view sworduploader.py @ 3:c1918aa337c4

v0.2 small bug fixes

#!usr/bin/env/ python

""" SWORD2 DSpace bulk uploader - v0.2

A python script to submit large numbers of files to a SWORD2-compatible repository, specifically DSpace 1.8x.
Built on the SWORD2 python client library: https://bitbucket.org/beno/python-sword2/overview

Dependencies:

- python 2.X

- sword2 library: https://bitbucket.org/beno/python-sword2/src

-----------------------------------
Copyright 2012 Marco Fabiani
Copyright 2012 Queen Mary, University of London
-----------------------------------
"""

import argparse, getpass, zipfile,os,sys
from sword2 import *

# 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("--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]
if args.dspaceurl == None:
	dspaceurl = "http://c4dm.eecs.qmul.ac.uk/smdmrd-test/swordv2/servicedocument"
else:
	dspaceurl = args.dspaceurl[0]


try:
	# Connect to SWORD server
	attempts = 3 #  Number of attempts left to connect to server
	connected = False
	while attempts>0 and not connected:
		print "Connecting to SWORD server. Remaining attempts: ", attempts
		# Try to login, get service document
		# Get username and password
		if args.user_name == None:
			user_name = raw_input("DSpace Username: ")
		else:
			user_name = args.user_name[0]
			print "DSpace Username: ",user_name
		user_pass = getpass.getpass("DSpace password:")
		# Connect to the server
		c = Connection(dspaceurl, user_name=user_name, user_pass=user_pass,keep_history=False)
		
		# Get service document
		try:
			c.get_service_document()
		except: # Server error
			print "Server unreachable!"
			break
		if c.sd != None:
			connected = True
		else:
			attempts-=1
			print "Incorrect username and/or password"
			
			
	if connected:
		# List available collections
		print "------------------------"
		print "Welcome to the ",c.workspaces[0][0], "repository"
		print "Available Collections: "
		numColl = len(c.workspaces[0][1])
		for ctr in range(numColl):
			coll = c.workspaces[0][1][ctr]
			print ctr+1,":",coll.title
		# Select a collection to deposit into
		sel = -1
		while (sel<=0 or sel>numColl):
			sel = input("Select a Collection to submit your files into: ")
		collection = c.workspaces[0][1][sel-1]
		print "Selected Collection: ",collection.title
		
		# Create a submission: build the zip files
		temp = True # delete the zip file at the end of the upload	
		if zipfile.is_zipfile(data):
			zipf = data
			temp = False
		elif os.path.isfile(data): # This is a single file
			dataname = os.path.basename(data)
			zipf = os.path.splitext(dataname)[0]+".zip"
			myzip = zipfile.ZipFile(zipf, "w")
			myzip.write(data,os.path.basename(data))
			myzip.close()
		elif os.path.isdir(data): # This is a directory, zip all the files and maintain the structure, but start from the base only...
			dataname = os.path.basename(os.path.normpath(data))
			zipf = dataname+".zip"
			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))
			myzip.close()
		else:
			print "Couldn't find the data."
			sys.exit()
		
		#Check if this is a METSDSpaceSIP: see if there is a mets.xml file in the zip
		myzip = zipfile.ZipFile(zipf)
		if "mets.xml" in myzip.namelist():
			packaging = "http://purl.org/net/sword/package/METSDSpaceSIP"
			type = "METS"
		else:
			packaging = "http://purl.org/net/sword/package/SimpleZip"
			type = "SimpleZip"
		
		print "------------------------"
		print "This is a ",type," submission"
		print "Uploading files..."
		myzip.close()

		payload = open(zipf, "rb")
		try:
			receipt_dep = c.create(col_iri = collection.href,
			payload = payload,
			filename = zipf,
			mimetype = "application/zip",
			packaging = packaging)
			print type, " submission successful."
			if type == "SimpleZip":
				# If some of the additional arguments for author, title, date etc. have been specified, update the metadata
				if (args.title != None) or (args.author != None) or (args.date != None):
					entry = Entry()	
					print "------------------------"
					print "Updating with additional metadata"
					if  args.title != None:
						entry.add_fields(dcterms_title = args.title[0])
					if args.author != None:
						for creator in args.author:
							entry.add_fields(dcterms_creator=creator)
					if args.date != None:
						entry.add_fields(dcterms_created = args.date[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
						print "Additional metadata updated successfully."
					except:
						print "Server error"
			print "------------------------"
			print "You will find this 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\"."			
		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 temp:
			os.remove(zipf)
		
	else: # Failed to connect to SWORDv2 Server
		print "Couldn't connect to the server."
		if attempts == 0:
			print "Invalid credentials entered 3 times."
	
except KeyboardInterrupt:
	print "------------------------"
	print "\nSubmission aborted by user."