pyDKB/storages: implement RucioClient.get() method.

Original Rucio `Client` does not have `get()` method.

Utils/Dataflow/pyDKB/storages/atlas/rucio.py

@@ -5,7 +5,7 @@
 import os
 
 from ..client import Client
-from ..exceptions import StorageException
+from ..exceptions import (StorageException, NotFound)
 from pyDKB.common.misc import (log, logLevel)
 from pyDKB.common.misc import try_to_import
 
@@ -50,3 +50,61 @@ class RucioClient(Client, ParentClientClass):
     def __init__(self, *args, **kwargs):
         """ Initialize parent client class. """
         ParentClientClass.__init__(self, *args, **kwargs)
+
+    def get(self, oid, **kwargs):
+        """ Get dataset metadata.
+
+        Implementation of interface method `Clent.get()`.
+
+        :param oid: dataset name
+        :type oid: str
+        :param fields: list of requested metadata fields
+                       (None = all metadata)
+        :type fields: list
+
+        :return: dataset metadata
+        :rtype: dict
+        """
+        scope, name = self._scope_and_name(oid)
+        try:
+            result = self.get_metadata(scope=scope, name=name)
+        except ValueError, err:
+            raise StorageException("Failed to get metadata from Rucio: %s"
+                                   % err)
+        except RucioException, err:
+            if 'Data identifier not found' in str(err):
+                raise NotFound(scope=scope, name=name)
+            raise StorageException("Failed to get metadata from Rucio: %s"
+                                   % err)
+        if kwargs.get('fields') is not None:
+            result = {f: result.get(f, None) for f in kwargs['fields']}
+        return result
+
+    def _scope_and_name(self, dsn):
+        """ Construct normalized scope and dataset name.
+
+        As input accepts dataset names in two forms:
+        * dot-separated string: "<XXX>.<YYY>[.<...>]";
+        * dot-separated string with prefix: "<scope>:<XXX>.<YYY>[.<...>]".
+
+        In first case ID is taken as a canonical dataset name and scope is set
+        to its first field (or two first fields, if the ID starts with 'user'
+        or 'group').
+        In second case prefix is taken as scope, and removed from ID to get the
+        canonical dataset name.
+
+        :param dsn: dataset name
+        :type dsn: str
+
+        :return: scope, datasetname
+        :rtype: tuple
+        """
+        result = dsn.split(':')
+        if len(result) < 2:
+            splitted = dsn.split('.')
+            if dsn.startswith('user') or dsn.startswith('group'):
+                scope = '.'.join(splitted[0:2])
+            else:
+                scope = splitted[0]
+            result = (scope, dsn)
+        return result