/**
* Attempt to filter a bitstream
*
* <p>An exception will be thrown if the media filter class cannot be instantiated, exceptions
* from filtering will be logged to STDOUT and swallowed.
*
* @return true if bitstream processed, false if no applicable filter or already processed
*/
public static boolean filterBitstream(Context c, Item myItem, Bitstream myBitstream)
throws Exception {
boolean filtered = false;
// iterate through filter classes. A single format may be actioned
// by more than one filter
for (int i = 0; i < filterClasses.length; i++) {
List fmts = (List) filterFormats.get(filterClasses[i].getClass().getName());
if (fmts.contains(myBitstream.getFormat().getShortDescription())) {
try {
// only update item if bitstream not skipped
if (filterClasses[i].processBitstream(c, myItem, myBitstream)) {
myItem.update(); // Make sure new bitstream has a sequence
// number
filtered = true;
}
} catch (Exception e) {
System.out.println(
"ERROR filtering, skipping bitstream #" + myBitstream.getID() + " " + e);
e.printStackTrace();
}
}
}
return filtered;
}
/**
* Remove a file from an item
*
* @param context current DSpace context
* @param item Item where file should be removed from
* @param bitstreamID The id of bitstream representing the file to remove
* @return Status or error flag which will be processed by UI-related code! (if STATUS_COMPLETE or
* 0 is returned, no errors occurred!)
*/
protected int processRemoveFile(Context context, Item item, int bitstreamID)
throws IOException, SQLException, AuthorizeException {
Bitstream bitstream;
// Try to find bitstream
try {
bitstream = Bitstream.find(context, bitstreamID);
} catch (NumberFormatException nfe) {
bitstream = null;
}
if (bitstream == null) {
// Invalid or mangled bitstream ID
// throw an error and return immediately
return STATUS_INTEGRITY_ERROR;
}
// remove bitstream from bundle..
// delete bundle if it's now empty
Bundle[] bundles = bitstream.getBundles();
bundles[0].removeBitstream(bitstream);
Bitstream[] bitstreams = bundles[0].getBitstreams();
// remove bundle if it's now empty
if (bitstreams.length < 1) {
item.removeBundle(bundles[0]);
item.update();
}
// no errors occurred
return STATUS_COMPLETE;
}
// Create workflow start provenance message
private static void recordStart(Context c, Item myitem)
throws SQLException, IOException, AuthorizeException {
// Get non-internal format bitstreams
Bitstream[] bitstreams = myitem.getNonInternalBitstreams();
// get date
DCDate now = DCDate.getCurrent();
// Create provenance description
String provmessage = "";
if (myitem.getSubmitter() != null) {
provmessage =
"Submitted by "
+ myitem.getSubmitter().getFullName()
+ " ("
+ myitem.getSubmitter().getEmail()
+ ") on "
+ now.toString()
+ "\n";
} else
// null submitter
{
provmessage = "Submitted by unknown (probably automated) on" + now.toString() + "\n";
}
// add sizes and checksums of bitstreams
provmessage += InstallItem.getBitstreamProvenanceMessage(myitem);
// Add message to the DC
myitem.addDC("description", "provenance", "en", provmessage);
myitem.update();
}
/** Update the workflow item, including the unarchived item. */
public void update() throws SQLException, IOException, AuthorizeException {
// FIXME check auth
log.info(
LogManager.getHeader(ourContext, "update_workflow_item", "workflow_item_id=" + getID()));
// Update the item
item.update();
// Update ourselves
DatabaseManager.update(ourContext, wfRow);
}
/**
* Look up the id of the template item for a given collection.
*
* @param context The current DSpace context.
* @param collectionID The collection id.
* @return The id of the template item.
* @throws IOException
*/
public static int getTemplateItemID(Context context, int collectionID)
throws SQLException, AuthorizeException, IOException {
Collection collection = Collection.find(context, collectionID);
Item template = collection.getTemplateItem();
if (template == null) {
collection.createTemplateItem();
template = collection.getTemplateItem();
collection.update();
template.update();
context.commit();
}
return template.getID();
}
/**
* rejects an item - rejection means undoing a submit - WorkspaceItem is created, and the
* WorkflowItem is removed, user is emailed rejection_message.
