def _create_experiment(self, data=None):
from tardis.tardis_portal import models, forms
from os.path import basename
from django.contrib.auth.models import User
data = self._data_to_post(data)
exp = models.Experiment(
title=data['title'],
institution_name=data['institution_name'],
description=data['description'],
created_by=User.objects.get(id=data['created_by']),
)
exp.save()
for i, a in enumerate(data['authors'].split(', ')):
ae = models.Author_Experiment(experiment=exp, author=a, order=i)
ae.save()
ds_desc = {
'first one': ['file/another.py'],
'second': ['second_ds/file.py', 'second_ds/file1.py']
}
for d, df in ds_desc.items():
dataset = models.Dataset(description=d, experiment=exp)
dataset.save()
for f in df:
d = models.Dataset_File(url='file://' + f,
dataset=dataset,
filename=basename(f))
d.save()
return exp
def test_authors(self):
from tardis.tardis_portal import models
exp = models.Experiment(title='test exp2',
institution_name='monash',
created_by=self.user,
)
exp.save()
models.Author_Experiment(experiment=exp,
author='nigel',
order=0).save()
exp = models.Experiment(title='test exp1',
institution_name='monash',
created_by=self.user,
)
exp.save()
ae1 = models.Author_Experiment(experiment=exp,
author='steve',
order=100)
ae1.save()
ae2 = models.Author_Experiment(experiment=exp,
author='russell',
order=1)
ae2.save()
ae3 = models.Author_Experiment(experiment=exp,
author='uli',
order=50)
ae3.save()
authors = exp.author_experiment_set.all()
# confirm that there are 2 authors
self.assertEqual(len(authors), 3)
self.assertTrue(ae1 in authors)
self.assertTrue(ae2 in authors)
self.assertTrue(ae3 == authors[1])
def _update_authors(self, data):
# For each author in the POST in a position
for data in self._parse_authors(data):
try:
# Get the current author for that position
o_ae = self.author_experiments[data['order']]
# Update the author form for that position with the new data
self.author_experiments[data['order']] = \
Author_Experiment(data=data,
instance=o_ae.instance)
except IndexError:
# Or create an author for that position
o_ae = Author_Experiment(data=data,
instance=models.Author_Experiment())
self.author_experiments.append(o_ae)
def _fill_authors(self, authors):
if self.instance:
o_author_experiments = \
self.instance.author_experiment_set.all()
else:
o_author_experiments = []
for num, author in enumerate(authors):
try:
o_ae = o_author_experiments[num]
except IndexError:
o_ae = models.Author_Experiment()
o_ae.experiment = self.instance
f = Author_Experiment(data={'author': author,
'order': num},
instance=o_ae)
self.author_experiments.append(f)
self.fields['authors'] = \
MultiValueCommaSeparatedField([author.fields['author'] for
author in self.author_experiments],
widget=CommaSeparatedInput())
def createhpcexperiment(request, user, dl):
from django.contrib.auth.models import User
from tardis.tardis_portal.views import _registerExperimentDocument
import os
import tempfile
#TODO
temp = tempfile.TemporaryFile()
for chunk in dl.chunks():
temp.write(chunk)
temp.seek(0)
metadata = {}
for line in temp:
key, value = line.split('~')
metadata[key] = value
temp.close()
author = metadata['Name']
title = metadata['Experiment']
ds_desc = metadata['Facility']
desc = metadata['Description']
fname = metadata['FolderName']
counter = metadata['Counter']
package = metadata['Package']
exp = Experiment(
title=title,
institution_name="RMIT University",
description=desc,
created_by=User.objects.get(id=user.pk),
)
exp.save()
eid = exp.id
# Store the author for the dataset
ae = models.Author_Experiment(experiment=exp, author=author, order='1')
ae.save()
auth_key = settings.DEFAULT_AUTH
try:
e = Experiment.objects.get(pk=eid)
except Experiment.DoesNotExist:
logger.exception(
'Experiment for eid %i in CreateHPCExperiment does not exist' %
eid)
acl = ExperimentACL(experiment=e,
pluginId=django_user,
entityId=str(user.id),
canRead=True,
canWrite=True,
canDelete=True,
isOwner=True,
aclOwnershipType=ExperimentACL.OWNER_OWNED)
acl.save()
folder_desc = "%s.%s.%s.%s" % (ds_desc.strip(), package.strip(),
fname.strip(), counter.strip())
logger.debug('folder_desc = %s' % folder_desc)
return eid, exp, folder_desc
def endElementNS(self, name, qname):
# just get the element name without the namespace
elName = name[1]
if elName == 'dmdSec':
self.inDmdSec = False
# if we currently processing an experiment structure, let's
# save the institution value before we finalise the experiment
if self.processExperimentStruct:
self.metsObject.institution = self.institution
# let's save the experiment in the DB
if self.tardisExpId:
self.modelExperiment = models.Experiment.objects.get(
pk=self.tardisExpId)
else:
self.modelExperiment = models.Experiment()
self.modelExperiment.id = self.tardisExpId
self.modelExperiment.url = self.metsObject.url
self.modelExperiment.approved = True
self.modelExperiment.title = self.metsObject.title
self.modelExperiment.institution_name = \
self.metsObject.institution
self.modelExperiment.description = self.metsObject.description
self.modelExperiment.start_time = self.metsObject.start_time
self.modelExperiment.end_time = self.metsObject.end_time
self.modelExperiment.created_by = self.createdBy
self.modelExperiment.save()
self.holder.experimentDatabaseId = self.modelExperiment.id
x = 0
for author in self.metsObject.authors:
author_experiment = models.Author_Experiment(
experiment=self.modelExperiment,
author=author, order=x)
author_experiment.save()
x = x + 1
elif self.processDatasetStruct:
# let's save the dataset in the DB
self.modelDataset = models.Dataset(
description=self.metsObject.title,
immutable=settings.IMMUTABLE_METS_DATASETS)
self.modelDataset.save()
self.modelDataset.experiments.add(self.modelExperiment)
self.modelDataset.save()
# let's also save the modelDataset in a dictionary so that we
# can look it up easily later on when we start processing
# the datafiles.
self.datasetLookupDict[self.metsObject.id] = self.modelDataset
self.metsObject = None
self.processExperimentStruct = False
self.processDatasetStruct = False
elif elName == 'title' and self.inDmdSec:
self.grabTitle = False
elif elName == 'startTime' and self.processExperimentStruct:
self.grabStartTime = False
elif elName == 'endTime' and self.processExperimentStruct:
self.grabEndTime = False
elif elName == 'url' and self.processExperimentStruct:
self.grabExperimentUrl = False
elif elName == 'abstract' and self.processExperimentStruct:
self.grabAbstract = False
elif elName == 'name' and self.processExperimentStruct:
self.inName = False
elif elName == 'namePart' and self.inName:
self.grabMightBeAuthor = False
elif elName == 'roleTerm' and self.inName:
self.grabRoleTerm = False
self.mightBeAuthor = None
elif elName == 'name' and self.inInstitution:
self.grabInstitution = False
elif elName == 'agent':
self.inInstitution = False
elif elName == 'amdSec':
# we're done processing the metadata entries
self.inAmdSec = False
# let's reset the cached experiment model object
self.modelExperiment = None
elif elName == 'techMD' and self.inAmdSec:
self.inTechMd = False
self.metadataId = None
self.metsObject = None
self.processMetadata = False
elif elName == 'xmlData' and self.inTechMd:
self.inXmlData = False
elif elName != self.xmlDataChildElement and \
self.customHandler is not None:
self.customHandler.endElement(elName)
elif elName == self.xmlDataChildElement and self.inXmlData:
if self.customHandler is not None:
self.tempMetadataHolder = self.customHandler.metadataDict
try:
schema = models.Schema.objects.get(
namespace__exact=self.elementNamespace)
# get the associated parameter names for the given schema
parameterNames = \
models.ParameterName.objects.filter(
schema__namespace__exact=schema.namespace).order_by('id')
# let's create a trigger holder which we can use to check
# if we still need to create another parameterset entry in the
# DB
createParamSetFlag = {'experiment': True, 'dataset': True,
'datafile': True}
datasetParameterSet = None
datafileParameterSet = None
if self.metadataId in self.holder.metadataMap:
for metsObject in self.holder.metadataMap[self.metadataId]:
self.metsObject = metsObject
metsObjectClassName = self.metsObject.__class__.__name__
if metsObjectClassName == 'Experiment':
if createParamSetFlag['experiment']:
# create a new parameter set for the metadata
parameterSet = \
models.ExperimentParameterSet(
schema=schema,
experiment=self.modelExperiment)
parameterSet.save()
# now let's process the experiment parameters
for parameterName in parameterNames:
if parameterName.name in \
self.tempMetadataHolder:
parameterValues = self.tempMetadataHolder[
parameterName.name]
self._saveParameters('ExperimentParameter',
parameterName, parameterValues,
parameterSet)
createParamSetFlag['experiment'] = False
else:
# this is not even allowed as there's only going
# to be one experiment per METS file
raise Exception('forbidden state!')
