How to DiGraph in networkx reserving weight of multiple edges and self-loops when removing all the multiple edges and self-loops

Question

I created a graph using MultiDiGraph function in networkx and add Weights to Edges by Frequency of Edge Occurance. Now I am thinking to create a DiGraph graph and remove all the multiple edges and self-loops while still reserving the weight of multiple edges and self-loops. For example, if one of the edge occurs 5 times in the MultiDiGraph and the corresponding weight is 5, then the weight of this edge should still be 5 when creating the DiGraph and removing all the multiple edges and self-loops. How can do to make that happen? Thank you so much!

create two graphs separately

G1 = nx.MultiDiGraph()
G1.add_edges_from(
    [(3,4),(3,5),(3,7),(4,7),(6,7),(4,5),(5,6),(3,6),(4,5),(4,5),(6,3),(3,3)],
    color='red'
)

G2 = nx.MultiDiGraph()
G2.add_edges_from(
    [(2,5),(2,8),(4,8),(6,8),(4,5),(4,5),(5,6),(2,6),(6,2)],
    color='red'
)

Extract union of nodes and edges of these two graphs

union_nodes=list(set(list(G1.nodes)+list(G2.nodes)))
union_edges=list(list(G1.edges())+list(G2.edges()))

create a new graph combing these two graphs

G=nx.MultiDiGraph() 

G.add_nodes_from(union_nodes)

G.add_edges_from(union_edges)

Adding weight to edge by the frequency of edge occurrence

from collections import Counter
c = Counter(G.edges())

for u, v, d in G.edges(data=True):
    d['weight'] = c[u, v]

print(list(G.edges(data=True)))

nx.draw_networkx(G, width=[d['weight'] for _, _, d in G.edges(data=True)])

Maybe edit your question to add some of the code you've tried so far. — Edgar Ramírez Mondragón
Thank you! I edited that and added some [email protected]ón — Fan Yang

Edgar Ramírez Mondragón Edgar Ramírez Mondragón · Accepted Answer · 2018-10-20T14:39:41

You can create the edges of the DiGraph directly assigning them the corresponding weight:

c = Counter(G.edges())
simple_digraph = nx.DiGraph()

for u, v, d in G.edges(data=True):
    # avoid repeating edges and self-loops
    if not simple_digraph.has_edge(u, v) and u != v:
        simple_digraph.add_edge(u, v, weight=c[u, v])

print(list(simple_digraph.edges(data=True)))

Output:

[
     (2, 5, {'weight': 1}), (2, 8, {'weight': 1}), (2, 6, {'weight': 1}),
     (5, 6, {'weight': 2}), (6, 7, {'weight': 1}), (6, 3, {'weight': 1}),
     (6, 8, {'weight': 1}), (6, 2, {'weight': 1}), (3, 4, {'weight': 1}),
     (3, 5, {'weight': 1}), (3, 7, {'weight': 1}), (3, 6, {'weight': 1}),
     (4, 7, {'weight': 1}), (4, 5, {'weight': 5}), (4, 8, {'weight': 1})
]

How to DiGraph in networkx reserving weight of multiple edges and self-loops when removing all the multiple edges and self-loops

1 Answers