
|
this.serverCode = serverCode;
|
203
|
this.xmlIndex = new Thread(this);
|
204
|
}
|
205
|
|
206
|
/** SAX Handler that receives notification of beginning of the document */
|
207
|
public void startDocument() throws SAXException
|
208
|
{
|
209
|
MetaCatUtil.debugMessage("start Document", 50);
|
210
|
|
211
|
// Create the document node representation as root
|
212
|
rootNode = new DBSAXNode(connection, this.docid);
|
213
|
// Add the node to the stack, so that any text data can be
|
214
|
// added as it is encountered
|
215
|
nodeStack.push(rootNode);
|
216
|
}
|
217
|
|
218
|
/** SAX Handler that receives notification of end of the document */
|
219
|
public void endDocument() throws SAXException
|
220
|
{
|
221
|
MetaCatUtil.debugMessage("end Document", 50);
|
222
|
// Starting new thread for writing XML Index.
|
223
|
// It calls the run method of the thread.
|
224
|
|
225
|
//if it is data package insert triple into relationtion table;
|
226
|
if (doctype != null
|
227
|
&& MetaCatUtil.getOptionList(
|
228
|
MetaCatUtil.getOption("packagedoctype")).contains(
|
229
|
doctype) && hasTriple) {
|
230
|
try {
|
231
|
//initial handler and write into relationdb
|
232
|
RelationHandler handler = new RelationHandler(docid, doctype,
|
233
|
connection, tripleList);
|
234
|
} catch (Exception e) {
|
235
|
MetaCatUtil.debugMessage(
|
236
|
"Failed to write triples into relation table"
|
237
|
+ e.getMessage(), 30);
|
238
|
throw new SAXException(
|
239
|
"Failed to write triples into relation table "
|
240
|
+ e.getMessage());
|
241
|
}
|
242
|
}
|
243
|
boolean useXMLIndex =
|
244
|
(new Boolean(MetaCatUtil.getOption("usexmlindex"))).booleanValue();
|
245
|
if (useXMLIndex) {
|
246
|
try {
|
247
|
xmlIndex.start();
|
248
|
} catch (NullPointerException e) {
|
249
|
xmlIndex = null;
|
250
|
throw new SAXException(
|
251
|
"Problem with starting thread for writing XML Index. "
|
252
|
+ e.getMessage());
|
253
|
}
|
254
|
}
|
255
|
}
|
256
|
|
257
|
/** SAX Handler that is called at the start of Namespace */
|
258
|
public void startPrefixMapping(String prefix, String uri)
|
259
|
throws SAXException
|
260
|
{
|
261
|
MetaCatUtil.debugMessage("NAMESPACE", 50);
|
262
|
|
263
|
namespaces.put(prefix, uri);
|
264
|
}
|
265
|
|
266
|
/** SAX Handler that is called at the start of each XML element */
|
267
|
public void startElement(String uri, String localName, String qName,
|
268
|
Attributes atts) throws SAXException
|
269
|
{
|
270
|
// for element <eml:eml...> qname is "eml:eml", local name is "eml"
|
271
|
// for element <acl....> both qname and local name is "eml"
|
272
|
// uri is namesapce
|
273
|
MetaCatUtil.debugMessage("Start ELEMENT(qName) " + qName, 50);
|
274
|
MetaCatUtil.debugMessage("Start ELEMENT(localName) " + localName, 50);
|
275
|
MetaCatUtil.debugMessage("Start ELEMENT(uri) " + uri, 50);
|
276
|
|
277
|
DBSAXNode parentNode = null;
|
278
|
DBSAXNode currentNode = null;
|
279
|
|
280
|
// Get a reference to the parent node for the id
|
281
|
try {
|
282
|
parentNode = (DBSAXNode) nodeStack.peek();
|
283
|
} catch (EmptyStackException e) {
|
284
|
parentNode = null;
|
285
|
}
|
286
|
|
287
|
// If hit a text node, we need write this text for current's parent
|
288
|
// node
|
289
|
// This will happend if the element is mixted
|
290
|
if (hitTextNode && parentNode != null) {
|
291
|
// write the textbuffer into db for parent node.
