/**
* Delete bitstream from item
*
* @param context
* @param ItemArchive
* @param isTest
* @param suppressUndo
* @throws IllegalArgumentException
* @throws ParseException
* @throws IOException
* @throws AuthorizeException
* @throws SQLException
*/
public void execute(Context context, ItemArchive itarch, boolean isTest, boolean suppressUndo)
throws IllegalArgumentException, IOException, SQLException, AuthorizeException,
ParseException {
File f = new File(itarch.getDirectory(), ItemUpdate.DELETE_CONTENTS_FILE);
if (!f.exists()) {
ItemUpdate.pr(
"Warning: Delete_contents file for item " + itarch.getDirectoryName() + " not found.");
} else {
List<Integer> list = MetadataUtilities.readDeleteContentsFile(f);
if (list.isEmpty()) {
ItemUpdate.pr("Warning: empty delete_contents file for item " + itarch.getDirectoryName());
} else {
for (int id : list) {
try {
Bitstream bs = Bitstream.find(context, id);
if (bs == null) {
ItemUpdate.pr("Bitstream not found by id: " + id);
} else {
Bundle[] bundles = bs.getBundles();
for (Bundle b : bundles) {
if (isTest) {
ItemUpdate.pr("Delete bitstream with id = " + id);
} else {
b.removeBitstream(bs);
ItemUpdate.pr("Deleted bitstream with id = " + id);
}
}
if (alterProvenance) {
DtoMetadata dtom = DtoMetadata.create("dc.description.provenance", "en", "");
String append =
"Bitstream " + bs.getName() + " deleted on " + DCDate.getCurrent() + "; ";
Item item = bundles[0].getItems()[0];
ItemUpdate.pr("Append provenance with: " + append);
if (!isTest) {
MetadataUtilities.appendMetadata(item, dtom, false, append);
}
}
}
} catch (SQLException e) {
ItemUpdate.pr("Error finding bitstream from id: " + id + " : " + e.toString());
}
}
}
}
}
/**
* Process Item, correcting CC-License if encountered.
*
* @param item
* @throws SQLException
* @throws AuthorizeException
* @throws IOException
*/
protected static void handleItem(Item item) throws SQLException, AuthorizeException, IOException {
Bundle[] bundles = item.getBundles("CC-LICENSE");
if (bundles == null || bundles.length == 0) return;
Bundle bundle = bundles[0];
Bitstream bitstream = bundle.getBitstreamByName("license_rdf");
String license_rdf = new String(copy(bitstream));
/* quickly fix xml by ripping out offensive parts */
license_rdf = license_rdf.replaceFirst("<license", "");
license_rdf = license_rdf.replaceFirst("</license>", "");
StringWriter result = new StringWriter();
try {
templates
.newTransformer()
.transform(
new StreamSource(new ByteArrayInputStream(license_rdf.getBytes())),
new StreamResult(result));
} catch (TransformerException e) {
throw new RuntimeException(e.getMessage(), e);
}
StringBuffer buffer = result.getBuffer();
Bitstream newBitstream =
bundle.createBitstream(new ByteArrayInputStream(buffer.toString().getBytes()));
newBitstream.setName(bitstream.getName());
newBitstream.setDescription(bitstream.getDescription());
newBitstream.setFormat(bitstream.getFormat());
newBitstream.setSource(bitstream.getSource());
newBitstream.setUserFormatDescription(bitstream.getUserFormatDescription());
newBitstream.update();
bundle.removeBitstream(bitstream);
bundle.update();
}
/**
* Create a new DSpace item out of a METS content package. All contents are dictated by the METS
* manifest. Package is a ZIP archive, all files relative to top level and the manifest (as per
* spec) in mets.xml.
*
* @param context - DSpace context.
* @param collection - collection under which to create new item.
* @param pkg - input stream containing package to ingest.
* @param license - may be null, which takes default license.
* @return workspace item created by ingest.
* @throws PackageValidationException if package is unacceptable or there is a fatal error turning
* it into an Item.
*/
public WorkspaceItem ingest(
Context context,
Collection collection,
InputStream pkg,
PackageParameters params,
String license)
throws PackageValidationException, CrosswalkException, AuthorizeException, SQLException,
IOException {
BitstreamDAO bsDAO = BitstreamDAOFactory.getInstance(context);
BitstreamFormatDAO bfDAO = BitstreamFormatDAOFactory.getInstance(context);
BundleDAO bundleDAO = BundleDAOFactory.getInstance(context);
WorkspaceItemDAO wsiDAO = WorkspaceItemDAOFactory.getInstance(context);
ZipInputStream zip = new ZipInputStream(pkg);
HashMap fileIdToBitstream = new HashMap();
WorkspaceItem wi = null;
boolean success = false;
HashSet packageFiles = new HashSet();
boolean validate = params.getBooleanProperty("validate", true);
try {
/* 1. Read all the files in the Zip into bitstreams first,
* because we only get to take one pass through a Zip input
* stream. Give them temporary bitstream names corresponding
* to the same names they had in the Zip, since those MUST
* match the URL references in <Flocat> and <mdRef> elements.
*/
METSManifest manifest = null;
wi = wsiDAO.create(collection, false);
Item item = wi.getItem();
Bundle contentBundle = item.createBundle(Constants.CONTENT_BUNDLE_NAME);
Bundle mdBundle = null;
ZipEntry ze;
while ((ze = zip.getNextEntry()) != null) {
if (ze.isDirectory()) continue;
Bitstream bs = null;
String fname = ze.getName();
if (fname.equals(MANIFEST_FILE)) {
if (preserveManifest) {
mdBundle = item.createBundle(Constants.METADATA_BUNDLE_NAME);
bs = mdBundle.createBitstream(new PackageUtils.UnclosableInputStream(zip));
bs.setName(fname);
bs.setSource(fname);
// Get magic bitstream format to identify manifest.
BitstreamFormat manifestFormat = null;
manifestFormat =
PackageUtils.findOrCreateBitstreamFormat(
context,
MANIFEST_BITSTREAM_FORMAT,
"application/xml",
MANIFEST_BITSTREAM_FORMAT + " package manifest");
bs.setFormat(manifestFormat);
manifest = METSManifest.create(bs.retrieve(), validate);
} else {
manifest = METSManifest.create(new PackageUtils.UnclosableInputStream(zip), validate);
continue;
}
} else {
bs = contentBundle.createBitstream(new PackageUtils.UnclosableInputStream(zip));
bs.setSource(fname);
bs.setName(fname);
}
packageFiles.add(fname);
bs.setSource(fname);
bsDAO.update(bs);
}
zip.close();
if (manifest == null)
throw new PackageValidationException(
"No METS Manifest found (filename=" + MANIFEST_FILE + "). Package is unacceptable.");
// initial sanity checks on manifest (in subclass)
checkManifest(manifest);
/* 2. Grovel a file list out of METS Manifest and compare
* it to the files in package, as an integrity test.
*/
List manifestContentFiles = manifest.getContentFiles();
// Compare manifest files with the ones found in package:
// a. Start with content files (mentioned in <fileGrp>s)
HashSet missingFiles = new HashSet();
for (Iterator mi = manifestContentFiles.iterator(); mi.hasNext(); ) {
// First locate corresponding Bitstream and make
// map of Bitstream to <file> ID.
Element mfile = (Element) mi.next();
String mfileId = mfile.getAttributeValue("ID");
if (mfileId == null)
throw new PackageValidationException(
"Invalid METS Manifest: file element without ID attribute.");
String path = METSManifest.getFileName(mfile);
Bitstream bs = contentBundle.getBitstreamByName(path);
if (bs == null) {
log.warn(
"Cannot find bitstream for filename=\""
+ path
+ "\", skipping it..may cause problems later.");
missingFiles.add(path);
} else {
fileIdToBitstream.put(mfileId, bs);
// Now that we're done using Name to match to <file>,
// set default bitstream Name to last path element;
// Zip entries all have '/' pathname separators
// NOTE: set default here, hopefully crosswalk of
// a bitstream techMD section will override it.
String fname = bs.getName();
int lastSlash = fname.lastIndexOf('/');
if (lastSlash >= 0 && lastSlash + 1 < fname.length())
bs.setName(fname.substring(lastSlash + 1));
// Set Default bitstream format:
// 1. attempt to guess from MIME type
// 2. if that fails, guess from "name" extension.
