[networkx] Use concrete node types for the atlas, cograph and duplication generators (#16793)

Author: dannyshafer

stubs/networkx/networkx/generators/atlas.pyi

@@ -1,5 +1,4 @@
 import sys
-from _typeshed import Incomplete
 from typing import Final
 
 from networkx.utils.backends import _dispatchable
@@ -17,6 +16,6 @@ NUM_GRAPHS: Final = 1253
 ATLAS_FILE: Final[Traversable]
 
 @_dispatchable
-def graph_atlas(i: int) -> Graph[Incomplete]: ...
+def graph_atlas(i: int) -> Graph[int]: ...
 @_dispatchable
-def graph_atlas_g() -> list[Graph[Incomplete]]: ...
+def graph_atlas_g() -> list[Graph[int]]: ...

stubs/networkx/networkx/generators/cographs.pyi

@@ -1,9 +1,7 @@
-from _typeshed import Incomplete
-
 from networkx.classes.graph import Graph
 from networkx.utils.backends import _dispatchable
 
 __all__ = ["random_cograph"]
 
 @_dispatchable
-def random_cograph(n: int, seed=None) -> Graph[Incomplete]: ...
+def random_cograph(n: int, seed=None) -> Graph[int]: ...

stubs/networkx/networkx/generators/duplication.pyi

@@ -1,12 +1,10 @@
-from _typeshed import Incomplete
-
 from networkx.utils.backends import _dispatchable
 
 from ..classes.graph import Graph
 
 __all__ = ["partial_duplication_graph", "duplication_divergence_graph"]
 
 @_dispatchable
-def partial_duplication_graph(N: int, n: int, p: float, q: float, seed=None): ...
+def partial_duplication_graph(N: int, n: int, p: float, q: float, seed=None) -> Graph[int]: ...
 @_dispatchable
-def duplication_divergence_graph(n: int, p: float, seed=None) -> Graph[Incomplete]: ...
+def duplication_divergence_graph(n: int, p: float, seed=None) -> Graph[int]: ...