def parse_pfns(self, pfns):
"""
Splits the given PFN into the parts known by the protocol. During parsing the PFN is also checked for
validity on the given RSE with the given protocol.
:param pfn: a fully qualified PFN
:returns: a dict containing all known parts of the PFN for the protocol e.g. scheme, path, filename
:raises RSEFileNameNotSupported: if the provided PFN doesn't match with the protocol settings
"""
ret = dict()
pfns = [pfns] if isinstance(pfns, string_types) else pfns
for pfn in pfns:
parsed = urlparse.urlparse(pfn)
if parsed.path.startswith('/srm/managerv2') or\
parsed.path.startswith('/srm/managerv1') or\
parsed.path.startswith('/srm/v2/server'):
scheme, hostname, port, service_path, path = re.findall(r"([^:]+)://([^:/]+):?(\d+)?([^:]+=)?([^:]+)", pfn)[0]
else:
scheme = parsed.scheme
hostname = parsed.netloc.partition(':')[0]
port = parsed.netloc.partition(':')[2]
path = parsed.path
service_path = ''
# force type conversion
try:
port = int(port)
except:
port = ''
if self.attributes['hostname'] != hostname and\
self.attributes['hostname'] != scheme + "://" + hostname:
raise exception.RSEFileNameNotSupported('Invalid hostname: provided \'%s\', expected \'%s\'' % (hostname,
self.attributes['hostname']))
if port != '' and str(self.attributes['port']) != str(port):
raise exception.RSEFileNameNotSupported('Invalid port: provided \'%s\', expected \'%s\'' % (port,
self.attributes['port']))
elif port == '':
port = self.attributes['port']
if not path.startswith(self.attributes['prefix']):
raise exception.RSEFileNameNotSupported('Invalid prefix: provided \'%s\', expected \'%s\'' % ('/'.join(path.split('/')[0:len(self.attributes['prefix'].split('/')) - 1]),
self.attributes['prefix'])) # len(...)-1 due to the leading '/
# Spliting path into prefix, path, filename
prefix = self.attributes['prefix']
path = path.partition(self.attributes['prefix'])[2]
name = path.split('/')[-1]
path = '/' + '/'.join(path.split('/')[:-1]) if not self.rse['staging_area'] else None
if path != '/' and path[:-1] != '/':
path += '/'
ret[pfn] = {'scheme': scheme, 'port': port, 'hostname': hostname,
'path': path, 'name': name, 'prefix': prefix,
'web_service_path': service_path}
return ret
def parse_pfns(self, pfns):
"""
Splits the given PFN into the parts known by the protocol. It is also checked if the provided protocol supportes the given PFNs.
:param pfns: a list of a fully qualified PFNs
:returns: dic with PFN as key and a dict with path and name as value
:raises RSEFileNameNotSupported: if the provided PFN doesn't match with the protocol settings
"""
ret = dict()
pfns = [pfns] if isinstance(pfns, string_types) else pfns
logging.debug('... Beginning GlobusRSEProtocol.parse_pfns ... ')
for pfn in pfns:
parsed = urlparse(pfn)
scheme = parsed.scheme
hostname = parsed.netloc.partition(':')[0]
port = int(parsed.netloc.partition(':')[2]) if parsed.netloc.partition(':')[2] != '' else 0
while '//' in parsed.path:
parsed = parsed._replace(path=parsed.path.replace('//', '/'))
path = parsed.path
# Protect against 'lazy' defined prefixes for RSEs in the repository
if not self.attributes['prefix'].startswith('/'):
self.attributes['prefix'] = '/' + self.attributes['prefix']
if not self.attributes['prefix'].endswith('/'):
self.attributes['prefix'] += '/'
if self.attributes['hostname'] != hostname:
if self.attributes['hostname'] != 'localhost': # In the database empty hostnames are replaced with localhost but for some URIs (e.g. file) a hostname is not included
raise exception.RSEFileNameNotSupported('Invalid hostname: provided \'%s\', expected \'%s\'' % (hostname, self.attributes['hostname']))
if self.attributes['port'] != port:
raise exception.RSEFileNameNotSupported('Invalid port: provided \'%s\', expected \'%s\'' % (port, self.attributes['port']))
if not path.startswith(self.attributes['prefix']):
raise exception.RSEFileNameNotSupported('Invalid prefix: provided \'%s\', expected \'%s\'' % ('/'.join(path.split('/')[0:len(self.attributes['prefix'].split('/')) - 1]),
self.attributes['prefix'])) # len(...)-1 due to the leading '/
# Spliting parsed.path into prefix, path, filename
prefix = self.attributes['prefix']
path = path.partition(self.attributes['prefix'])[2]
name = path.split('/')[-1]
path = '/'.join(path.split('/')[:-1])
if not path.startswith('/'):
path = '/' + path
if path != '/' and not path.endswith('/'):
path = path + '/'
ret[pfn] = {'path': path, 'name': name, 'scheme': scheme, 'prefix': prefix, 'port': port, 'hostname': hostname, }
return ret
def split_pfn(self, pfn):
"""
Splits the given PFN into the parts known by the protocol. During parsing the PFN is also checked for
validity on the given RSE with the given protocol.
:param pfn: a fully qualified PFN
:returns: a dict containing all known parts of the PFN for the protocol e.g. scheme, hostname, port, prefix, path, filename
:raises RSEFileNameNotSupported: if the provided PFN doesn't match with the protocol settings
"""
parsed = urlparse(pfn)
ret = dict()
ret['scheme'] = parsed.scheme
ret['hostname'] = parsed.netloc.partition(':')[0]
ret['port'] = int(parsed.netloc.partition(':')
[2]) if parsed.netloc.partition(':')[2] != '' else 0
ret['path'] = parsed.path
# Protect against 'lazy' defined prefixes for RSEs in the repository
self.rse['prefix'] = '' if self.rse['prefix'] is None else self.rse[
'prefix']
if not self.rse['prefix'].startswith('/'):
self.rse['prefix'] = '/' + self.rse['prefix']
if not self.rse['prefix'].endswith('/'):
self.rse['prefix'] += '/'
if self.rse['hostname'] != ret['hostname']:
raise exception.RSEFileNameNotSupported(
'Invalid hostname: provided \'%s\', expected \'%s\'' %
(ret['hostname'], self.rse['hostname']))
if not ret['path'].startswith(self.rse['prefix']):
raise exception.RSEFileNameNotSupported(
'Invalid prefix: provided \'%s\', expected \'%s\'' %
('/'.join(ret['path'].split('/')
[0:len(self.rse['prefix'].split('/')) - 1]),
self.rse['prefix'])) # len(...)-1 due to the leading '/
# Spliting parsed.path into prefix, path, filename
ret['prefix'] = self.rse['prefix']
ret['path'] = ret['path'].partition(self.rse['prefix'])[2]
ret['name'] = ret['path'].split('/')[-1]
ret['path'] = ret['path'].partition(ret['name'])[0]
return ret