|
292
|
endNodeId = writeTextForDBSAXNode(endNodeId, textBuffer, parentNode);
|
293
|
// rest hitTextNode
|
294
|
hitTextNode = false;
|
295
|
// reset textbuffer
|
296
|
textBuffer = null;
|
297
|
textBuffer = new StringBuffer();
|
298
|
|
299
|
}
|
300
|
|
301
|
// Document representation that points to the root document node
|
302
|
if (atFirstElement) {
|
303
|
atFirstElement = false;
|
304
|
// If no DOCTYPE declaration: docname = root element
|
305
|
// doctype = root element name or name space
|
306
|
if (docname == null) {
|
307
|
docname = localName;
|
308
|
// if uri isn't null doctype = uri(namespace)
|
309
|
// othewise root element
|
310
|
if (uri != null && !(uri.trim()).equals("")) {
|
311
|
doctype = uri;
|
312
|
} else {
|
313
|
doctype = docname;
|
314
|
}
|
315
|
MetaCatUtil.debugMessage("DOCNAME-a: " + docname, 30);
|
316
|
MetaCatUtil.debugMessage("DOCTYPE-a: " + doctype, 30);
|
317
|
} else if (doctype == null) {
|
318
|
// because docname is not null and it is declared in dtd
|
319
|
// so could not be in schema, no namespace
|
320
|
doctype = docname;
|
321
|
MetaCatUtil.debugMessage("DOCTYPE-b: " + doctype, 30);
|
322
|
}
|
323
|
rootNode.writeNodename(docname);
|
324
|
try {
|
325
|
// for validated XML Documents store a reference to XML DB
|
326
|
// Catalog
|
327
|
// Because this is select statement and it needn't to roll back
|
328
|
// if
|
329
|
// insert document action fialed.
|
330
|
// In order to decrease DBConnection usage count, we get a new
|
331
|
// dbconnection from pool
|
332
|
String catalogid = null;
|
333
|
DBConnection dbConn = null;
|
334
|
int serialNumber = -1;
|
335
|
|
336
|
if (systemid != null) {
|
337
|
try {
|
338
|
// Get dbconnection
|
339
|
dbConn = DBConnectionPool
|
340
|
.getDBConnection("DBSAXHandler.startElement");
|
341
|
serialNumber = dbConn.getCheckOutSerialNumber();
|
342
|
|
343
|
Statement stmt = dbConn.createStatement();
|
344
|
ResultSet rs = stmt
|
345
|
.executeQuery("SELECT catalog_id FROM xml_catalog "
|
346
|
+ "WHERE entry_type = 'DTD' "
|
347
|
+ "AND public_id = '" + doctype + "'");
|
348
|
boolean hasRow = rs.next();
|
349
|
if (hasRow) {
|
350
|
catalogid = rs.getString(1);
|
351
|
}
|
352
|
stmt.close();
|
353
|
}//try
|
354
|
finally {
|
355
|
// Return dbconnection
|
356
|
DBConnectionPool.returnDBConnection(dbConn,
|
357
|
serialNumber);
|
358
|
}//finally
|
359
|
}
|
360
|
|
361
|
//create documentImpl object by the constructor which can
|
362
|
// specify
|
363
|
//the revision
|
364
|
currentDocument = new DocumentImpl(connection, rootNode
|
365
|
.getNodeID(), docname, doctype, docid, revision,
|
366
|
action, user, this.pub, catalogid, this.serverCode);
|
367
|
|
368
|
} catch (Exception ane) {
|
369
|
throw (new SAXException("Error in DBSaxHandler.startElement "
|
370
|
+ action, ane));
|
371
|
}
|
372
|
}
|
373
|
|
374
|
// Create the current node representation
|
375
|
currentNode = new DBSAXNode(connection, qName, localName, parentNode,
|
376
|
currentDocument.getRootNodeID(), docid, currentDocument
|
377
|
.getDoctype());
