Project

General

Profile

1 72 bojilova
/**
2 203 jones
 *  '$RCSfile$'
3
 *    Purpose: A Class that implements org.xml.sax.EntityResolver interface
4
 *             for resolving external entities
5
 *  Copyright: 2000 Regents of the University of California and the
6
 *             National Center for Ecological Analysis and Synthesis
7 243 jones
 *    Authors: Jivka Bojilova, Matt Jones
8 72 bojilova
 *
9 203 jones
 *   '$Author$'
10
 *     '$Date$'
11
 * '$Revision$'
12 669 jones
 *
13
 * This program is free software; you can redistribute it and/or modify
14
 * it under the terms of the GNU General Public License as published by
15
 * the Free Software Foundation; either version 2 of the License, or
16
 * (at your option) any later version.
17
 *
18
 * This program is distributed in the hope that it will be useful,
19
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
20
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
21
 * GNU General Public License for more details.
22
 *
23
 * You should have received a copy of the GNU General Public License
24
 * along with this program; if not, write to the Free Software
25
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
26 72 bojilova
 */
27
28 75 jones
package edu.ucsb.nceas.metacat;
29 72 bojilova
30 2663 sgarg
import org.apache.log4j.Logger;
31 72 bojilova
import org.xml.sax.*;
32 598 bojilova
import org.xml.sax.helpers.DefaultHandler;
33 72 bojilova
34 5090 daigle
import edu.ucsb.nceas.metacat.accesscontrol.AccessControlList;
35 5015 daigle
import edu.ucsb.nceas.metacat.database.DBConnection;
36
import edu.ucsb.nceas.metacat.database.DBConnectionPool;
37 4080 daigle
import edu.ucsb.nceas.metacat.util.SystemUtil;
38
import edu.ucsb.nceas.utilities.PropertyNotFoundException;
39
40 72 bojilova
import java.sql.*;
41 243 jones
import java.io.File;
42 598 bojilova
import java.io.Reader;
43
import java.io.BufferedReader;
44 694 bojilova
import java.io.BufferedInputStream;
45 598 bojilova
import java.io.FileWriter;
46
import java.io.BufferedWriter;
47 243 jones
import java.io.InputStream;
48
import java.io.IOException;
49 72 bojilova
import java.net.URL;
50
import java.net.MalformedURLException;
51
52 1358 tao
/**
53
 * A database aware Class implementing EntityResolver interface for the SAX
54
 * parser to call when processing the XML stream and intercepting any
55
 * external entities (including the external DTD subset and external
56 122 jones
 * parameter entities, if any) before including them.
57 72 bojilova
 */
58
public class DBEntityResolver implements EntityResolver
59
{
60 1217 tao
  private DBConnection connection = null;
61 598 bojilova
  private DefaultHandler handler = null;
62
  private String docname = null;
63
  private String doctype = null;
64
  private String systemid = null;
65
  private Reader dtdtext = null;
66 2663 sgarg
  private static Logger logMetacat = Logger.getLogger(DBEntityResolver.class);
67
68 1358 tao
  /**
69 598 bojilova
   * Construct an instance of the DBEntityResolver class
70
   *
71
   * @param conn the JDBC connection to which information is written
72 599 bojilova
   */
73 1217 tao
  public DBEntityResolver(DBConnection conn)
74 599 bojilova
  {
75 1217 tao
    this.connection= conn;
76 599 bojilova
  }
77 1358 tao
  /**
78 599 bojilova
   * Construct an instance of the DBEntityResolver class
79
   *
80
