def create_datafile(index): testfile = path.join(path.dirname(__file__), 'fixtures', 'jeol_sem_test%d.txt' % index) size, sha512sum = get_size_and_sha512sum(testfile) datafile = Dataset_File(dataset=dataset, filename=path.basename(testfile), size=size, sha512sum=sha512sum) datafile.save() base_url = 'file://' + path.abspath(path.dirname(testfile)) location = Location.load_location({ 'name': 'test-jeol', 'url': base_url, 'type': 'external', 'priority': 10, 'transfer_provider': 'local' }) replica = Replica(datafile=datafile, url='file://' + path.abspath(testfile), protocol='file', location=location) replica.verify() replica.save() return Dataset_File.objects.get(pk=datafile.pk)
def process_enclosure(self, dataset, enclosure): filename = getattr(enclosure, 'title', basename(enclosure.href)) datafile = Dataset_File(filename=filename, dataset=dataset) try: datafile.mimetype = enclosure.mime except AttributeError: pass try: datafile.size = enclosure.length except AttributeError: pass try: hash = enclosure.hash # Split on white space, then ':' to get tuples to feed into dict hashdict = dict([s.partition(':')[::2] for s in hash.split()]) # Set SHA-512 sum datafile.sha512sum = hashdict['sha-512'] except AttributeError: pass datafile.save() url = enclosure.href # This means we will allow the atom feed to feed us any enclosure # URL that matches a registered location. Maybe we should restrict # this to a specific location. location = Location.get_location_for_url(url) if not location: logger.error('Rejected ingestion for unknown location %s' % url) return replica = Replica(datafile=datafile, url=url, location=location) replica.protocol = enclosure.href.partition('://')[0] replica.save() self.make_local_copy(replica)
def _build(dataset, filename, url, protocol): from tardis.tardis_portal.models import \ Dataset_File, Replica, Location datafile = Dataset_File(dataset=dataset, filename=filename) datafile.save() replica = Replica(datafile=datafile, url=url, protocol=protocol, location=Location.get_default_location()) replica.save() return datafile
def _create_test_dataset(nosDatafiles): ds_ = Dataset(description='happy snaps of plumage') ds_.save() for i in range(0, nosDatafiles): df_ = Dataset_File(dataset=ds_, size='21', sha512sum='bogus') df_.save() rep_ = Replica(datafile=df_, url='http://planet-python.org/' + str(_next_id()), location=Location.get_default_location()) rep_.save() ds_.save() return ds_
def generate_datafile(path, dataset, content=None, size=-1, verify=True, verified=True, verify_checksums_req=False): '''Generates a datafile AND a replica to hold its contents''' from tardis.tardis_portal.models import Dataset_File, Replica, Location saved = settings.REQUIRE_DATAFILE_CHECKSUMS settings.REQUIRE_DATAFILE_CHECKSUMS = False try: datafile = Dataset_File() if content: datafile.size = str(len(content)) else: datafile.size = str(size) # Normally we use any old string for the datafile path, but some # tests require the path to be the same as what 'staging' would use if path == None: datafile.dataset_id = dataset.id datafile.save() path = "%s/%s/%s" % (dataset.get_first_experiment().id, dataset.id, datafile.id) filepath = os.path.normpath(settings.FILE_STORE_PATH + '/' + path) if content: try: os.makedirs(os.path.dirname(filepath)) os.remove(filepath) except: pass gen_file = open(filepath, 'wb+') gen_file.write(content) gen_file.close() datafile.mimetype = "application/unspecified" datafile.filename = os.path.basename(filepath) datafile.dataset_id = dataset.id datafile.save() settings.REQUIRE_DATAFILE_CHECKSUMS = verify_checksums_req location = _infer_location(path) replica = Replica(datafile=datafile, url=path, protocol='', location=location) if verify and content: if not replica.verify(): raise RuntimeError('verify failed!?!') replica.save() replica.verified = verified replica.save(update_fields=['verified']) # force no verification return (datafile, replica) finally: settings.REQUIRE_DATAFILE_CHECKSUMS = saved