|
378
|
|
379
|
// Add all of the namespaces
|
380
|
String prefix;
|
381
|
String nsuri;
|
382
|
Enumeration prefixes = namespaces.keys();
|
383
|
while (prefixes.hasMoreElements()) {
|
384
|
prefix = (String) prefixes.nextElement();
|
385
|
nsuri = (String) namespaces.get(prefix);
|
386
|
currentNode.setNamespace(prefix, nsuri, docid);
|
387
|
}
|
388
|
namespaces = null;
|
389
|
namespaces = new Hashtable();
|
390
|
|
391
|
// Add all of the attributes
|
392
|
for (int i = 0; i < atts.getLength(); i++) {
|
393
|
String attributeName = atts.getQName(i);
|
394
|
String attributeValue = atts.getValue(i);
|
395
|
endNodeId = currentNode.setAttribute(attributeName, attributeValue,
|
396
|
docid);
|
397
|
|
398
|
// To handle name space and schema location if the attribute name
|
399
|
// is
|
400
|
// xsi:schemaLocation. If the name space is in not in catalog table
|
401
|
// it will be regeistered.
|
402
|
if (attributeName != null
|
403
|
&& attributeName
|
404
|
.indexOf(MetaCatServlet.SCHEMALOCATIONKEYWORD) != -1) {
|
405
|
SchemaLocationResolver resolver = new SchemaLocationResolver(
|
406
|
attributeValue);
|
407
|
resolver.resolveNameSpace();
|
408
|
|
409
|
}
|
410
|
}
|
411
|
|
412
|
// Add the node to the stack, so that any text data can be
|
413
|
// added as it is encountered
|
414
|
nodeStack.push(currentNode);
|
415
|
// Add the node to the vector used by thread for writing XML Index
|
416
|
nodeIndex.addElement(currentNode);
|
417
|
// start parsing triple
|
418
|
if (doctype != null
|
419
|
&& MetaCatUtil.getOptionList(
|
420
|
MetaCatUtil.getOption("packagedoctype")).contains(
|
421
|
doctype) && localName.equals("triple")) {
|
422
|
startParseTriple = true;
|
423
|
hasTriple = true;
|
424
|
currentTriple = new Triple();
|
425
|
}
|
426
|
}
|
427
|
|
428
|
/* The run method of xmlIndex thread. It writes XML Index for the document. */
|
429
|
public void run()
|
430
|
{
|
431
|
DBSAXNode currNode = null;
|
432
|
DBSAXNode prevNode = null;
|
433
|
DBConnection dbConn = null;
|
434
|
int serialNumber = -1;
|
435
|
String doctype = currentDocument.getDoctype();
|
436
|
int step = 0;
|
437
|
int counter = 0;
|
438
|
|
439
|
try {
|
440
|
// stop 5 second
|
441
|
Thread.sleep(5000);
|
442
|
// Opening separate db connection for writing XML Index
|
443
|
dbConn = DBConnectionPool.getDBConnection("DBSAXHandler.run");
|
444
|
serialNumber = dbConn.getCheckOutSerialNumber();
|
445
|
dbConn.setAutoCommit(false);
|
446
|
//make sure record is done
|
447
|
checkDocumentTable();
|
448
|
// Going through the elements of the document and writing its Index
|
449
|
Enumeration nodes = nodeIndex.elements();
|
450
|
while (nodes.hasMoreElements()) {
|
451
|
currNode = (DBSAXNode) nodes.nextElement();
|
452
|
currNode.updateNodeIndex(dbConn, docid, doctype);
|
453
|
}
|
454
|
dbConn.commit();
|
455
|
|
456
|
} catch (Exception e) {
|
457
|
try {
|
458
|
dbConn.rollback();
|
459
|
//dbconn.close();
|
460
|
} catch (SQLException sqle) {
|
461
|
}
|
462
|
MetaCatUtil.debugMessage("Error in DBSAXHandler.run "
|
463
|
