def __init__(self, graph: Union[PyGraph, "nx.Graph"]) -> None:
"""
Args:
graph: Input graph for Lattice. Can be provided as ``retworkx.PyGraph``, which is
used internally, or, for convenience, as ``networkx.Graph``. The graph
cannot be a multigraph.
Raises:
ValueError: If the input graph is a multigraph.
ValueError: If the graph edges are non-numeric.
"""
if not isinstance(graph, PyGraph):
_optionals.HAS_NETWORKX.require_now(
"Lattice construction from networkx.Graph")
graph = networkx_converter(graph)
if graph.multigraph:
raise ValueError(
f"Invalid `graph.multigraph` {graph.multigraph} is given. "
"`graph.multigraph` must be `False`.")
# validate the edge weights
for edge_index, edge in graph.edge_index_map().items():
weight = edge[2]
if weight is None or weight == {}:
# None or {} is updated to be 1
graph.update_edge_by_index(edge_index, 1)
elif not isinstance(weight, numbers.Number):
raise ValueError(
f"Unsupported weight {weight} on edge with index {edge_index}."
)
self._graph = graph
self.pos: Optional[dict] = None
def test_random_k_out_graph(self):
g = networkx.random_k_out_graph(100, 50, 3.14159, True, 42)
out_graph = retworkx.networkx_converter(g)
self.assertIsInstance(out_graph, retworkx.PyDiGraph)
self.assertEqual(out_graph.nodes(), list(g.nodes))
self.assertEqual(out_graph.weighted_edge_list(),
list(g.edges(data=True)))
self.assertEqual(out_graph.multigraph, g.is_multigraph())
def test_cubical_multigraph(self):
g = networkx.cubical_graph(networkx.MultiGraph)
out_graph = retworkx.networkx_converter(g)
self.assertIsInstance(out_graph, retworkx.PyGraph)
self.assertEqual(out_graph.nodes(), list(g.nodes))
self.assertEqual(out_graph.weighted_edge_list(),
list(g.edges(data=True)))
self.assertEqual(out_graph.multigraph, g.is_multigraph())
def test_empty_directed_graph(self):
g = networkx.DiGraph()
out_graph = retworkx.networkx_converter(g)
self.assertIsInstance(out_graph, retworkx.PyDiGraph)
self.assertEqual(out_graph.nodes(), list(g.nodes))
self.assertEqual(out_graph.weighted_edge_list(),
list(g.edges(data=True)))
self.assertEqual(out_graph.multigraph, g.is_multigraph())
def test_undirected_gnm_graph(self):
g = networkx.gnm_random_graph(10, 10, seed=42)
out_graph = retworkx.networkx_converter(g)
self.assertIsInstance(out_graph, retworkx.PyGraph)
self.assertEqual(out_graph.nodes(), list(g.nodes))
self.assertEqual(out_graph.weighted_edge_list(),
list(g.edges(data=True)))
self.assertEqual(out_graph.multigraph, g.is_multigraph())
def __call__(self, D: np.ndarray, M: np.ndarray, time_step: int,
sir: np.ndarray) -> np.ndarray:
pressured_nodes = self._pressure_handler(sir)
D = self._call(D, M, time_step, pressured_nodes)
R = M - D
# Collect Data
self._last_pressured_nodes = pressured_nodes
self._last_removed_edges = R
G = nx.from_numpy_array(D)
self._last_comps = tuple(nx.connected_components(G))
self._last_diameter = np.max(
rx.distance_matrix(rx.networkx_converter(G)))
self._last_perc_edges_removed = np.sum(R, axis=0) / np.sum(M, axis=0)
return D
def __call__(self, D: np.ndarray, M: np.ndarray, time_step: int,
sir: np.ndarray) -> np.ndarray:
final_ND = np.zeros(D.shape, dtype=bool)
final_pressured_nodes = np.zeros(D.shape[0], dtype=bool)
for behavior in self._behaviors:
pressured_nodes = behavior._pressure_handler(sir)
ND = behavior._call(D, M, time_step, pressured_nodes)
final_ND = final_ND | (ND > 0)
final_pressured_nodes = final_pressured_nodes | pressured_nodes
# Collect Data
R = M - final_ND
self._last_pressured_nodes = final_pressured_nodes
self._last_removed_edges = R
G = nx.from_numpy_array(final_ND)
self._last_comps = tuple(nx.connected_components(G))
self._last_diameter = np.max(
rx.distance_matrix(rx.networkx_converter(G)))
self._last_perc_edges_removed = np.sum(R, axis=0) / np.sum(M, axis=0)
return ND