   * @param conn the JDBC connection to which information is written
81 598 bojilova
   * @param handler the SAX handler to determine parsing context
82
   * @param dtd Reader of new dtd to be uploaded on server's file system
83
   */
84 1217 tao
  public DBEntityResolver(DBConnection conn, DefaultHandler handler, Reader dtd)
85 598 bojilova
  {
86 1217 tao
    this.connection = conn;
87 598 bojilova
    this.handler = handler;
88
    this.dtdtext = dtd;
89
  }
90 1358 tao
91
  /**
92
   * The Parser call this method before opening any external entity
93 598 bojilova
   * except the top-level document entity (including the external DTD subset,
94 1358 tao
   * external entities referenced within the DTD, and external entities
95 598 bojilova
   * referenced within the document element)
96
   */
97
  public InputSource resolveEntity (String publicId, String systemId)
98
                     throws SAXException
99
  {
100 5177 daigle
    logMetacat.debug("DBEntityResolver.resolveEntity - in resolveEntity");
101 598 bojilova
    String dbSystemID;
102
    String doctype = null;
103 1358 tao
104 598 bojilova
    // Won't have a handler under all cases
105
    if ( handler != null ) {
106
      if ( handler instanceof DBSAXHandler ) {
107
        DBSAXHandler dhandler = null;
108
        dhandler = (DBSAXHandler)handler;
109
        if ( dhandler.processingDTD() ) {
110 1363 tao
111 598 bojilova
          // public ID is doctype
112 1358 tao
          if (publicId != null) {
113 598 bojilova
            doctype = publicId;
114 5177 daigle
            logMetacat.debug("DBEntityResolver.resolveEntity - in get type from publicId and doctype is: "
115
                                     + doctype);
116 598 bojilova
          // assume public ID (doctype) is docname
117
          } else if (systemId != null) {
118
            doctype = dhandler.getDocname();
119 72 bojilova
          }
120
        }
121 598 bojilova
      } else if ( handler instanceof AccessControlList ) {
122
        AccessControlList ahandler = null;
123
        ahandler = (AccessControlList)handler;
124
        //if ( ahandler.processingDTD() ) {
125
          // public ID is doctype
126 1358 tao
          if (publicId != null) {
127 598 bojilova
            doctype = publicId;
128
          // assume public ID (doctype) is docname
129
          } else if (systemId != null) {
130
            doctype = ahandler.getDocname();
131
          }
132
        //}
133
      }
134
    }
135 72 bojilova
136 598 bojilova
    // get System ID for doctype
137
    if (doctype != null) {
138 1358 tao
      // look at db XML Catalog for System ID
139 5177 daigle
      logMetacat.info("DBEntityResolver.resolveEntity - get systemId from doctype: " + doctype);
140 598 bojilova
      dbSystemID = getDTDSystemID(doctype);
141 5177 daigle
      logMetacat.info("DBEntityResolver.resolveEntity - The Systemid is: " + dbSystemID);
142 4803 leinfelder
      // check that it is accessible on our system before getting too far
143
      try {
144
    	  InputStream in = checkURLConnection(dbSystemID);
145 5177 daigle
	  } catch (SAXException se) {
146 4978 daigle
		  // after an upgrade, the dtd will not exist on disk, but it is in xml catalog.  The db system id may be pointing
147
		  // back at this system  Try and download it from the original system id and see if we still have a problem
148
		  // checking the URL connection.