def testMirror(self): dest = Location.get_location('test') datafile, replica = generate_datafile(None, self.dataset, "Hi granny") path = datafile.get_absolute_filepath() self.assertTrue(os.path.exists(path)) dummy_replica = Replica() dummy_replica.datafile = datafile dummy_replica.location = Location.objects.get(name='test') dummy_replica.url = dummy_replica.generate_default_url() with self.assertRaises(TransferError): dest.provider.get_length(dummy_replica) self.assertTrue(migrate_replica(replica, dest, mirror=True)) datafile = Dataset_File.objects.get(id=datafile.id) self.assertTrue(datafile.is_local()) self.assertEquals(dest.provider.get_length(dummy_replica), 9)
def create_staging_datafile(filepath, username, dataset_id): from tardis.tardis_portal.models import Dataset_File, Dataset, Replica, \ Location dataset = Dataset.objects.get(id=dataset_id) url, size = get_staging_url_and_size(username, filepath) datafile = Dataset_File(dataset=dataset, filename=path.basename(filepath), size=size) replica = Replica(datafile=datafile, protocol='staging', url=url, location=Location.get_location('staging')) replica.verify(allowEmptyChecksums=True) datafile.save() replica.datafile = datafile replica.save()
def _create_datafile(): user = User.objects.create_user('testuser', '*****@*****.**', 'pwd') user.save() UserProfile(user=user).save() Location.force_initialize() full_access = Experiment.PUBLIC_ACCESS_FULL experiment = Experiment.objects.create(title="IIIF Test", created_by=user, public_access=full_access) experiment.save() ObjectACL(content_object=experiment, pluginId='django_user', entityId=str(user.id), isOwner=True, canRead=True, canWrite=True, canDelete=True, aclOwnershipType=ObjectACL.OWNER_OWNED).save() dataset = Dataset() dataset.save() dataset.experiments.add(experiment) dataset.save() # Create new Datafile tempfile = TemporaryUploadedFile('iiif_stored_file', None, None, None) with Image(filename='magick:rose') as img: img.format = 'tiff' img.save(file=tempfile.file) tempfile.file.flush() datafile = Dataset_File(dataset=dataset, size=os.path.getsize(tempfile.file.name), filename='iiif_named_file') replica = Replica(datafile=datafile, url=write_uploaded_file_to_dataset(dataset, tempfile), location=Location.get_default_location()) replica.verify(allowEmptyChecksums=True) datafile.save() replica.datafile = datafile replica.save() return datafile
def _build_datafile(self, testfile, filename, dataset, url, protocol='', checksum=None, size=None, mimetype=''): filesize, sha512sum = get_size_and_sha512sum(testfile) datafile = Dataset_File( dataset=dataset, filename=filename, mimetype=mimetype, size=str(size if size != None else filesize), sha512sum=(checksum if checksum else sha512sum)) datafile.save() if urlparse.urlparse(url).scheme == '': location = Location.get_location('local') else: location = Location.get_location_for_url(url) if not location: location = Location.load_location({ 'name': filename, 'url': urlparse.urljoin(url, '.'), 'type': 'external', 'priority': 10, 'transfer_provider': 'local' }) replica = Replica(datafile=datafile, protocol=protocol, url=url, location=location) replica.verify() replica.save() return Dataset_File.objects.get(pk=datafile.pk)
def fpupload(request, dataset_id): """ Uploads all files picked by filepicker to the dataset :param request: a HTTP Request instance :type request: :class:`django.http.HttpRequest` :param dataset_id: the dataset_id :type dataset_id: integer :returns: boolean true if successful :rtype: bool """ dataset = Dataset.objects.get(id=dataset_id) logger.debug('called fpupload') if request.method == 'POST': logger.debug('got POST') for key, val in request.POST.items(): splits = val.split(",") for url in splits: try: fp = FilepickerFile(url) except ValueError: pass else: picked_file = fp.get_file() filepath = write_uploaded_file_to_dataset( dataset, picked_file) datafile = Dataset_File(dataset=dataset, filename=picked_file.name, size=picked_file.size) replica = Replica(datafile=datafile, url=filepath, protocol='', location=Location.get_default_location()) replica.verify(allowEmptyChecksums=True) datafile.save() replica.datafile = datafile replica.save() return HttpResponse(json.dumps({"result": True}))
