Revision 6962
Added by ben leinfelder about 12 years ago
| src/edu/ucsb/nceas/metacat/dataone/SystemMetadataFactory.java | ||
|---|---|---|
| 25 | 25 |
*/ |
| 26 | 26 |
package edu.ucsb.nceas.metacat.dataone; |
| 27 | 27 |
|
| 28 |
import java.io.BufferedInputStream; |
|
| 29 |
import java.io.File; |
|
| 28 | 30 |
import java.io.IOException; |
| 29 | 31 |
import java.io.InputStream; |
| 30 | 32 |
import java.math.BigInteger; |
| ... | ... | |
| 78 | 80 |
import edu.ucsb.nceas.metacat.AccessionNumber; |
| 79 | 81 |
import edu.ucsb.nceas.metacat.AccessionNumberException; |
| 80 | 82 |
import edu.ucsb.nceas.metacat.DBUtil; |
| 83 |
import edu.ucsb.nceas.metacat.DocumentImpl; |
|
| 81 | 84 |
import edu.ucsb.nceas.metacat.IdentifierManager; |
| 82 | 85 |
import edu.ucsb.nceas.metacat.McdbDocNotFoundException; |
| 83 | 86 |
import edu.ucsb.nceas.metacat.McdbException; |
| ... | ... | |
| 169 | 172 |
sysMeta.setSerialVersion(BigInteger.valueOf(1)); |
| 170 | 173 |
sysMeta.setArchived(false); |
| 171 | 174 |
} |
| 172 |
|
|
| 173 |
// get the data or metadata object |
|
| 174 |
InputStream inputStream; |
|
| 175 |
try {
|
|
| 176 |
inputStream = MetacatHandler.read(localId); |
|
| 177 |
} catch (ParseLSIDException ple) {
|
|
| 178 |
logMetacat.debug("There was a problem parsing the LSID from "
|
|
| 179 |
+ localId + ". The error message was: " + ple.getMessage()); |
|
| 180 |
throw ple; |
|
| 181 |
|
|
| 182 |
} catch (PropertyNotFoundException pnfe) {
|
|
| 183 |
logMetacat.debug("There was a problem finding a property. "
|
|
| 184 |
+ "The error message was: " + pnfe.getMessage()); |
|
| 185 |
throw pnfe; |
|
| 186 |
|
|
| 187 |
} catch (McdbException me) {
|
|
| 188 |
logMetacat.debug("There was a Metacat problem. "
|
|
| 189 |
+ "The error message was: " + me.getMessage()); |
|
| 190 |
throw me; |
|
| 191 |
|
|
| 192 |
} catch (SQLException sqle) {
|
|
| 193 |
logMetacat.debug("There was a SQL problem. "
|
|
| 194 |
+ "The error message was: " + sqle.getMessage()); |
|
| 195 |
throw sqle; |
|
| 196 |
|
|
| 197 |
} catch (ClassNotFoundException cnfe) {
|
|
| 198 |
logMetacat.debug("There was a problem finding a class. "
|
|
| 199 |
+ "The error message was: " + cnfe.getMessage()); |
|
| 200 |
throw cnfe; |
|
| 201 |
|
|
| 202 |
} catch (IOException ioe) {
|
|
| 203 |
logMetacat.debug("There was an I/O exception. "
|
|
| 204 |
+ "The error message was: " + ioe.getMessage()); |
|
| 205 |
throw ioe; |
|
| 206 |
|
|
| 207 |
} |
|
| 208 |
|
|
| 175 |
|
|
| 209 | 176 |
// get additional docinfo |
| 210 | 177 |
Hashtable<String, String> docInfo = ReplicationService.getDocumentInfoMap(localId); |
| 211 | 178 |
// set the default object format |
| ... | ... | |
| 240 | 207 |
sysMeta.setFormatId(fmtid); |
| 241 | 208 |
logMetacat.debug("The ObjectFormat for " + localId + " is " + fmtid.getValue());
|
| 242 | 209 |
|
| 210 |
// for retrieving the actual object |
|
| 211 |