149 5177 daigle
		  logMetacat.warn("DBEntityResolver.resolveEntity - Problem when checking URL Connection: " + se.getMessage());
150
		  logMetacat.warn("DBEntityResolver.resolveEntity - Probably, dtd for doc type " + doctype + " existed in xml catalog, but not on disk.  Uploading from: " + systemId);
151 4978 daigle
		  InputStream istream = checkURLConnection(systemId);
152
		  uploadDTDFromURL(istream, systemId);
153
		  try {
154 5177 daigle
			  Thread.currentThread().sleep(6000);
155 4978 daigle
			  checkURLConnection(dbSystemID);
156
		  } catch (Exception e2) {
157 5177 daigle
			  logMetacat.error("DBEntityResolver.resolveEntity - still could not find dtd for doc type " + doctype + " at "
158 4978 daigle
					  + dbSystemID + " : " + e2.getMessage());
159
			  dbSystemID = null;
160
		  }
161 5177 daigle
	  }
162 598 bojilova
      boolean doctypeIsInDB = true;
163
      // no System ID found in db XML Catalog
164
      if (dbSystemID == null) {
165
        doctypeIsInDB = false;
166
        // use the provided System ID
167
        if (systemId != null) {
168
          dbSystemID = systemId;
169
        }
170 5177 daigle
        logMetacat.info("DBEntityResolver.resolveEntity - If above Systemid is null, then get "
171
                                 + "system id from file: " + dbSystemID);
172 598 bojilova
      }
173
      // there are dtd text provided; try to upload on Metacat
174
      if ( dtdtext != null ) {
175
        dbSystemID = uploadDTD(dbSystemID);
176
      }
177
178 694 bojilova
      // open URLConnection to check first
179 598 bojilova
      InputStream istream = checkURLConnection(dbSystemID);
180
181
      // need to register System ID in db XML Catalog if not yet
182
      if ( !doctypeIsInDB ) {
183 694 bojilova
        // new DTD from outside URL location; try to upload on Metacat
184
        if ( dtdtext == null ) {
185
          dbSystemID = uploadDTDFromURL(istream, dbSystemID);
186
        }
187 598 bojilova
        registerDTD(doctype, dbSystemID);
188
      }
189
      // return a byte-input stream for use
190 1358 tao
      InputSource is = new InputSource(dbSystemID);
191 694 bojilova
192
      // close and open URLConnection again
193
      try {
194
        istream.close();
195
      } catch (IOException e) {
196 1358 tao
        throw new SAXException
197 4967 daigle
        ("DBEntityResolver.resolveEntity - I/O issue when resolving entity: " + e.getMessage());
198 1358 tao
      }
199 694 bojilova
      istream = checkURLConnection(dbSystemID);
200 598 bojilova
      is.setByteStream(istream);
201
      return is;
202
    } else {
203 694 bojilova
      // use provided systemId for the other cases
204 5177 daigle
      logMetacat.info("DBEntityResolver.resolveEntity - doctype is null and using system id from file");
205 598 bojilova
      InputStream istream = checkURLConnection(systemId);
206
      return null;
207 1358 tao
208 598 bojilova
    }
209 1358 tao
210 598 bojilova
  }
211
212 1358 tao
  /**
213 598 bojilova
   * Look at db XML Catalog to get System ID (if any) for @doctype.
214
   * Return null if there are no System ID found for @doctype
215
   */
216 1895 tao
  public static String getDTDSystemID( String doctype )
217 598 bojilova
                 throws SAXException
218
  {
219
    String systemid = null;
220 1217 tao
    Statement stmt = null;
221
    DBConnection conn = null;
222
    int serialNumber = -1;
223 598 bojilova
    try {
224 1217 tao
      //check out DBConnection
225
      conn=DBConnectionPool.getDBConnection("DBEntityResolver.getDTDSystemID");
226
      serialNumber=conn.getCheckOutSerialNumber();
227 1358 tao
228 598 bojilova
      stmt = conn.createStatement();
229 1358 tao
      stmt.execute("SELECT system_id FROM xml_catalog " +
230 598 bojilova
                   "WHERE entry_type = 'DTD' AND public_id = '" +
231
                   doctype + "'");
232
      ResultSet rs = stmt.getResultSet();
233
      boolean tableHasRows = rs.next();
234
      if (tableHasRows) {
235
        systemid = rs.getString(1);
236 4080 daigle
        // system id may not have server url on front.  Add it if not.