elif metsObjectClassName == 'Dataset':
if createParamSetFlag['dataset']:
dataset = self.datasetLookupDict[
self.metsObject.id]
# create a new parameter set for the
# dataset metadata
datasetParameterSet = \
models.DatasetParameterSet(schema=schema,
dataset=dataset)
datasetParameterSet.save()
# now let's process the dataset parameters
for parameterName in parameterNames:
if parameterName.name in \
self.tempMetadataHolder:
parameterValues = self.tempMetadataHolder[
parameterName.name]
self._saveParameters('DatasetParameter',
parameterName, parameterValues,
datasetParameterSet)
# disable creation for the next visit
createParamSetFlag['dataset'] = False
elif metsObjectClassName == 'Datafile':
# this will be a good time to save the
# "hard" metadata of this datafile so that
# when we start adding "soft" metadata
# parameters to it, we already have an
# entry for it in the DB
# look up the dataset this file belongs to
thisFilesDataset = self.datasetLookupDict[
self.metsObject.dataset.id]
# also check if the file already exists
datafile = thisFilesDataset.dataset_file_set.filter(
filename=self.metsObject.name, size=self.metsObject.size)
if datafile.count() == 0:
size = self.metsObject.size
if not self.metsObject.size:
size = 0
def checksum(obj, type_):
# Check if the checksum is of type
if obj.checksumType != type_:
return ''
checksum = obj.checksum.lower()
# Ensure the checksum is hexdecimal
if not re.match('[0-9a-f]+$', checksum):
return ''
# Get algorithm
try:
name = type_.replace('-','').lower()
alg = getattr(hashlib, name)
except:
return ''
# Check checksum is the correct length
hex_length = alg('').digest_size * 2
if hex_length != len(checksum):
return ''
# Should be valid checksum of given type
return checksum
sync_url, proto = get_sync_url_and_protocol(
self.syncRootDir,
self.metsObject.url)
self.modelDatafile = models.Dataset_File(
dataset=thisFilesDataset,
filename=self.metsObject.name,
url=sync_url,
size=size,
md5sum=checksum(self.metsObject, 'MD5'),
sha512sum=checksum(self.metsObject,
'SHA-512'),
protocol=proto)
self.modelDatafile.save()
else:
self.modelDatafile = thisFilesDataset.dataset_file_set.get(
filename=self.metsObject.name, size=self.metsObject.size)
# TODO: we need to note here that we are
# only creating a datafile entry in the DB
# for files that have corresponding
# metadata. if we are to create a file
# entry for files with no metadata, we'll
# need to get the unaccessed datafiles
# from datasetLookupDict.
if createParamSetFlag['datafile']:
# create a new parameter set for the metadata
datafileParameterSet = \
models.DatafileParameterSet(schema=schema,
dataset_file=self.modelDatafile)
datafileParameterSet.save()
# now let's process the datafile parameters
for parameterName in parameterNames:
if parameterName.name in \
self.tempMetadataHolder:
parameterValues = self.tempMetadataHolder[
parameterName.name]
self._saveParameters('DatafileParameter',
parameterName, parameterValues,
datafileParameterSet)
createParamSetFlag['datafile'] = False
except models.Schema.DoesNotExist:
logger.warning('unsupported schema being ingested ' +
self.elementNamespace)
# reset the current xmlData child element so that if a new
# parameter set is read, we can process it again
self.xmlDataChildElement = None
self.customHandler = None
elif elName == self.parameterName and \
self.xmlDataChildElement is not None:
# reset self.parameterName to None so the next parameter can be
# processed
self.parameterName = None