+ e.getMessage(), 30);
|
464
|
|
465
|
} finally {
|
466
|
DBConnectionPool.returnDBConnection(dbConn, serialNumber);
|
467
|
}//finally
|
468
|
}
|
469
|
|
470
|
/*
|
471
|
* method to make sure insert is finished before create index table If new
|
472
|
* version of record is in xml_documents every thing will be fine
|
473
|
*/
|
474
|
private void checkDocumentTable() throws Exception
|
475
|
{
|
476
|
|
477
|
DBConnection dbConn = null;
|
478
|
int serialNumber = -1;
|
479
|
|
480
|
try {
|
481
|
// Opening separate db connection for writing XML Index
|
482
|
dbConn = DBConnectionPool
|
483
|
.getDBConnection("DBSAXHandler.checkDocumentTable");
|
484
|
serialNumber = dbConn.getCheckOutSerialNumber();
|
485
|
|
486
|
// the following while loop construct checks to make sure that
|
487
|
// the docid of the document that we are trying to index is already
|
488
|
// in the xml_documents table. if this is not the case, the foreign
|
489
|
// key relationship between xml_documents and xml_index is
|
490
|
// temporarily broken causing multiple problems.
|
491
|
boolean inxmldoc = false;
|
492
|
long startTime = System.currentTimeMillis();
|
493
|
while (!inxmldoc) {
|
494
|
String xmlDocumentsCheck = "select distinct docid from xml_documents"
|
495
|
+ " where docid ='"
|
496
|
+ docid
|
497
|
+ "' and "
|
498
|
+ " rev ='"
|
499
|
+ revision + "'";
|
500
|
|
501
|
PreparedStatement xmlDocCheck = dbConn
|
502
|
.prepareStatement(xmlDocumentsCheck);
|
503
|
// Increase usage count
|
504
|
dbConn.increaseUsageCount(1);
|
505
|
xmlDocCheck.execute();
|
506
|
ResultSet doccheckRS = xmlDocCheck.getResultSet();
|
507
|
boolean tableHasRows = doccheckRS.next();
|
508
|
if (tableHasRows) {
|
509
|
MetaCatUtil.debugMessage(
|
510
|
"=========== find the correct document", 35);
|
511
|
inxmldoc = true;
|
512
|
}
|
513
|
doccheckRS.close();
|
514
|
xmlDocCheck.close();
|
515
|
// make sure the while loop will be ended in reseaonable time
|
516
|
long stopTime = System.currentTimeMillis();
|
517
|
if ((stopTime - startTime) > INDEXDELAY) { throw new Exception(
|
518
|
"Couldn't find the docid for index build in"
|
519
|
+ "reseaonable time!"); }
|
520
|
}//while
|
521
|
} catch (Exception e) {
|
522
|
try {
|
523
|
dbConn.rollback();
|
524
|
//dbconn.close();
|
525
|
} catch (SQLException sqle) {
|
526
|
}
|
527
|
MetaCatUtil.debugMessage("Error in DBSAXHandler.run "
|
528
|
+ e.getMessage(), 30);
|
529
|
|
530
|
} finally {
|
531
|
DBConnectionPool.returnDBConnection(dbConn, serialNumber);
|
532
|
}//finally
|
533
|
|
534
|
}
|
535
|
|
536
|
/** SAX Handler that is called for each XML text node */
|
537
|
public void characters(char[] cbuf, int start, int len) throws SAXException
|
538
|
{
|
539
|
MetaCatUtil.debugMessage("CHARACTERS", 50);
|
540
|
// buffer all text nodes for same element. This is for text was splited
|
541
|
// into different nodes
|
542
|
textBuffer.append(new String(cbuf, start, len));
|
543
|
// set hittextnode true
|
544
|
hitTextNode = true;
|
545
|
// if text buffer .size is greater than max, write it to db.
|
546
|
// so we can save memory
|
547
|
if (textBuffer.length() > MAXDATACHARS) {
|
548
|
MetaCatUtil.debugMessage("Write text into DB in charaters"
|
549
|
+ " when text buffer size is greater than maxmum number",
|
550
|
50);
|
551
|
DBSAXNode currentNode = (DBSAXNode) nodeStack.peek();
|
552
|
endNodeId = writeTextForDBSAXNode(endNodeId, textBuffer,
|
553
|
currentNode);
|
554
|
textBuffer = null;
|
555
|
textBuffer = new StringBuffer();
|
556
|
}
|
557
|
}
|
558
|
|
559
|
/**
|
560
|
* SAX Handler that is called for each XML text node that is Ignorable
|
561
|
* white space
|
562
|
*/
|
563
|
public void ignorableWhitespace(char[] cbuf, int start, int len)
|
564
|
throws SAXException
|
565
|
{
|
566
|
// When validation is turned "on", white spaces are reported here
|
567
|
// When validation is turned "off" white spaces are not reported here,
|
568
|
// but through characters() callback
|
569
|
MetaCatUtil.debugMessage("IGNORABLEWHITESPACE", 50);
|
570
|
|
571
|
DBSAXNode currentNode = (DBSAXNode) nodeStack.peek();
|
572
|
String data = null;
|
573
|
int leftover = len;
|
574
|
int offset = start;
|
575
|
boolean moredata = true;
|
576
|
|
577
|
// This loop deals with the case where there are more characters
|
578
|
// than can fit in a single database text field (limit is
|
579
|
// MAXDATACHARS). If the text to be inserted exceeds MAXDATACHARS,
|
580
|
// write a series of nodes that are MAXDATACHARS long, and then the
|
581
|
// final node contains the remainder
|
582
|
while (moredata) {
|
583
|
if (leftover > MAXDATACHARS) {
|
584
|
data = new String(cbuf, offset, MAXDATACHARS);
|
585
|
leftover -= MAXDATACHARS;
|
586
|
offset += MAXDATACHARS;
|
587
|
} else {
|
588
|
data = new String(cbuf, offset, leftover);
|
589
|
moredata = false;
|
590
|
}
|
591
|
|
592
|
// Write the content of the node to the database
|
593
|
endNodeId = currentNode.writeChildNodeToDB("TEXT", null, data,
|
594
|
docid);
|
595
|
}
|
596
|
}
|
597
|
|
598
|
/**
|
599
|
* SAX Handler called once for each processing instruction found: node that
|
600
|
* PI may occur before or after the root element.