237
        if (!systemid.startsWith("http://")) {
238 4123 daigle
        	systemid = SystemUtil.getContextURL() + systemid;
239 4080 daigle
        }
240 598 bojilova
      }
241
      stmt.close();
242
    } catch (SQLException e) {
243
      throw new SAXException
244 4967 daigle
      ("DBEntityResolver.getDTDSystemID - SQL error when getting DTD system ID: " + e.getMessage());
245 4080 daigle
    } catch (PropertyNotFoundException pnfe) {
246
        throw new SAXException
247 4967 daigle
        ("DBEntityResolver.getDTDSystemID - Property error when getting DTD system ID:  " + pnfe.getMessage());
248 4080 daigle
      }
249 1217 tao
    finally
250
    {
251
      try
252
      {
253
        stmt.close();
254
      }//try
255
      catch (SQLException sqlE)
256
      {
257 5177 daigle
        logMetacat.error("DBEntityResolver.getDTDSystemId - SQL error: " + sqlE.getMessage());
258 1217 tao
      }//catch
259
      finally
260
      {
261
        DBConnectionPool.returnDBConnection(conn, serialNumber);
262
      }//finally
263
    }//finally
264 598 bojilova
265
    // return the selected System ID
266
    return systemid;
267
  }
268
269 1358 tao
  /**
270
   * Register new DTD identified by @systemId in Metacat XML Catalog
271 598 bojilova
   * . make a reference with @systemId for @doctype in Metacat DB
272
   */
273
  private void registerDTD ( String doctype, String systemId )
274
                 throws SAXException
275
  {
276 4803 leinfelder
	  String existingSystemId = getDTDSystemID(doctype);
277
	  if (existingSystemId != null && existingSystemId.equals(systemId)) {
278 5177 daigle
		  logMetacat.warn("DBEntityResolver.registerDTD - doctype/systemId already registered in DB: " + doctype);
279 4803 leinfelder
		  return;
280
	  }
281 1217 tao
    //DBConnection conn = null;
282
    //int serialNumber = -1;
283
    PreparedStatement pstmt = null;
284 598 bojilova
    // make a reference in db catalog table with @systemId for @doctype
285
    try {
286 1217 tao
      //check out DBConnection
287
      //conn=DBConnectionPool.getDBConnection("DBEntityResolver.registerDTD");
288
      //serialNumber=conn.getCheckOutSerialNumber();
289 1358 tao
290
291 1217 tao
      pstmt = connection.prepareStatement(
292 243 jones
             "INSERT INTO xml_catalog " +
293 4800 leinfelder
             "(entry_type, public_id, system_id) " +
294
             "VALUES ('DTD', ?, ?)");
295 1217 tao
      // Increase usage count
296
      connection.increaseUsageCount(1);
297 598 bojilova
      // Bind the values to the query
298
      pstmt.setString(1, doctype);
299
      pstmt.setString(2, systemId);
300
      // Do the insertion
301
      pstmt.execute();
302 4803 leinfelder
      int updateCnt = pstmt.getUpdateCount();
303 5177 daigle
      logMetacat.debug("DBEntityReolver.registerDTD - DTDs registered: " + updateCnt);
304 598 bojilova
      pstmt.close();
305
    } catch (SQLException e) {
306
      throw new SAXException
307 4967 daigle
      ("DBEntityResolver.registerDTD - SQL issue when registering DTD: " + e.getMessage());
308 598 bojilova
    }
309 1217 tao
    finally
310
    {
311
      try
312
      {
313
        pstmt.close();
314
      }//try
315
      catch (SQLException sqlE)
316
      {
317 5177 daigle
        logMetacat.error("DBEntityResolver.registerDTD - SQL error: " + sqlE.getMessage());
318 1217 tao
      }//catch
319
      //DBConnectionPool.returnDBConnection(conn, serialNumber);
320
    }//finally
321 1358 tao
322 598 bojilova
  }
323 243 jones
324 1358 tao
  /**
325 4080 daigle
	 * Upload new DTD text identified by
326
	 *
327
	 * @systemId to Metacat file system
328
	 */
329
	private String uploadDTD(String systemId) throws SAXException {
330
		String dtdPath = null;
331
		String dtdURL = null;
332
		try {
333
			dtdPath = SystemUtil.getContextDir() + "/dtd/";
334
			dtdURL = SystemUtil.getContextURL() + "/dtd/";
335
		} catch (PropertyNotFoundException pnfe) {
336
			throw new SAXException("DBEntityResolver.uploadDTD: " + pnfe.getMessage());
337
		}
338 1358 tao
339 4080 daigle
		// get filename from systemId
340
		String filename = systemId;
341
		int slash = Math.max(filename.lastIndexOf('/'), filename.lastIndexOf('\\'));
342
		if (slash > -1) {
343
			filename = filename.substring(slash + 1);
344
		}
345 245 jones
346 4080 daigle
		// writing dtd text on Metacat file system as filename
347
		try {
348
			// create a buffering character-input stream
349
			// that uses a default-sized input buffer
350
			BufferedReader in = new BufferedReader(dtdtext);
351 245 jones
352 4080 daigle
			// open file writer to write the input into it
353
			// String dtdPath = "/opt/tomcat/webapps/bojilova/dtd/";
354
			File f = new File(dtdPath, filename);
355
			synchronized (f) {
356
				try {
357
					if (f.exists()) {
358
						throw new IOException("File already exist: "
359
								+ f.getCanonicalFile());
360
						// if ( f.exists() && !f.canWrite() ) {
361
						// throw new IOException("Not writable: " +
362
						// f.getCanonicalFile());
363
					}
364
				} catch (SecurityException se) {
365
					// if a security manager exists,
366
					// its checkRead method is called for f.exist()
367
					// or checkWrite method is called for f.canWrite()
368
					throw se;
369
				}
370
				// create a buffered character-output stream
371
				// that uses a default-sized output buffer
372
				FileWriter fw = new FileWriter(f);
373
				BufferedWriter out = new BufferedWriter(fw);
374 598 bojilova
375 4080 daigle
				// read the input and write into the file writer
376
				String inputLine;
377
				while ((inputLine = in.readLine()) != null) {
378
					out.write(inputLine, 0, inputLine.length());
379
					out.newLine(); // instead of out.write('\r\n');
380
				}
381 598 bojilova
382 4080 daigle
				// the input and the output streams must be closed
383
				in.close();
384
				out.flush();
385
				out.close();
386
				fw.close();
387
			} // end of synchronized
388
		} catch (MalformedURLException e) {
389 4967 daigle
			throw new SAXException("DBEntityResolver.uploadDTD() - Malformed URL when uploading DTD: " + e.getMessage());
390 4080 daigle
		} catch (IOException e) {
391 4967 daigle
			throw new SAXException("DBEntityResolver.uploadDTD - I/O issue when uploading DTD: " + e.getMessage());
392 4080 daigle
		} catch (SecurityException e) {
393 4967 daigle
			throw new SAXException("DBEntityResolver.uploadDTD() - Security issue when uploading DTD: " + e.getMessage());
394 4080 daigle
		}
395 1358 tao
396 4080 daigle
		// String dtdURL = "http://dev.nceas.ucsb.edu/bojilova/dtd/";
397
		return dtdURL + filename;
398
	}
399 598 bojilova
400 694 bojilova
401 1358 tao
  /**
402 4080 daigle
	 * Upload new DTD located at outside URL to Metacat file system
403
	 */
404
	private String uploadDTDFromURL(InputStream istream, String systemId)
405
			throws SAXException {
406
		String dtdPath = null;
407
		String dtdURL = null;
408
		try {
409
			dtdPath = SystemUtil.getContextDir() + "/dtd/";
410
			dtdURL = SystemUtil.getContextURL() + "/dtd/";
411
		} catch (PropertyNotFoundException pnfe) {
412 4967 daigle
			throw new SAXException("DBEntityResolver.uploadDTDFromURL - Property issue when uploading DTD from URL: "
413 4080 daigle
					+ pnfe.getMessage());
414
		}
415 1358 tao
416 4080 daigle
		// get filename from systemId
417
		String filename = systemId;
