def get_results_dir(result, request):
swiftsettings = getUtility(IRegistry).forInterface(ISwiftSettings)
# swift only if it is remote dataset. For blob and multi-species dataset, store locally.
# For other dataset type, store at swift if possible.
do_swift = IRemoteDataset.providedBy(result) or \
(not IMultiSpeciesDataset.providedBy(result) and \
not IBlobDataset.providedBy(result) and \
swiftsettings.storage_url)
if do_swift:
if swiftsettings.storage_url:
results_dir = 'swift+{storage_url}/{container}/{path}/'.format(
storage_url=swiftsettings.storage_url,
container=swiftsettings.result_container,
path=IUUID(result)
)
else:
raise Exception("Remote dataset requires swift url to be set")
else:
# if swift is not setup we use local storage
results_dir = 'scp://{uid}@{ip}:{port}{path}/'.format(
uid=pwd.getpwuid(os.getuid()).pw_name,
# FIXME: hostname from request is not good enough...
# need to get ip or host from plone_worker that does actual
# import
# store in registry?
# (is ok for testing)
# ip=get_public_ip(),
ip=get_hostname(request),
port=os.environ.get('SSH_PORT', 22),
path=tempfile.mkdtemp(prefix='result_import_')
)
return results_dir
def get_results_dir(result, request, childSpecies=False):
swiftsettings = getUtility(IRegistry).forInterface(ISwiftSettings)
# swift only if it is remote dataset. For blob and multi-species dataset, store locally.
# For other dataset type (including the child species of multispecies), store at swift if possible.
do_swift = IRemoteDataset.providedBy(result) or \
((childSpecies or (not IMultiSpeciesDataset.providedBy(result))) and \
not IBlobDataset.providedBy(result) and \
swiftsettings.storage_url)
if do_swift:
if swiftsettings.storage_url:
results_dir = 'swift+{storage_url}/{container}/{path}/'.format(
storage_url=swiftsettings.storage_url,
container=swiftsettings.result_container,
path=IUUID(result))
else:
raise Exception("Remote dataset requires swift url to be set")
else:
# if swift is not setup we use local storage
results_dir = 'scp://{uid}@{ip}:{port}{path}/'.format(
uid=pwd.getpwuid(os.getuid()).pw_name,
# FIXME: hostname from request is not good enough...
# need to get ip or host from plone_worker that does actual
# import
# store in registry?
# (is ok for testing)
# ip=get_public_ip(),
ip=get_hostname(request),
port=os.environ.get('SSH_PORT', 22),
path=tempfile.mkdtemp(prefix='result_import_'))
return results_dir
def getGenreSchemata(self):
schemata = []
md = IBCCVLMetadata(self.context)
genre = md.get('genre')
if genre in self.genre_interface_map:
schemata.append(self.genre_interface_map[genre])
if IBlobDataset.providedBy(self.context):
schemata.append(IBlobDataset)
if IRemoteDataset.providedBy(self.context):
schemata.append(IRemoteDataset)
return schemata
def getGenreSchemata(self):
schemata = []
md = IBCCVLMetadata(self.context)
genre = md.get('genre')
if genre in self.genre_interface_map:
schemata.append(self.genre_interface_map[genre])
if IBlobDataset.providedBy(self.context):
schemata.append(IBlobDataset)
if IRemoteDataset.providedBy(self.context):
schemata.append(IRemoteDataset)
return schemata
def __call__(self, **kw):
jt = IJobTracker(self.context)
# TODO: if state is empty check if there is a downloadable file
# Yes: COMPLETED
# No: FAILED
state = jt.state
if not state:
if IBlobDataset.providedBy(self.context):
# we have no state, may happen for imported datasets,
# let's check if we have a file
if self.context.file is not None:
state = 'COMPLETED'
else:
state = 'FAILED'
elif IRemoteDataset.providedBy(self.context):
if self.context.remoteUrl:
state = 'COMPLETED'
else:
state = 'FAILED'
return state
def _download_results(self, context, zfile):
# FIXME: This is a rather lengthy process, and should probably be turned into a background task... (maybe as part of a datamanager service?)
# 1. find all IBlobDataset/ IRemotedDataset/ IDataset objects within context
pc = getToolByName(context, 'portal_catalog')
brains = pc.searchResults(path='/'.join(context.getPhysicalPath()),
object_provides=[
IBlobDataset.__identifier__,
IRemoteDataset.__identifier__
])
metadata = {}
# the file/folder name for the zip
zfilename = context.title
# iterate over files and add to zip
for brain in brains:
content = brain.getObject()
if IBlobDataset.providedBy(content):
# If data is stored locally:
arcname = '/'.join((zfilename, 'data', content.file.filename))
# ob.file should be a NamedFile ... need to get fs name for that
blobfile = content.file.openDetached()
zfile.write(blobfile.name, arcname)
blobfile.close()
elif IRemoteDataset.providedBy(content):
# TODO: duplicate code from
remoteUrl = getattr(content, 'remoteUrl', None)
if remoteUrl is None:
raise NotFound(self, 'remoteUrl', self.request)
# get arcname from remoteUrl
arcname = '/'.join(
(zfilename, 'data', os.path.basename(remoteUrl)))
# FIXME: should check dataset downloaiable flag here,
# but assumption is, that this function can only be called on an experiment result folder....
# TODO: duplicate code in browser/dataset.py:RemoteDatasetDownload.__call__
# TODO: may not work in general... it always uses swift as remote url
tool = getUtility(ISwiftUtility)
try:
url = tool.generate_temp_url(url=remoteUrl)
except:
url = remoteUrl
# url is now the location from which we can fetch the file
temp_file, _ = urlretrieve(url)
zfile.write(temp_file, arcname)
os.remove(temp_file)
else:
# unknown type of Dataset
# just skip it
# TODO: Log warning or debug?
continue
metadata[arcname] = getdsmetadata(content)
# all files are in ....
# TODO: add experiment result metadata
# put metadata into zip
# provenance data stored on result container
provdata = IProvenanceData(context)
if not provdata.data is None:
zfile.writestr('/'.join((zfilename, 'prov.ttl')),
provdata.data.encode('utf-8'))
# add experiment metadata
expmetadata = IExperimentMetadata(context)
if not expmetadata.data is None:
zfile.writestr('/'.join((zfilename, 'expmetadata.txt')),
expmetadata.data.encode('utf-8'))
# add mets.xml
metsview = getMultiAdapter((context, self.request), name="mets.xml")
zfile.writestr('/'.join((zfilename, 'mets.xml')),
metsview.render().encode('utf-8'))
# add experiment parameters
params = IExperimentParameter(context)
if not params.data is None:
zfile.writestr('/'.join((zfilename, 'params.json')),
params.data.encode('utf-8'))
def _download_results(self, context, zfile):
# FIXME: This is a rather lengthy process, and should probably be turned into a background task... (maybe as part of a datamanager service?)
# 1. find all IBlobDataset/ IRemotedDataset/ IDataset objects within context
pc = getToolByName(context, 'portal_catalog')
brains = pc.searchResults(path='/'.join(context.getPhysicalPath()),
object_provides=[IBlobDataset.__identifier__,
IRemoteDataset.__identifier__])
metadata = {}
# the file/folder name for the zip
zfilename = context.title
# iterate over files and add to zip
for brain in brains:
content = brain.getObject()
if IBlobDataset.providedBy(content):
# If data is stored locally:
arcname = '/'.join((zfilename, 'data', content.file.filename))
# ob.file should be a NamedFile ... need to get fs name for that
blobfile = content.file.openDetached()
zfile.write(blobfile.name, arcname)
blobfile.close()
elif IRemoteDataset.providedBy(content):
# TODO: duplicate code from
remoteUrl = getattr(content, 'remoteUrl', None)
if remoteUrl is None:
raise NotFound(self, 'remoteUrl', self.request)
# get arcname from remoteUrl
arcname = '/'.join((zfilename, 'data', os.path.basename(remoteUrl)))
# FIXME: should check dataset downloaiable flag here,
# but assumption is, that this function can only be called on an experiment result folder....
# TODO: duplicate code in browser/dataset.py:RemoteDatasetDownload.__call__
# TODO: may not work in general... it always uses swift as remote url
tool = getUtility(ISwiftUtility)
try:
url = tool.generate_temp_url(url=remoteUrl)
except:
url = remoteUrl
# url is now the location from which we can fetch the file
temp_file, _ = urlretrieve(url)
zfile.write(temp_file, arcname)
os.remove(temp_file)
else:
# unknown type of Dataset
# just skip it
# TODO: Log warning or debug?
continue
metadata[arcname] = getdsmetadata(content)
