NetworkXGraphEdgeDescriptor: Use __set_name__.

Still need the __init__ use to handle a single case of overriding the name.

Signed-off-by: Chris PeBenito <chpebeni@linux.microsoft.com>

setools/descriptors.py

@@ -189,8 +189,8 @@ class NetworkXGraphEdgeDescriptor(ABC):
     """
     Descriptor abstract base class for NetworkX graph edge attributes.
 
-    Parameter:
-    name        The edge property name
+    Keyword Parameter:
+    name        Override the graph edge property name.
 
     Instance class attribute use (obj parameter):
     G           The NetworkX graph
@@ -198,8 +198,11 @@ class NetworkXGraphEdgeDescriptor(ABC):
     target      The edge's target node
     """
 
-    def __init__(self, propname: str) -> None:
-        self.name = propname
+    def __init__(self, propname: Optional[str] = None) -> None:
+        self.override_name = propname
+
+    def __set_name__(self, owner, name):
+        self.name = self.override_name if self.override_name else name
 
     def __get__(self, obj, objtype=None):
         if obj is None:

setools/dta.py

@@ -595,13 +595,13 @@ class Edge:
                 The default is False.
     """
 
-    transition = EdgeAttrList('transition')
-    setexec = EdgeAttrList('setexec')
-    dyntransition = EdgeAttrList('dyntransition')
-    setcurrent = EdgeAttrList('setcurrent')
-    entrypoint = EdgeAttrDict('entrypoint')
-    execute = EdgeAttrDict('execute')
-    type_transition = EdgeAttrDict('type_transition')
+    transition = EdgeAttrList()
+    setexec = EdgeAttrList()
+    dyntransition = EdgeAttrList()
+    setcurrent = EdgeAttrList()
+    entrypoint = EdgeAttrDict()
+    execute = EdgeAttrDict()
+    type_transition = EdgeAttrDict()
 
     def __init__(self, graph, source: Type, target: Type, create: bool = False) -> None:
         self.G = graph

setools/infoflow.py

@@ -405,7 +405,7 @@ class InfoFlowStep:
                 The default is False.
     """
 
-    rules = EdgeAttrList('rules')
+    rules = EdgeAttrList()
 
     # use capacity to store the info flow weight so
     # we can use network flow algorithms naturally.