comparison sworduploader.py @ 8:ff51b8204ad4

If using zip files, the zip is stored in DSpace as well
Small changes (in_progress explicitly)

 # Parse arguments
 parser = argparse.ArgumentParser(description="Bulk upload to DSpace using SWORDv2.",epilog="If the submission is created successfully, it will remain open to be completed with the necessary metadata and licenses, using the DSpace web interface. The submission can be found in the \"My Account -> Submissions\" section of the user's area.")
 parser.add_argument("data", type=str, nargs=1,
                    help="Accepts: METSDSpaceSIP packages, zip files, directories, single files. NOTE: METSDSpaceSIP packages are only accepted by Collections with a workflow!")
 parser.add_argument("--username", dest="user_name", type=str,nargs=1, help="DSpace username.")
-parser.add_argument("--zip", action="store_true",dest="zip",default=False, help="If \"data\" is a directory, send it as a single zip archive to preserve its structure.")
+parser.add_argument("--zip", action="store_true",dest="zip",default=False, help="If \"data\" is a directory, send it as a single zip archive to preserve its structure. The zip file will be saved along with the individual files.")
 parser.add_argument("--title", dest="title", type=str,nargs=1, help="Title (ignored for METS packages).")
 parser.add_argument("--author", dest="author", type=str,nargs="+", help="Author(s) (ignored for METS packages). Accepts multiple entries in the format \"Surname, Name\"")
 parser.add_argument("--date", dest="date", type=str,nargs=1, help="Date of creation (string) (ignored for METS packages).")
 parser.add_argument("--servicedoc", dest="dspaceurl", type=str,nargs=1, help="Url of the SWORDv2 service document (default: http://c4dm.eecs.qmul.ac.uk/smdmrd-test/swordv2/servicedocument")
 
 args = parser.parse_args()
 data = args.data[0]
+storezip = True
+
 if args.dspaceurl == None:
 	dspaceurl = "http://c4dm.eecs.qmul.ac.uk/smdmrd-test/swordv2/servicedocument"
 else:
 	dspaceurl = args.dspaceurl[0]
 

 				myzip = zipfile.ZipFile(zipf, "w")
 				# get the directory structure
 				print "Creating a zip archive for submission..."
 				for root, dirs, files in os.walk(data):
 					for name in files:
-						myzip.write(os.path.join(root,name),
-							os.path.relpath(os.path.join(root,name),data))
+						if not name.startswith('.'): # Do not upload hidden files, OSX/linux
+							myzip.write(os.path.join(root,name),
+								os.path.relpath(os.path.join(root,name),data))
 				fileslist.append(zipf)
 				myzip.close()
 				packaging = "http://purl.org/net/sword/package/SimpleZip"
 				type = "SimpleZip"
 				temp = True

 			fileslist.append(data)
 			myzip = zipfile.ZipFile(data)
 			if "mets.xml" in myzip.namelist(): #This is a METS package
 				packaging = "http://purl.org/net/sword/package/METSDSpaceSIP"
 				type = "METS"
+				in_progress = False
 			else: #THis is a simple zip file
 				packaging = "http://purl.org/net/sword/package/SimpleZip"
 				type = "SimpleZip"
 			myzip.close()
 		elif os.path.isfile(data): # This is a single file

 			if args.date != None:
 				entry.add_fields(dcterms_created = args.date[0])
 		else:
 			entry = None
 		# Select what to do
-		if (type is "single file") or (type is "multiple files"):
+		if (type is "single file") or (type is "multiple files"): # Use the single file upload procedure
 			try:
 				# Create the metadata entry with ATOM
 				print "------------------------"
 				print "Creating the item..."
 				if entry is None:
 					entry = Entry(dcterms_title=(os.path.basename(data)))
-				creation_receipt = c.create(col_iri = collection.href, metadata_entry = entry)
+				creation_receipt = c.create(col_iri = collection.href, metadata_entry = entry, in_progress=True)
 				
 				# Add the files
 				for f in fileslist:
 					print "Uploading file ",os.path.basename(f)
 					payload = open(f,"rb")

 						packaging = 'http://purl.org/net/sword/package/Binary')
 					payload.close()
 			except HTTPResponseError:
 				print "Bad request"
 		else:
-			# PUT the data
+			# Send the zip file and let the ingester do its job
 			payload = open(fileslist[0], "rb")
+			if type == "SimpleZip":
+				in_progress = True
+			else:
+				in_progress = False
 			try:
 				deposit_receipt = c.create(col_iri = collection.href,
 					payload = payload,
 					filename = fileslist[0],
 					mimetype = "application/zip",
-					packaging = packaging)
+					packaging = packaging,
+					in_progress = in_progress)
 				print type, " submission successful."
 			except:
 				print "Error! Couldn't submit the file!"						
 				if type == "METS": # Just guessing: not sure this is the problem...
 					print "To submit a METS package, the collection MUST have a workflow!"
 			payload.close()
 			
-			# If some of the additional arguments for author, title, date etc. have been specified, update the metadata
+			# If some of the additional arguments for author, title, date etc. have been specified, update the metadata (only SimpleZip)
 			if type == "SimpleZip":
 				if entry is None:
 					entry = Entry(dcterms_title=(os.path.basename(fileslist[0])))
 				try:
-					receipt_update = c.update(dr = receipt_dep , metadata_entry = entry, in_progress = True) # in_progress is True: we don't want to close the submission
+					update_receipt = c.update(dr = deposit_receipt , metadata_entry = entry, in_progress = True) # in_progress is True: we don't want to close the submission
 					print "Metadata update successfull."
 				except:
 					print "Server error"
+				# If we want to store the zip file along with the individual files (Only SimpleZip)
+				if storezip:
+					try:
+						payload = open(fileslist[0],"rb")
+						zipdeposit_receipt = c.add_file_to_resource(edit_media_iri = deposit_receipt.edit_media,
+							payload = payload,
+							filename = os.path.basename(fileslist[0]).replace(" ","_"),
+							mimetype = 'application/zip',
+							packaging = 'http://purl.org/net/sword/package/Binary')
+						payload.close()
+						print "Zip file successfully added to the bitstreams."
+					except:
+						print "Server error: could not add the zip file to the resources"
 			if temp:
 				os.remove(fileslist[0])
 
 		print "------------------------"
 		print "You will find the submission in the \"Submissions\" list in your DSpace account. To complete/edit it with metadata and licenses, click on the title and then on \"Resume\"."