*
* @param c Context
* @param wi WorkflowItem to operate on
* @param e EPerson doing the operation
* @param rejection_message message to email to user
*/
public static WorkspaceItem reject(
Context c, WorkflowItem wi, EPerson e, String rejection_message)
throws SQLException, AuthorizeException, IOException {
// authorize a DSpaceActions.REJECT
// stop workflow
deleteTasks(c, wi);
// rejection provenance
Item myitem = wi.getItem();
// Get current date
String now = DCDate.getCurrent().toString();
// Get user's name + email address
String usersName = getEPersonName(e);
// Here's what happened
String provDescription =
"Rejected by " + usersName + ", reason: " + rejection_message + " on " + now + " (GMT) ";
// Add to item as a DC field
myitem.addDC("description", "provenance", "en", provDescription);
myitem.update();
// convert into personal workspace
WorkspaceItem wsi = returnToWorkspace(c, wi);
// notify that it's been rejected
notifyOfReject(c, wi, e, rejection_message);
log.info(
LogManager.getHeader(
c,
"reject_workflow",
"workflow_item_id="
+ wi.getID()
+ "item_id="
+ wi.getItem().getID()
+ "collection_id="
+ wi.getCollection().getID()
+ "eperson_id="
+ e.getID()));
return wsi;
}
/*
If we created a new Bitstream but now realised there is a problem then remove it.
*/
private void backoutBitstream(SubmissionInfo subInfo, Bitstream b, Item item)
throws SQLException, AuthorizeException, IOException {
// remove bitstream from bundle..
// delete bundle if it's now empty
Bundle[] bnd = b.getBundles();
bnd[0].removeBitstream(b);
Bitstream[] bitstreams = bnd[0].getBitstreams();
// remove bundle if it's now empty
if (bitstreams.length < 1) {
item.removeBundle(bnd[0]);
item.update();
}
subInfo.setBitstream(null);
}
// Record approval provenance statement
private static void recordApproval(Context c, WorkflowItem wi, EPerson e)
throws SQLException, IOException, AuthorizeException {
Item item = wi.getItem();
// Get user's name + email address
String usersName = getEPersonName(e);
// Get current date
String now = DCDate.getCurrent().toString();
// Here's what happened
String provDescription =
"Approved for entry into archive by " + usersName + " on " + now + " (GMT) ";
// add bitstream descriptions (name, size, checksums)
provDescription += InstallItem.getBitstreamProvenanceMessage(item);
// Add to item as a DC field
item.addDC("description", "provenance", "en", provDescription);
item.update();
}
/**
* Perform a deposit, using the supplied SWORD Deposit object.
*
* @param deposit
* @throws SWORDErrorException
* @throws DSpaceSWORDException
*/
public DepositResult doDeposit(Deposit deposit) throws SWORDErrorException, DSpaceSWORDException {
// get the things out of the service that we need
Context context = swordService.getContext();
SWORDConfiguration swordConfig = swordService.getSwordConfig();
SWORDUrlManager urlManager = swordService.getUrlManager();
// FIXME: the spec is unclear what to do in this situation. I'm going
// the throw a 415 (ERROR_CONTENT) until further notice
//
// determine if this is an acceptable file format
if (!swordConfig.isAcceptableContentType(context, deposit.getContentType(), collection)) {
log.error(
"Unacceptable content type detected: "
+ deposit.getContentType()
+ " for collection "
+ collection.getID());
throw new SWORDErrorException(
ErrorCodes.ERROR_CONTENT,
"Unacceptable content type in deposit request: " + deposit.getContentType());
}
// determine if this is an acceptable packaging type for the deposit
// if not, we throw a 415 HTTP error (Unsupported Media Type, ERROR_CONTENT)
if (!swordConfig.isSupportedMediaType(deposit.getPackaging(), this.collection)) {
log.error(
"Unacceptable packaging type detected: "
+ deposit.getPackaging()
+ "for collection"
+ collection.getID());
throw new SWORDErrorException(
ErrorCodes.ERROR_CONTENT,
"Unacceptable packaging type in deposit request: " + deposit.getPackaging());
}
// Obtain the relevant ingester from the factory
SWORDIngester si = SWORDIngesterFactory.getInstance(context, deposit, collection);
swordService.message("Loaded ingester: " + si.getClass().getName());
// do the deposit
DepositResult result = si.ingest(swordService, deposit, collection);
swordService.message("Archive ingest completed successfully");
// if there's an item availalble, and we want to keep the original
// then do that
try {
if (swordConfig.isKeepOriginal()) {
swordService.message(
"DSpace will store an original copy of the deposit, "
+ "as well as ingesting the item into the archive");
// in order to be allowed to add the file back to the item, we need to ignore authorisations
// for a moment
boolean ignoreAuth = context.ignoreAuthorization();
context.setIgnoreAuthorization(true);
String bundleName = ConfigurationManager.getProperty("sword-server", "bundle.name");
if (bundleName == null || "".equals(bundleName)) {
bundleName = "SWORD";
}
Item item = result.getItem();
Bundle[] bundles = item.getBundles(bundleName);
Bundle swordBundle = null;
if (bundles.length > 0) {
swordBundle = bundles[0];
}
if (swordBundle == null) {
swordBundle = item.createBundle(bundleName);
}
String fn = swordService.getFilename(context, deposit, true);
Bitstream bitstream;
FileInputStream fis = null;
try {
fis = new FileInputStream(deposit.getFile());
bitstream = swordBundle.createBitstream(fis);
} finally {
if (fis != null) {
fis.close();
}
}
bitstream.setName(fn);
bitstream.setDescription("SWORD deposit package");
BitstreamFormat bf = BitstreamFormat.findByMIMEType(context, deposit.getContentType());
if (bf != null) {
bitstream.setFormat(bf);
}
bitstream.update();
swordBundle.update();
item.update();
swordService.message(
"Original package stored as " + fn + ", in item bundle " + swordBundle);
// now reset the context ignore authorisation
context.setIgnoreAuthorization(ignoreAuth);
// set the media link for the created item
result.setMediaLink(urlManager.getMediaLink(bitstream));
} else {
// set the vanilla media link, which doesn't resolve to anything
result.setMediaLink(urlManager.getBaseMediaLinkUrl());
}
} catch (SQLException e) {
log.error("caught exception: ", e);
throw new DSpaceSWORDException(e);
} catch (AuthorizeException e) {
log.error("caught exception: ", e);
throw new DSpaceSWORDException(e);
} catch (FileNotFoundException e) {
log.error("caught exception: ", e);
throw new DSpaceSWORDException(e);
} catch (IOException e) {
log.error("caught exception: ", e);
throw new DSpaceSWORDException(e);
}
return result;
}
/**
* perform the ingest using the given deposit object onto the specified target dspace object,
* using the sword service implementation
*
* @param service
* @param deposit
* @param target
* @return
* @throws DSpaceSWORDException
* @throws SWORDErrorException
*/
public DepositResult ingest(SWORDService service, Deposit deposit, DSpaceObject target)
throws DSpaceSWORDException, SWORDErrorException {
try {
if (!(target instanceof Item)) {
throw new DSpaceSWORDException(
"SimpleFileIngester can only be loaded for deposit onto DSpace Items");
}
Item item = (Item) target;
// now set the sword service
SWORDService swordService = service;
// get the things out of the service that we need
Context context = swordService.getContext();
SWORDUrlManager urlManager = swordService.getUrlManager();
Bundle[] bundles = item.getBundles("ORIGINAL");
Bundle original;
if (bundles.length > 0) {
original = bundles[0];
} else {
original = item.createBundle("ORIGINAL");
}
Bitstream bs;
FileInputStream fis = null;
try {
fis = new FileInputStream(deposit.getFile());
bs = original.createBitstream(fis);
} finally {
if (fis != null) {
fis.close();
}
}
String fn = swordService.getFilename(context, deposit, false);
bs.setName(fn);
swordService.message("File created in item with filename " + fn);
BitstreamFormat bf = BitstreamFormat.findByMIMEType(context, deposit.getContentType());
if (bf != null) {
bs.setFormat(bf);
}
// to do the updates, we need to ignore authorisation in the context
boolean ignoreAuth = context.ignoreAuthorization();
context.setIgnoreAuthorization(true);
bs.update();
original.update();
item.update();
// reset the ignore authorisation
context.setIgnoreAuthorization(ignoreAuth);
DepositResult result = new DepositResult();
result.setHandle(urlManager.getBitstreamUrl(bs));
result.setTreatment(this.getTreatment());
result.setBitstream(bs);
return result;
} catch (SQLException e) {
throw new DSpaceSWORDException(e);
} catch (AuthorizeException e) {
throw new DSpaceSWORDException(e);
} catch (IOException e) {
throw new DSpaceSWORDException(e);
}
}
private void processArchive(
Context context,
String sourceDirPath,
String itemField,
String metadataIndexName,
boolean alterProvenance,
boolean isTest)
throws Exception {
// open and process the source directory
File sourceDir = new File(sourceDirPath);
if ((sourceDir == null) || !sourceDir.exists() || !sourceDir.isDirectory()) {
pr("Error, cannot open archive source directory " + sourceDirPath);
throw new Exception("error with archive source directory " + sourceDirPath);
}
String[] dircontents = sourceDir.list(directoryFilter); // just the names, not the path
Arrays.sort(dircontents);
// Undo is suppressed to prevent undo of undo
boolean suppressUndo = false;
File fSuppressUndo = new File(sourceDir, SUPPRESS_UNDO_FILENAME);
if (fSuppressUndo.exists()) {
suppressUndo = true;
}
File undoDir = null; // sibling directory of source archive
if (!suppressUndo && !isTest) {
undoDir = initUndoArchive(sourceDir);
}
int itemCount = 0;
int successItemCount = 0;
for (String dirname : dircontents) {
itemCount++;
pr("");
pr("processing item " + dirname);
try {
ItemArchive itarch = ItemArchive.create(context, new File(sourceDir, dirname), itemField);
for (UpdateAction action : actionMgr) {
pr("action: " + action.getClass().getName());
action.execute(context, itarch, isTest, suppressUndo);
if (!isTest && !suppressUndo) {
itarch.writeUndo(undoDir);
}
}
if (!isTest) {
Item item = itarch.getItem();
item.update(); // need to update before commit
context.commit();
item.decache();
}
ItemUpdate.pr("Item " + dirname + " completed");
successItemCount++;
} catch (Exception e) {
pr("Exception processing item " + dirname + ": " + e.toString());
}
}
if (!suppressUndo && !isTest) {
StringBuilder sb = new StringBuilder("dsrun org.dspace.app.itemupdate.ItemUpdate ");
sb.append(" -e ").append(this.eperson);
sb.append(" -s ").append(undoDir);
if (itemField != null) {
sb.append(" -i ").append(itemField);
}
if (!alterProvenance) {
sb.append(" -P ");
}
if (isTest) {
sb.append(" -t ");
}
for (String actionOption : undoActionList) {
sb.append(actionOption);
}
PrintWriter pw = null;
try {
File cmdFile = new File(undoDir.getParent(), undoDir.getName() + "_command.sh");
pw = new PrintWriter(new BufferedWriter(new FileWriter(cmdFile)));
pw.println(sb.toString());
} finally {
pw.close();
}
}
pr("");
pr(
"Done processing. Successful items: "
+ successItemCount
+ " of "
+ itemCount
+ " items in source archive");
pr("");
}
private void crosswalkPDF(Context context, Item item, InputStream metadata)
throws CrosswalkException, IOException, SQLException, AuthorizeException {
COSDocument cos = null;
try {
PDFParser parser = new PDFParser(metadata);
parser.parse();
cos = parser.getDocument();
// sanity check: PDFBox breaks on encrypted documents, so give up.
if (cos.getEncryptionDictionary() != null)
throw new MetadataValidationException(
"This packager cannot accept an encrypted PDF document.");
/* PDF to DC "crosswalk":
*
* NOTE: This is not in a crosswalk plugin because (a) it isn't
* useful anywhere else, and more importantly, (b) the source
* data is not XML so it doesn't fit the plugin's interface.
*
* pattern of crosswalk -- PDF dict entries to DC:
* Title -> title.null
* Author -> contributor.author
* CreationDate -> date.created
* ModDate -> date.created
* Creator -> description.provenance (application that created orig)
* Producer -> description.provenance (convertor to pdf)
* Subject -> description.abstract
* Keywords -> subject.other
* date is java.util.Calendar
*/
PDDocument pd = new PDDocument(cos);
PDDocumentInformation docinfo = pd.getDocumentInformation();
String title = docinfo.getTitle();
// sanity check: item must have a title.
if (title == null)
throw new MetadataValidationException(
"This PDF file is unacceptable, it does not have a value for \"Title\" in its Info dictionary.");
log.debug("PDF Info dict title=\"" + title + "\"");
item.addDC("title", null, "en", title);
String value;
Calendar date;
if ((value = docinfo.getAuthor()) != null) {
item.addDC("contributor", "author", null, value);
log.debug("PDF Info dict author=\"" + value + "\"");
}
if ((value = docinfo.getCreator()) != null)
item.addDC(
"description",
"provenance",
"en",
"Application that created the original document: " + value);
if ((value = docinfo.getProducer()) != null)
item.addDC(
"description", "provenance", "en", "Original document converted to PDF by: " + value);
if ((value = docinfo.getSubject()) != null)
item.addDC("description", "abstract", null, value);
if ((value = docinfo.getKeywords()) != null) item.addDC("subject", "other", null, value);
// Take either CreationDate or ModDate as "date.created",
// Too bad there's no place to put "last modified" in the DC.
