如何在 Graph Networkx 中找到具有公共节点的边？

Question

I have this graph:

G = nx.MultiGraph()

source_nodes = ['A', 'A', 'B', 'D']
target_nodes = ['B', 'C', 'D', 'E']

for u,v in zip(source_nodes, target_nodes):
    G.add_edge(u, v)

and the edges I get are:

(A, B)
(A, C)
(B, D)
(D, E)

if the target node of an edge is the source node from another edge, I would like to create an edge between them. For instance, for the above example the new edges will be:

(A, D) and (A, E) .

I'm new in Networkx and I'm not sure how to do that...

Answer 1

1. Firstly, you use MultiGraph . It is not directed and doesn't store info about sources and targets (they become equal connected nodes). You should use DiGraphs or MultiDiGraphs to store info in a directed graph.

2. NetworkX DiGraphs and MultiDiGraphs has functions predecessors and successors , that can be applied to a particular node. You can use them to check if an edge from predecessor to successor exists, and if not, create it.

import networkx as nx

G = nx.DiGraph()  # Not a MultiGraph! (can be MultiDiGraph)

source_nodes = ['A', 'A', 'B', 'D']
target_nodes = ['B', 'C', 'D', 'E']

for u,v in zip(source_nodes, target_nodes):
    G.add_edge(u, v)

# For each node
for n in G.nodes:
    # We check all predecessors and successors
    for p in G.predecessors(n):
        for s in G.successors(n):
            # If graph has no edge from predecessor to successor
            if (p, s) not in G.edges:
                print(p, s)  # Print, can be removed
                # We add this node
                G.add_edge(p, s)

If you want to create edges not for all nodes, but for a set of nodes, you can replace G.nodes: in the line for n in G.nodes: with a given set of nodes, like:

for n in [1, 2, 3]:

If you want to create edges only for one node, you don't need the main loop:

NODE_ID = 2

# We check all predecessors and successors of a node
for p in G.predecessors(NODE_ID):
    for s in G.successors(NODE_ID):
        # If graph has no edge from predecessor to successor
        if (p, s) not in G.edges:
            print(p, s)  # Print, can be removed
            # We add this node
            G.add_edge(p, s)