|
601
|
*/
|
602
|
public void processingInstruction(String target, String data)
|
603
|
throws SAXException
|
604
|
{
|
605
|
MetaCatUtil.debugMessage("PI", 50);
|
606
|
DBSAXNode currentNode = (DBSAXNode) nodeStack.peek();
|
607
|
endNodeId = currentNode.writeChildNodeToDB("PI", target, data, docid);
|
608
|
}
|
609
|
|
610
|
/** SAX Handler that is called at the end of each XML element */
|
611
|
public void endElement(String uri, String localName, String qName)
|
612
|
throws SAXException
|
613
|
{
|
614
|
MetaCatUtil.debugMessage("End ELEMENT " + qName, 50);
|
615
|
|
616
|
// write buffered text nodes into db (so no splited)
|
617
|
DBSAXNode currentNode = (DBSAXNode) nodeStack.peek();
|
618
|
|
619
|
// If before the end element, the parser hit text nodes and store them
|
620
|
// into the buffer, write the buffer to data base. The reason we put
|
621
|
// write database here is for xerces some time split text node
|
622
|
if (hitTextNode) {
|
623
|
MetaCatUtil.debugMessage("Write text into DB in End Element", 50);
|
624
|
endNodeId = writeTextForDBSAXNode(endNodeId, textBuffer,
|
625
|
currentNode);
|
626
|
|
627
|
//if it is triple parsing process
|
628
|
if (startParseTriple) {
|
629
|
|
630
|
String content = textBuffer.toString().trim();
|
631
|
if (localName.equals("subject")) { //get the subject content
|
632
|
currentTriple.setSubject(content);
|
633
|
} else if (localName.equals("relationship")) { //get the
|
634
|
// relationship
|
635
|
// content
|
636
|
currentTriple.setRelationship(content);
|
637
|
} else if (localName.equals("object")) { //get the object
|
638
|
// content
|
639
|
currentTriple.setObject(content);
|
640
|
}
|
641
|
}
|
642
|
|
643
|
}//if
|
644
|
|
645
|
//set hitText false
|
646
|
hitTextNode = false;
|
647
|
// reset textbuff
|
648
|
textBuffer = null;
|
649
|
textBuffer = new StringBuffer();
|
650
|
|
651
|
// Get the node from the stack
|
652
|
currentNode = (DBSAXNode) nodeStack.pop();
|
653
|
//finishing parsing single triple
|
654
|
if (startParseTriple && localName.equals("triple")) {
|
655
|
// add trip to triple collection
|
656
|
tripleList.addTriple(currentTriple);
|
657
|
//rest variable
|
658
|
currentTriple = null;
|
659
|
startParseTriple = false;
|
660
|
}
|
661
|
}
|
662
|
|
663
|
//
|
664
|
// the next section implements the LexicalHandler interface
|
665
|
//
|
666
|
|
667
|
/** SAX Handler that receives notification of DOCTYPE. Sets the DTD */
|
668
|
public void startDTD(String name, String publicId, String systemId)
|
669
|
throws SAXException
|
670
|
{
|
671
|
docname = name;
|
672
|
doctype = publicId;
|
673
|
systemid = systemId;
|
674
|
|
675
|
processingDTD = true;
|
676
|
DBSAXNode currentNode = (DBSAXNode) nodeStack.peek();
|
677
|
//create a DTD node and write docname,publicid and system id into db
|
678
|
// we don't put the dtd node into node stack
|
679
|
DBSAXNode dtdNode = new DBSAXNode(connection, name, publicId, systemId,
|
680
|
currentNode, currentNode.getRootNodeID(), docid);
|
681
|
MetaCatUtil.debugMessage("Start DTD", 50);
|
682
|
MetaCatUtil.debugMessage("Setting processingDTD to true", 50);
|
683
|
MetaCatUtil.debugMessage("DOCNAME: " + docname, 50);
|
684
|
MetaCatUtil.debugMessage("DOCTYPE: " + doctype, 50);
|
685
|
MetaCatUtil.debugMessage(" SYSID: " + systemid, 50);
|
686
|
}
|
687
|
|
688
|
/**
|
689
|
* SAX Handler that receives notification of end of DTD
|
690
|
*/
|
691
|
public void endDTD() throws SAXException
