def get(self, path):
try:
parent = self.sess.collections.get(dirname(path))
except CollectionDoesNotExist:
raise DataObjectDoesNotExist()
query = self.sess.query(DataObject)\
.filter(DataObject.name == basename(path))\
.filter(DataObject.collection_id == parent.id)
results = query.all()
if len(results) <= 0:
raise DataObjectDoesNotExist()
return iRODSDataObject(self, parent, results)
def setFileMetadata(self, destination_dir, file_name, metadata):
"""
This function sets the metadata for a given file
@param destination_dir, path to destination dir in iRODS
@param file_name, the name for the file in Galaxy
@param metadata, the metadata for the file (including provenance)
@return True if finished successfully
@throws DataObjectDoesNotExist if the file does not exists in iRODS
"""
# Step 1. Check if file exists
path = os.path.join(destination_dir, file_name)
obj = self.session.data_objects.get(path)
if obj == None:
raise DataObjectDoesNotExist("The object " + path +
" was not found at iRODS server.")
# Step 2. Clear previous metadata
obj.metadata.remove_all()
print("Setting metadata to file: " + str(metadata))
# Step 3. Set metadata
for key, value in metadata.iteritems():
obj.metadata.add(key, value)
return True
def get(self, path):
name = os.path.basename(path)
coll = self.session.collections.get(os.path.dirname(path))
for obj in coll.data_objects:
if obj.name == name:
return obj
raise DataObjectDoesNotExist(path)
def download_irods(path, output_folder=None, overwrite=False):
"""
Download file from ALTA using iRODS
:param path: Path to file on iRODS server
:param str output_folder: Output folder (Default: current directory)
:param bool overwrite: Overwrite output file if it already exists
"""
# check if iRODS tools are available
if not have_irods:
print("iRODS tools not available; reinstall arts_tools "
"with pip install --upgrade arts_tools[irods]")
return
# get the irods environment file
try:
env_file = os.environ['IRODS_ENVIRONMENT_FILE']
except KeyError:
env_file = os.path.expanduser('~/.irods/irods_environment.json')
if not os.path.isfile(env_file):
print(
"iRODS environment file not found. Set IRODS_ENVIRONMENT_FILE if the file is not in the default "
"location. Also make sure your iRODS environment is set up properly (with iinit)"
)
return
# create output folder and get full path to output file
if output_folder is None:
output_folder = os.getcwd()
os.makedirs(output_folder, exist_ok=True)
output_file = os.path.join(output_folder, os.path.basename(path))
# check if the file already exists
if os.path.isfile(output_file) and not overwrite:
print(f"File already exists: {output_file}")
return
if overwrite:
download_options = {FORCE_FLAG_KW: ''}
else:
download_options = {}
# create SSL context
ssl_context = ssl.create_default_context(purpose=ssl.Purpose.SERVER_AUTH,
cafile=None,
capath=None,
cadata=None)
# open an iRODS session
with iRODSSession(irods_env_file=env_file,
ssl_context=ssl_context) as session:
try:
session.data_objects.get(path, output_folder, **download_options)
except DataObjectDoesNotExist:
# add filename to error message
raise DataObjectDoesNotExist(f"File not found: {path}")
def get_collection_tree(self, base, path, sessions_cleanup=True):
"""
Lists the folders and files attributes at the input 'path'
Parameters
----------
base : str
The base path to validate ; eg. P000000001/C000000001
path : str
The collection's id; eg. P000000001/C000000001/SubFolder1/Experiment1/
sessions_cleanup: bool
If true, the session will be closed after retrieving the values.
Returns
-------
dict
The folders and files attributes at the requested path
"""
output = []
base_path = "/nlmumc/projects/" + base
absolute_path = "/nlmumc/projects/" + path
if not is_safe_path(base_path, absolute_path):
raise CAT_NO_ACCESS_PERMISSION
collection = self.session.collections.get(absolute_path)
for coll in collection.subcollections:
# query extra collection info: ctime
query = self.session.query(iRODSCollection).filter(
iRODSCollection.id == coll.id)
try:
result = query.one()
except NoResultFound:
raise CollectionDoesNotExist()
name = irods_basename(result[iRODSCollection.name])
ctime = result[iRODSCollection.create_time]
relative_path = path + "/" + name
folder_node = {
"name": name,
"path": relative_path,
"type": "folder",
"size": "--",
"rescname": "--",
"ctime": ctime.strftime("%Y-%m-%d %H:%M:%S"),
}
output.append(folder_node)
for data in collection.data_objects:
# query extra data info: ctime
query = self.session.query(DataObject).filter(
DataObject.id == data.id)
try:
result = query.first()
except NoResultFound:
raise DataObjectDoesNotExist()
ctime = result[DataObject.create_time]
relative_path = path + "/" + data.name
data_node = {
"name": data.name,
"path": relative_path,
"type": "file",
"size": data.size,
"rescname": data.resource_name,
"offlineResource": data.resource_name == "arcRescSURF01",
"ctime": ctime.strftime("%Y-%m-%d %H:%M:%S"),
}
output.append(data_node)
if sessions_cleanup:
self.session.cleanup()
return output