Ignore non-connected nodes in NetworkX Graph

Question

I have a network diagram where a node is connected to another node which in-turn is connected to another node, worked through a logic to ignore when there is only one neighbour but that is still leaving few bits back. The attached diagram does a better job explaining it

I need to only keep where a red-node is connected to a blue-node or another red-node through a green node. Data used

for u, v in CA_new.edges:
    
    if len(list(CA_new.neighbors(v))) == 1 and len(list(CA_new.neighbors(u))) == 1:
        removeNodeslist.append(v)
        removeNodeslist.append(u)
    else:
        keepNodeslist.append(v)
        keepNodeslist.append(u)
CA_new.remove_nodes_from(removeNodeslist)

Asked By: john_sab1

||

Source

Answer 1

From your comments, it sounds like you’re looking for the connected components that contain at least two parent nodes, and at least one of those parents has a colored edge. With that in mind, you can do something like the following:

import numpy as np
import networkx as nx
import pandas as pd

data = {
    'Parent': list("EEABEHDILGKCDBLLFBCCJ"),
    'Child':  ["X1","X2","Y1","Y1","Y1","M1","N3","N4","N5","N7","N8","M1","M2",
               "M3","M4","M5","M6","M7","M8","M9","P7"],
    'Colour': list("NNNNYYNYNNNYNNNNYNNNN")
}
df = pd.DataFrame(data)

G = nx.from_pandas_edgelist(df, source = 'Parent', target = 'Child')
parent_set = set(df['Parent'])
colored_parent_set = set(df.loc[df['Colour']=='Y','Parent'])

node_set = set()
for comp in nx.connected_components(G):
    if (len(comp & parent_set) >= 2 and
        comp & colored_parent_set):
        node_set |= comp

H = G.subgraph(node_set)

colors = ['limegreen']*len(H)
for i,n in enumerate(H.nodes):
    if n in colored_parent_set:
        colors[i] = "red"
    elif n in parent_set:
        colors[i] = "deepskyblue"

nx.draw(H, node_color = colors, with_labels = True)

Here’s the result that I get:

Answered By: Ben Grossmann

Ignore non-connected nodes in NetworkX Graph

Question:

Answers: