How do I increase the maximum number of communities in a graph visualization?

Hi, May I please ask you two questions in graph visualization?

1. The maximum number of communities is 16 (the same as the number of COLOR elements); is there a reason why it’s limited to 16?
2. Do you plan to apply a lighter color to a label if the node is a dark color?

---

Is your feature request related to a problem? Please describe.

• I’m stuck with a limited number of communities
• If the node is dark, the label is hard to see

Describe the solution you’d like

• I don’t want the number of communities to be limited.
• I’d like to determine the color of the label based on the relative luminance of the node’s color.
• I’d like to apply a color map if possible.

Describe alternatives you’ve considered I’ve tried to create an implementation.

Additional context

G = nx.karate_club_graph()
coms = algorithms.louvain(G)
pos = nx.spring_layout(G, k=0.3)

# viz.plot_network_clusters(G, coms, pos, plot_overlaps=True, plot_labels=True, cmap=None)
# viz.plot_network_clusters(G, coms, pos, plot_overlaps=True, plot_labels=True, cmap="jet")
viz.plot_network_clusters(G, coms, pos, plot_overlaps=True, plot_labels=True, cmap=plt.cm.Wistia)

[GiulioRossetti]

That’s great! I didn’t notice the luminance fix (I only quickly skimmed the code, my fault!) 😃

Filtering out small communities should be trivial: the list that contains them in the CDlib clustering object is already ordered by (decreasing) size.

[eternity1984]

Thank you! I’ll send a pull request later.😄