|
692
|
{
|
693
|
|
694
|
processingDTD = false;
|
695
|
MetaCatUtil.debugMessage("Setting processingDTD to false", 50);
|
696
|
MetaCatUtil.debugMessage("end DTD", 50);
|
697
|
}
|
698
|
|
699
|
/**
|
700
|
* SAX Handler that receives notification of comments in the DTD
|
701
|
*/
|
702
|
public void comment(char[] ch, int start, int length) throws SAXException
|
703
|
{
|
704
|
MetaCatUtil.debugMessage("COMMENT", 50);
|
705
|
if (!processingDTD) {
|
706
|
DBSAXNode currentNode = (DBSAXNode) nodeStack.peek();
|
707
|
endNodeId = currentNode.writeChildNodeToDB("COMMENT", null,
|
708
|
new String(ch, start, length), docid);
|
709
|
}
|
710
|
}
|
711
|
|
712
|
/**
|
713
|
* SAX Handler that receives notification of the start of CDATA sections
|
714
|
*/
|
715
|
public void startCDATA() throws SAXException
|
716
|
{
|
717
|
MetaCatUtil.debugMessage("start CDATA", 50);
|
718
|
}
|
719
|
|
720
|
/**
|
721
|
* SAX Handler that receives notification of the end of CDATA sections
|
722
|
*/
|
723
|
public void endCDATA() throws SAXException
|
724
|
{
|
725
|
MetaCatUtil.debugMessage("end CDATA", 50);
|
726
|
}
|
727
|
|
728
|
/**
|
729
|
* SAX Handler that receives notification of the start of entities
|
730
|
*/
|
731
|
public void startEntity(String name) throws SAXException
|
732
|
{
|
733
|
MetaCatUtil.debugMessage("start ENTITY: " + name, 50);
|
734
|
//System.out.println("start ENTITY: " + name);
|
735
|
if (name.equals("[dtd]")) {
|
736
|
processingDTD = true;
|
737
|
}
|
738
|
}
|
739
|
|
740
|
/**
|
741
|
* SAX Handler that receives notification of the end of entities
|
742
|
*/
|
743
|
public void endEntity(String name) throws SAXException
|
744
|
{
|
745
|
MetaCatUtil.debugMessage("end ENTITY: " + name, 50);
|
746
|
//System.out.println("end ENTITY: " + name);
|
747
|
if (name.equals("[dtd]")) {
|
748
|
processingDTD = false;
|
749
|
}
|
750
|
}
|
751
|
|
752
|
/**
|
753
|
* SAX Handler that receives notification of element declarations
|
754
|
*/
|
755
|
public void elementDecl(String name, String model)
|
756
|
throws org.xml.sax.SAXException
|
757
|
{
|
758
|
//System.out.println("ELEMENTDECL: " + name + " " + model);
|
759
|
MetaCatUtil.debugMessage("ELEMENTDECL: " + name + " " + model, 50);
|
760
|
}
|
761
|
|
762
|
/**
|
763
|
* SAX Handler that receives notification of attribute declarations
|
764
|
*/
|
765
|
public void attributeDecl(String eName, String aName, String type,
|
766
|
String valueDefault, String value) throws org.xml.sax.SAXException
|
767
|
{
|
768
|
|
769
|
//System.out.println("ATTRIBUTEDECL: " + eName + " "
|
770
|
// + aName + " " + type + " " + valueDefault + " "
|
771
|
// + value);
|
772
|
MetaCatUtil.debugMessage("ATTRIBUTEDECL: " + eName + " " + aName + " "
|
773
|
+ type + " " + valueDefault + " " + value, 50);
|
774
|
}
|
775
|
|
776
|
/**
|
777
|
* SAX Handler that receives notification of internal entity declarations
|
778
|
*/
|
779
|
public void internalEntityDecl(String name, String value)
|
780
|
throws org.xml.sax.SAXException
|
781
|
{
|
782
|
//System.out.println("INTERNENTITYDECL: " + name + " " + value);
|
783
|
MetaCatUtil.debugMessage("INTERNENTITYDECL: " + name + " " + value, 50);
|
784
|
}
|
785
|
|
786
|
/**
|
787
|
* SAX Handler that receives notification of external entity declarations
|
788
|
*/
|
789
|
public void externalEntityDecl(String name, String publicId, String systemId)
|
790
|
throws org.xml.sax.SAXException