# all files are in ....
# TODO: add experiment result metadata
# put metadata into zip
# provenance data stored on result container
provdata = IProvenanceData(context)
if not provdata.data is None:
zfile.writestr('/'.join((zfilename, 'prov.ttl')),
provdata.data.encode('utf-8'))
# add experiment metadata
expmetadata = IExperimentMetadata(context)
if not expmetadata.data is None:
zfile.writestr('/'.join((zfilename, 'expmetadata.txt')),
expmetadata.data.encode('utf-8'))
# add mets.xml
metsview = getMultiAdapter((context, self.request), name="mets.xml")
zfile.writestr('/'.join((zfilename, 'mets.xml')),
metsview.render().encode('utf-8'))
# add experiment parameters
params = IExperimentParameter(context)
if not params.data is None:
zfile.writestr('/'.join((zfilename, 'params.json')),
params.data.encode('utf-8'))
def __call__(self):
# FIXME: This is a rather lengthy process, and should probably be turned into a background task... (maybe as part of a datamanager service?)
# 1. find all IBlobDataset/ IRemotedDataset/ IDataset objects within context
pc = getToolByName(self.context, 'portal_catalog')
brains = pc.searchResults(path='/'.join(self.context.getPhysicalPath()),
object_provides=[IBlobDataset.__identifier__,
IRemoteDataset.__identifier__])
fname = None
try:
# create tmp file
fd, fname = tempfile.mkstemp()
fo = os.fdopen(fd, 'w')
zfile = zipfile.ZipFile(fo, 'w')
metadata = {}
# the file/folder name for the zip
zfilename = self.context.title
# iterate over files and add to zip
for brain in brains:
content = brain.getObject()
if IBlobDataset.providedBy(content):
# If data is stored locally:
arcname = '/'.join((zfilename, 'data', content.file.filename))
# ob.file should be a NamedFile ... need to get fs name for that
blobfile = content.file.openDetached()
zfile.write(blobfile.name, arcname)
blobfile.close()
elif IRemoteDataset.providedBy(content):
# TODO: duplicate code from
remoteUrl = getattr(content, 'remoteUrl', None)
if remoteUrl is None:
raise NotFound(self, 'remoteUrl', self.request)
# get arcname from remoteUrl
arcname = '/'.join((zfilename, 'data', os.path.basename(remoteUrl)))
# FIXME: should check dataset downloaiable flag here,
# but assumption is, that this function can only be called on an experiment result folder....
# TODO: duplicate code in browser/dataset.py:RemoteDatasetDownload.__call__
# TODO: may not work in general... it always uses swift as remote url
tool = getUtility(ISwiftUtility)
try:
url = tool.generate_temp_url(url=remoteUrl)
except:
url = remoteUrl
# url is now the location from which we can fetch the file
temp_file, _ = urlretrieve(url)
zfile.write(temp_file, arcname)
os.remove(temp_file)
else:
# unknown type of Dataset
# just skip it
# TODO: Log warning or debug?
continue
metadata[arcname] = getdsmetadata(content)
# all files are in ....
# TODO: add experiment result metadata
# put metadata into zip
# provenance data stored on result container
provdata = IProvenanceData(self.context)
if not provdata.data is None:
zfile.writestr('/'.join((zfilename, 'prov.ttl')),
provdata.data)
# add mets.xml
metsview = getMultiAdapter((self.context, self.request), name="mets.xml")
zfile.writestr('/'.join((zfilename, 'mets.xml')),
metsview.render())
# finish zip file
zfile.close()
fo.close()
# create response
self.request.response.setHeader('Content-Type', 'application/zip')
self.request.response.setHeader('Content-Disposition', 'attachment; filename="{}.zip"'.format(zfilename))
self.request.response.setHeader('Content-Length', '{}'.format(os.path.getsize(fname)))
return tmpfile_stream_iterator(fname)
except Exception as e:
# something went wrong ...
# clean up and re-raise
if os.path.exists(fname):
os.remove(fname)
raise e