String mimeType = mfile.getAttributeValue("MIMETYPE");
BitstreamFormat bf = (mimeType == null) ? null : bfDAO.retrieveByMimeType(mimeType);
if (bf == null) bf = FormatIdentifier.guessFormat(context, bs);
bs.setFormat(bf);
// if this bitstream belongs in another Bundle, move it:
String bundleName = manifest.getBundleName(mfile);
if (!bundleName.equals(Constants.CONTENT_BUNDLE_NAME)) {
Bundle bn;
Bundle bns[] = item.getBundles(bundleName);
if (bns != null && bns.length > 0) bn = bns[0];
else bn = item.createBundle(bundleName);
bn.addBitstream(bs);
contentBundle.removeBitstream(bs);
}
// finally, build compare lists by deleting matches.
if (packageFiles.contains(path)) packageFiles.remove(path);
else missingFiles.add(path);
}
}
// b. Process files mentioned in <mdRef>s - check and move
// to METADATA bundle.
for (Iterator mi = manifest.getMdFiles().iterator(); mi.hasNext(); ) {
Element mdref = (Element) mi.next();
String path = METSManifest.getFileName(mdref);
// finally, build compare lists by deleting matches.
if (packageFiles.contains(path)) packageFiles.remove(path);
else missingFiles.add(path);
// if there is a bitstream with that name in Content, move
// it to the Metadata bundle:
Bitstream mdbs = contentBundle.getBitstreamByName(path);
if (mdbs != null) {
if (mdBundle == null) mdBundle = item.createBundle(Constants.METADATA_BUNDLE_NAME);
mdBundle.addBitstream(mdbs);
contentBundle.removeBitstream(mdbs);
}
}
// KLUDGE: make sure Manifest file doesn't get flagged as missing
// or extra, since it won't be mentioned in the manifest.
if (packageFiles.contains(MANIFEST_FILE)) packageFiles.remove(MANIFEST_FILE);
// Give subclass a chance to refine the lists of in-package
// and missing files, delete extraneous files, etc.
checkPackageFiles(packageFiles, missingFiles, manifest);
// Any discrepency in file lists is a fatal error:
if (!(packageFiles.isEmpty() && missingFiles.isEmpty())) {
StringBuffer msg =
new StringBuffer("Package is unacceptable: contents do not match manifest.");
if (!missingFiles.isEmpty()) {
msg.append("\nPackage is missing these files listed in Manifest:");
for (Iterator mi = missingFiles.iterator(); mi.hasNext(); )
msg.append("\n\t" + (String) mi.next());
}
if (!packageFiles.isEmpty()) {
msg.append("\nPackage contains extra files NOT in manifest:");
for (Iterator mi = packageFiles.iterator(); mi.hasNext(); )
msg.append("\n\t" + (String) mi.next());
}
throw new PackageValidationException(msg.toString());
}
/* 3. crosswalk the metadata
*/
// get mdref'd streams from "callback" object.
MdrefManager callback = new MdrefManager(mdBundle);
chooseItemDmd(context, item, manifest, callback, manifest.getItemDmds());
// crosswalk content bitstreams too.
for (Iterator ei = fileIdToBitstream.entrySet().iterator(); ei.hasNext(); ) {
Map.Entry ee = (Map.Entry) ei.next();
manifest.crosswalkBitstream(
context, (Bitstream) ee.getValue(), (String) ee.getKey(), callback);
}
// Take a second pass over files to correct names of derived files
// (e.g. thumbnails, extracted text) to what DSpace expects:
for (Iterator mi = manifestContentFiles.iterator(); mi.hasNext(); ) {
Element mfile = (Element) mi.next();
String bundleName = manifest.getBundleName(mfile);
if (!bundleName.equals(Constants.CONTENT_BUNDLE_NAME)) {
Element origFile = manifest.getOriginalFile(mfile);
if (origFile != null) {
String ofileId = origFile.getAttributeValue("ID");
Bitstream obs = (Bitstream) fileIdToBitstream.get(ofileId);
String newName = makeDerivedFilename(bundleName, obs.getName());
if (newName != null) {
String mfileId = mfile.getAttributeValue("ID");
Bitstream bs = (Bitstream) fileIdToBitstream.get(mfileId);
bs.setName(newName);
bsDAO.update(bs);
}
}
}
}
// Sanity-check the resulting metadata on the Item:
PackageUtils.checkMetadata(item);
/* 4. Set primary bitstream; same Bundle
*/
Element pbsFile = manifest.getPrimaryBitstream();
if (pbsFile != null) {
Bitstream pbs = (Bitstream) fileIdToBitstream.get(pbsFile.getAttributeValue("ID"));
if (pbs == null)
log.error(
"Got Primary Bitstream file ID="
+ pbsFile.getAttributeValue("ID")
+ ", but found no corresponding bitstream.");
else {
List<Bundle> bn = bundleDAO.getBundles(pbs);
if (bn.size() > 0) bn.get(0).setPrimaryBitstreamID(pbs.getID());
else log.error("Sanity check, got primary bitstream without any parent bundle.");
}
}
// have subclass manage license since it may be extra package file.