def setUp(self): """ setting up essential objects, copied from tests above """ Location.force_initialize() self.location = Location.get_location('local') user = '******' pwd = 'secret' email = '' self.user = User.objects.create_user(user, email, pwd) self.userProfile = UserProfile(user=self.user).save() self.exp = Experiment(title='test exp1', institution_name='monash', created_by=self.user) self.exp.save() self.acl = ObjectACL( pluginId=django_user, entityId=str(self.user.id), content_object=self.exp, canRead=True, isOwner=True, aclOwnershipType=ObjectACL.OWNER_OWNED, ) self.acl.save() self.dataset = Dataset(description='dataset description...') self.dataset.save() self.dataset.experiments.add(self.exp) self.dataset.save() self.dataset_file = Dataset_File(dataset=self.dataset, size=42, filename="foo", md5sum="junk") self.dataset_file.save() self.replica = Replica(datafile=self.dataset_file, url="http://foo", location=self.location, verified=False) self.replica.save()
def do_provider(self, loc): provider = loc.provider base_url = loc.url datafile, replica = generate_datafile("1/1/3", self.dataset, "Hi mum") self.assertEquals(replica.verify(allowEmptyChecksums=True), True) target_replica = Replica() target_replica.datafile = datafile target_replica.location = loc url = provider.generate_url(target_replica) self.assertEquals(url, base_url + '1/1/3') target_replica.url = url provider.put_file(replica, target_replica) self.assertEqual( replica.location.provider.get_file(replica).read(), "Hi mum") self.assertEqual(provider.get_file(target_replica).read(), "Hi mum") self.assertEqual(provider.get_length(target_replica), 6) try: self.maxDiff = None self.assertEqual( provider.get_metadata(target_replica), { 'sha512sum': '2274cc8c16503e3d182ffaa835c543b' + 'ce278bc8fc971f3bf38b94b4d9db44cd89c8f36d4006e' + '5abea29bc05f7f0ea662cb4b0e805e56bbce97f00f94e' + 'a6e6498', 'md5sum': '3b6b51114c3d0ad347e20b8e79765951', 'length': '6' }) except NotImplementedError: pass provider.remove_file(target_replica) with self.assertRaises(TransferError): provider.get_length(target_replica)
def migrate_replica(replica, location, noRemove=False, mirror=False): """ Migrate the replica to a different storage location. The overall effect will be that the datafile will be stored at the new location and removed from the current location, and the datafile metadata will be updated to reflect this. """ from tardis.tardis_portal.models import Replica, Location with transaction.commit_on_success(): replica = Replica.objects.select_for_update().get(pk=replica.pk) source = Location.get_location(replica.location.name) if not replica.verified or location.provider.trust_length: raise MigrationError('Only verified datafiles can be migrated' \ ' to this destination') filename = replica.get_absolute_filepath() try: newreplica = Replica.objects.get(datafile=replica.datafile, location=location) created_replica = False # We've most likely mirrored this file previously. But if # we are about to delete the source Replica, we need to check # that the target Replica still verifies. if not mirror and not check_file_transferred(newreplica, location): raise MigrationError('Previously mirrored / migrated Replica' \ ' no longer verifies locally!') except Replica.DoesNotExist: newreplica = Replica() newreplica.location = location newreplica.datafile = replica.datafile newreplica.protocol = '' newreplica.stay_remote = location != Location.get_default_location( ) newreplica.verified = False url = location.provider.generate_url(newreplica) if newreplica.url == url: # We should get here ... raise MigrationError('Cannot migrate a replica to its' \ ' current location') newreplica.url = url location.provider.put_file(replica, newreplica) verified = False try: verified = check_file_transferred(newreplica, location) except: # FIXME - should we always do this? location.provider.remove_file(newreplica) raise newreplica.verified = verified newreplica.save() logger.info('Transferred file %s for replica %s' % (filename, replica.id)) created_replica = True if mirror: return created_replica # FIXME - do this more reliably ... replica.delete() if not noRemove: source.provider.remove_file(replica) logger.info('Removed local file %s for replica %s' % (filename, replica.id)) return True