Calendar calValue;
if ((calValue = docinfo.getCreationDate()) == null) calValue = docinfo.getModificationDate();
if (calValue != null)
item.addDC("date", "created", null, (new DCDate(calValue.getTime())).toString());
item.update();
} finally {
if (cos != null) cos.close();
}
}
/**
* Process the upload of a new file!
*
* @param context current DSpace context
* @param request current servlet request object
* @param response current servlet response object
* @param subInfo submission info object
* @return Status or error flag which will be processed by UI-related code! (if STATUS_COMPLETE or
* 0 is returned, no errors occurred!)
*/
protected int processUploadFile(
Context context,
HttpServletRequest request,
HttpServletResponse response,
SubmissionInfo subInfo)
throws ServletException, IOException, SQLException, AuthorizeException {
boolean formatKnown = true;
boolean fileOK = false;
BitstreamFormat bf = null;
Bitstream b = null;
// NOTE: File should already be uploaded.
// Manakin does this automatically via Cocoon.
// For JSP-UI, the SubmissionController.uploadFiles() does the actual upload
Enumeration attNames = request.getAttributeNames();
// loop through our request attributes
while (attNames.hasMoreElements()) {
String attr = (String) attNames.nextElement();
// if this ends with "-path", this attribute
// represents a newly uploaded file
if (attr.endsWith("-path")) {
// strip off the -path to get the actual parameter
// that the file was uploaded as
String param = attr.replace("-path", "");
// Load the file's path and input stream and description
String filePath = (String) request.getAttribute(param + "-path");
InputStream fileInputStream = (InputStream) request.getAttribute(param + "-inputstream");
// attempt to get description from attribute first, then direct from a parameter
String fileDescription = (String) request.getAttribute(param + "-description");
if (fileDescription == null || fileDescription.length() == 0) {
fileDescription = request.getParameter("description");
}
// if information wasn't passed by User Interface, we had a problem
// with the upload
if (filePath == null || fileInputStream == null) {
return STATUS_UPLOAD_ERROR;
}
if (subInfo == null) {
// In any event, if we don't have the submission info, the request
// was malformed
return STATUS_INTEGRITY_ERROR;
}
// Create the bitstream
Item item = subInfo.getSubmissionItem().getItem();
// do we already have a bundle?
Bundle[] bundles = item.getBundles("ORIGINAL");
if (bundles.length < 1) {
// set bundle's name to ORIGINAL
b = item.createSingleBitstream(fileInputStream, "ORIGINAL");
} else {
// we have a bundle already, just add bitstream
b = bundles[0].createBitstream(fileInputStream);
}
// Strip all but the last filename. It would be nice
// to know which OS the file came from.
String noPath = filePath;
while (noPath.indexOf('/') > -1) {
noPath = noPath.substring(noPath.indexOf('/') + 1);
}
while (noPath.indexOf('\\') > -1) {
noPath = noPath.substring(noPath.indexOf('\\') + 1);
}
b.setName(noPath);
b.setSource(filePath);
b.setDescription(fileDescription);
// Identify the format
bf = FormatIdentifier.guessFormat(context, b);
b.setFormat(bf);
// Update to DB
b.update();
item.update();
if ((bf != null) && (bf.isInternal())) {
log.warn("Attempt to upload file format marked as internal system use only");
backoutBitstream(subInfo, b, item);
return STATUS_UPLOAD_ERROR;
}
// Check for virus
if (ConfigurationManager.getBooleanProperty("submission-curation", "virus-scan")) {
Curator curator = new Curator();
curator.addTask("vscan").curate(item);
int status = curator.getStatus("vscan");
if (status == Curator.CURATE_ERROR) {
backoutBitstream(subInfo, b, item);
return STATUS_VIRUS_CHECKER_UNAVAILABLE;
} else if (status == Curator.CURATE_FAIL) {
backoutBitstream(subInfo, b, item);
return STATUS_CONTAINS_VIRUS;
}
}
// If we got this far then everything is more or less ok.
// Comment - not sure if this is the right place for a commit here
// but I'm not brave enough to remove it - Robin.
context.commit();
// save this bitstream to the submission info, as the
// bitstream we're currently working with
subInfo.setBitstream(b);
// if format was not identified
if (bf == null) {
return STATUS_UNKNOWN_FORMAT;
}
} // end if attribute ends with "-path"
} // end while
return STATUS_COMPLETE;
}