addLicense(context, collection, item, manifest, callback, license);
// subclass hook for final checks and rearrangements
finishItem(context, item);
// commit any changes to bundles
Bundle allBn[] = item.getBundles();
for (int i = 0; i < allBn.length; ++i) {
bundleDAO.update(allBn[i]);
}
wsiDAO.update(wi);
success = true;
log.info(
LogManager.getHeader(
context,
"ingest",
"Created new Item, db ID="
+ String.valueOf(item.getID())
+ ", WorkspaceItem ID="
+ String.valueOf(wi.getID())));
return wi;
} catch (SQLException se) {
// disable attempt to delete the workspace object, since
// database may have suffered a fatal error and the
// transaction rollback will get rid of it anyway.
wi = null;
// Pass this exception on to the next handler.
throw se;
} finally {
// kill item (which also deletes bundles, bitstreams) if ingest fails
if (!success && wi != null) wsiDAO.deleteAll(wi.getID());
}
}
/**
* Create new Item out of the ingested package, in the indicated collection. It creates a
* workspace item, which the application can then install if it chooses to bypass Workflow.
*
* <p>This is a VERY crude import of a single Adobe PDF (Portable Document Format) file, using the
* document's embedded metadata for package metadata. If the PDF file hasn't got the minimal
* metadata available, it is rejected.
*
* <p>
*
* @param context DSpace context.
* @param collection collection under which to create new item.
* @param pkg input stream containing package to ingest.
* @param params package parameters (none recognized)
* @param license may be null, which takes default license.
* @return workspace item created by ingest.
* @throws PackageException if package is unacceptable or there is a fatal error turning it into
* an Item.
*/
public WorkspaceItem ingest(
Context context,
Collection collection,
InputStream pkg,
PackageParameters params,
String license)
throws PackageValidationException, CrosswalkException, AuthorizeException, SQLException,
IOException {
InputStream bis = null;
COSDocument cos = null;
boolean success = false;
Bundle original = null;
Bitstream bs = null;
WorkspaceItem wi = null;
/**
* XXX comment out for now // XXX for debugging of parameter handling if (params != null) {
* Enumeration pe = params.propertyNames(); while (pe.hasMoreElements()) { String name =
* (String)pe.nextElement(); String v[] = params.getProperties(name); StringBuffer msg = new
* StringBuffer("PackageParam: "); msg.append(name).append(" = "); for (int i = 0; i < v.length;
* ++i) { if (i > 0) msg.append(", "); msg.append(v[i]); } log.debug(msg); } }
*/
try {
// Save the PDF in a bitstream first, since the parser
// has to read it as well, and we cannot "rewind" it after that.
wi = WorkspaceItem.create(context, collection, false);
Item myitem = wi.getItem();
original = myitem.createBundle("ORIGINAL");
bs = original.createBitstream(pkg);
pkg.close();
bs.setName("package.pdf");
setFormatToMIMEType(context, bs, "application/pdf");
bs.update();
log.debug("Created bitstream ID=" + String.valueOf(bs.getID()) + ", parsing...");
crosswalkPDF(context, myitem, bs.retrieve());
wi.update();
context.commit();
success = true;
log.info(
LogManager.getHeader(
context,
"ingest",
"Created new Item, db ID="
+ String.valueOf(myitem.getID())
+ ", WorkspaceItem ID="
+ String.valueOf(wi.getID())));
return wi;
} finally {
try {
// Close bitstream input stream and PDF file.
if (bis != null) bis.close();
if (cos != null) cos.close();
} catch (IOException ie) {
}
// get rid of bitstream and item if ingest fails
if (!success) {
if (original != null && bs != null) original.removeBitstream(bs);
if (wi != null) wi.deleteAll();
}
context.commit();
}
}