|
791
|
{
|
792
|
//System.out.println("EXTERNENTITYDECL: " + name + " " + publicId
|
793
|
// + " " + systemId);
|
794
|
MetaCatUtil.debugMessage("EXTERNENTITYDECL: " + name + " " + publicId
|
795
|
+ " " + systemId, 50);
|
796
|
// it processes other external entity, not the DTD;
|
797
|
// it doesn't signal for the DTD here
|
798
|
processingDTD = false;
|
799
|
}
|
800
|
|
801
|
//
|
802
|
// the next section implements the ErrorHandler interface
|
803
|
//
|
804
|
|
805
|
/**
|
806
|
* SAX Handler that receives notification of fatal parsing errors
|
807
|
*/
|
808
|
public void fatalError(SAXParseException exception) throws SAXException
|
809
|
{
|
810
|
MetaCatUtil.debugMessage("FATALERROR: " + exception.getMessage(), 50);
|
811
|
throw (new SAXException("Fatal processing error.", exception));
|
812
|
}
|
813
|
|
814
|
/**
|
815
|
* SAX Handler that receives notification of recoverable parsing errors
|
816
|
*/
|
817
|
public void error(SAXParseException exception) throws SAXException
|
818
|
{
|
819
|
MetaCatUtil.debugMessage("ERROR: " + exception.getMessage(), 50);
|
820
|
throw (new SAXException("Processing error.", exception));
|
821
|
}
|
822
|
|
823
|
/**
|
824
|
* SAX Handler that receives notification of warnings
|
825
|
*/
|
826
|
public void warning(SAXParseException exception) throws SAXException
|
827
|
{
|
828
|
MetaCatUtil.debugMessage("WARNING: " + exception.getMessage(), 50);
|
829
|
throw (new SAXException("Warning.", exception));
|
830
|
}
|
831
|
|
832
|
//
|
833
|
// Helper, getter and setter methods
|
834
|
//
|
835
|
|
836
|
/**
|
837
|
* get the document name
|
838
|
*/
|
839
|
public String getDocname()
|
840
|
{
|
841
|
return docname;
|
842
|
}
|
843
|
|
844
|
/**
|
845
|
* get the document processing state
|
846
|
*/
|
847
|
public boolean processingDTD()
|
848
|
{
|
849
|
return processingDTD;
|
850
|
}
|
851
|
|
852
|
/* Method to write a text buffer for DBSAXNode */
|
853
|
protected long writeTextForDBSAXNode(long previousEndNodeId,
|
854
|
StringBuffer strBuffer, DBSAXNode node) throws SAXException
|
855
|
{
|
856
|
long nodeId = previousEndNodeId;
|
857
|
// Check parameter
|
858
|
if (strBuffer == null || node == null) { return nodeId; }
|
859
|
boolean moredata = true;
|
860
|
String data = null;
|
861
|
int bufferSize = strBuffer.length();
|
862
|
int start = 0;
|
863
|
|
864
|
// if there are some cotent in buffer, write it
|
865
|
if (bufferSize > 0) {
|
866
|
MetaCatUtil.debugMessage("Write text into DB", 50);
|
867
|
// This loop deals with the case where there are more characters
|
868
|
// than can fit in a single database text field (limit is
|
869
|
// MAXDATACHARS). If the text to be inserted exceeds MAXDATACHARS,
|
870
|
// write a series of nodes that are MAXDATACHARS long, and then the
|
871
|
// final node contains the remainder
|
872
|
while (moredata) {
|
873
|
bufferSize = strBuffer.length();
|
874
|
if (bufferSize > MAXDATACHARS) {
|
875
|
data = strBuffer.substring(start, MAXDATACHARS);
|
876
|
// cut the stringbuffer part that already written into db
|
877
|
strBuffer = strBuffer.delete(start, MAXDATACHARS);
|
878
|
} else {
|
879
|
data = strBuffer.substring(start, bufferSize);
|
880
|
moredata = false;
|
881
|
}
|
882
|
|
883
|
// Write the content of the node to the database
|
884
|
nodeId = node.writeChildNodeToDB("TEXT", null, data, docid);
|
885
|
}//while
|
886
|
}//if
|
887
|
return nodeId;
|
888
|
}
|
889
|
}
|