418
		int slash = Math.max(filename.lastIndexOf('/'), filename.lastIndexOf('\\'));
419
		if (slash > -1) {
420
			filename = filename.substring(slash + 1);
421
		}
422 694 bojilova
423 4080 daigle
		// writing dtd text on Metacat file system as filename
424
		try {
425
			// create a buffering character-input stream
426
			// that uses a default-sized input buffer
427
			BufferedInputStream in = new BufferedInputStream(istream);
428 694 bojilova
429 4080 daigle
			// open file writer to write the input into it
430
			//String dtdPath = "/opt/tomcat/webapps/bojilova/dtd/";
431
			File f = new File(dtdPath, filename);
432
			synchronized (f) {
433
				try {
434
					if (f.exists()) {
435 5177 daigle
						logMetacat.warn("DBEntityResolver.uploadDTDFromURL - File already exists: " + f.getCanonicalFile());
436 4803 leinfelder
						//return dtdURL + filename;
437
						//throw new IOException("File already exist: "
438
						//		+ f.getCanonicalFile());
439 4080 daigle
						//if ( f.exists() && !f.canWrite() ) {
440
						//  throw new IOException("Not writable: " + f.getCanonicalFile());
441
					}
442
				} catch (SecurityException se) {
443
					// if a security manager exists,
444
					// its checkRead method is called for f.exist()
445
					// or checkWrite method is called for f.canWrite()
446
					throw se;
447
				}
448
				// create a buffered character-output stream
449
				// that uses a default-sized output buffer
450
				FileWriter fw = new FileWriter(f);
451
				BufferedWriter out = new BufferedWriter(fw);
452 694 bojilova
453 4080 daigle
				// read the input and write into the file writer
454
				int inputByte;
455
				while ((inputByte = in.read()) != -1) {
456
					out.write(inputByte);
457
					//out.newLine(); //instead of out.write('\r\n');
458
				}
459 694 bojilova
460 4080 daigle
				// the input and the output streams must be closed
461
				in.close();
462
				out.flush();
463
				out.close();
464
				fw.close();
465
			} // end of synchronized
466
		} catch (MalformedURLException e) {
467 4967 daigle
			throw new SAXException("DBEntityResolver.uploadDTDFromURL - Malformed URL when uploading DTD from URL: "
468 4080 daigle
					+ e.getMessage());
469
		} catch (IOException e) {
470 4967 daigle
			throw new SAXException("DBEntityResolver.uploadDTDFromURL - I/O issue when uploading DTD from URL:  "
471 4080 daigle
					+ e.getMessage());
472
		} catch (SecurityException e) {
473 4967 daigle
			throw new SAXException("DBEntityResolver.uploadDTDFromURL - Security issue when uploading DTD from URL:  "
474 4080 daigle
					+ e.getMessage());
475
		}
476 1358 tao
477 4080 daigle
		//String dtdURL = "http://dev.nceas.ucsb.edu/bojilova/dtd/";
478
		return dtdURL + filename;
479
	}
480 694 bojilova
481 4080 daigle
	/**
482
	 * Check URL Connection for @systemId, and return an InputStream
483
	 * that can be used to read from the systemId URL.  The parser ends
484
	 * up using this via the InputSource to read the DTD.
485
	 *
486
	 * @param systemId a URI (in practice URL) to be checked and opened
487
	 */
488
	public static InputStream checkURLConnection(String systemId) throws SAXException {
489
		try {
490
			return (new URL(systemId).openStream());
491 598 bojilova
492 4080 daigle
		} catch (MalformedURLException e) {
493 4967 daigle
			throw new SAXException("DBEntityResolver.checkURLConnection - Malformed URL when checking URL Connection: "
494 4080 daigle
					+ e.getMessage());
495
		} catch (IOException e) {
496 4967 daigle
			throw new SAXException("DBEntityResolver.checkURLConnection - I/O issue when checking URL Connection: "
497 4080 daigle
					+ e.getMessage());
498
		}
499
	}
500 72 bojilova
}