InputStream inputStream = null; |
|
| 212 |
inputStream = MetacatHandler.read(localId); |
|
| 213 |
|
|
| 243 | 214 |
// create the checksum |
| 244 |
inputStream = MetacatHandler.read(localId); |
|
| 245 | 215 |
String algorithm = "MD5"; |
| 246 | 216 |
Checksum checksum = ChecksumUtil.checksum(inputStream, algorithm); |
| 247 | 217 |
sysMeta.setChecksum(checksum); |
| 248 | 218 |
|
| 249 |
// set the size |
|
| 250 |
inputStream = MetacatHandler.read(localId); |
|
| 251 |
String sizeStr = new Long(sizeOfStream(inputStream)).toString(); |
|
| 252 |
sysMeta.setSize(new BigInteger(sizeStr)); |
|
| 219 |
// set the size from file on disk, don't read bytes again |
|
| 220 |
File fileOnDisk = getFileOnDisk(localId); |
|
| 221 |
long fileSize = 0; |
|
| 222 |
if (fileOnDisk.exists()) {
|
|
| 223 |
fileSize = fileOnDisk.length(); |
|
| 224 |
} |
|
| 225 |
sysMeta.setSize(BigInteger.valueOf(fileSize)); |
|
| 253 | 226 |
|
| 254 | 227 |
// submitter |
| 255 | 228 |
Subject submitter = new Subject(); |
| ... | ... | |
| 337 | 310 |
"eml://ecoinformatics.org/eml-2.1.1").getFormatId()) {
|
| 338 | 311 |
|
| 339 | 312 |
try {
|
| 313 |
|
|
| 314 |
// get it again to parse the document |
|
| 315 |
logMetacat.debug("Re-reading document inputStream");
|
|
| 340 | 316 |
inputStream = MetacatHandler.read(localId); |
| 341 | 317 |
|
| 342 | 318 |
DataoneEMLParser emlParser = DataoneEMLParser.getInstance(); |
| ... | ... | |
| 400 | 376 |
} |
| 401 | 377 |
} |
| 402 | 378 |
|
| 403 |
// TODO: any other special processing (csv, images, etc)? |
|
| 404 | 379 |
} else {
|
| 405 | 380 |
// if we don't know what it is, should we skip it? |
| 406 | 381 |
dataObject = null; |
| ... | ... | |
| 717 | 692 |
return size; |
| 718 | 693 |
|
| 719 | 694 |
} |
| 695 |
|
|
| 696 |
private static File getFileOnDisk(String docid) throws McdbException, PropertyNotFoundException {
|
|
| 697 |
|
|
| 698 |
DocumentImpl doc = new DocumentImpl(docid, false); |
|
| 699 |
String filepath = null; |
|
| 700 |
String filename = null; |
|
| 701 |
|
|
| 702 |
// deal with data or metadata cases |
|
| 703 |
if (doc.getRootNodeID() == 0) {
|
|
| 704 |
// this is a data file |
|
| 705 |
filepath = PropertyService.getProperty("application.datafilepath");
|
|
| 706 |
} else {
|
|
| 707 |
filepath = PropertyService.getProperty("application.documentfilepath");
|
|
| 708 |
} |
|
| 709 |
// ensure it is a directory path |
|
| 710 |
if (!(filepath.endsWith("/"))) {
|
|
| 711 |
filepath += "/"; |
|
| 712 |
} |
|
| 713 |
filename = filepath + docid; |
|
| 714 |
File documentFile = new File(filename); |
|
| 715 |
|
|
| 716 |
return documentFile; |
|
| 717 |
} |
|
| 720 | 718 |
} |
Also available in: Unified diff
calculate object size using the size on the file system rather than re-reading as an input stream.
Now only EML document bytes will be read twice: once for the checksum and again for parsing out datapackage details