graphistry/pygraphistry

Question: what community detection coloring is being used

Closed this issue · 2 comments

Hi there,

When plotting using encode_point_color on a non-existent categorical field, Graphistry assigns node colors that closely match communities that I would like to be using.

I have looked through the code but haven't been able to find which community detection algo (and params) is being used on this failure case. Note, they are not random, which is what happens when I don't call encode_point_color.

I have also checked the g object and don't see and info about the layout or colors.

Any help would be greatly appreciated.

Thanks

I was able to roughly reproduce these communities using Louvain with max 1 level

Yes, exactly, we do a local vote to ensure speed by default

compute_cugraph(alg) can enrich quickly with alternative community labels