Python Igraph community cluster colors

Question

I want to have clusters in different colours after community infomap, but problem is when I deleted single nodes it makes a mess an each node is different color or everything is red. How to do that in python?

Code:

E = ig.Graph(edges)
E.vs\['label'\] = labels
degree = 0
community = E.community_infomap()
cg = community.graph
singletons = cg.vs.select(_degree = 0)
cg.delete_vertices(singletons)
color_list =['red','blue','green','cyan','pink','orange','grey','yellow','white','black','purple' ]

ig.plot(cg)

Answer 1

It is not clear how did you try to assign colors to vertices. You should be aware that igraph reindexes vertices and edges upon deletion and addition of either vertices or edges. This reindexing should be considered unpredictable, the only things we know that indices go from 0 to n-1 at all time, and attributes remain assigned to the correct vertex or edge. Considering these, you can either do the deletion before or after the community detection, only you need to assign colors to a vertex attribute:

import igraph
g = igraph.Graph.Barabasi(n = 20, m = 1)
i = g.community_infomap()
pal = igraph.drawing.colors.ClusterColoringPalette(len(i))
g.vs['color'] = pal.get_many(i.membership)
igraph.plot(g)

Now let's see what happens if we delete a vertex:

colors_original = pal.get_many(i.membership)
g.delete_vertices([7])
# the clustering object is still the same length
# (so it is not valid any more, you can't be sure
# if the i.th vertex in the clustering is the 
# i.th one in the graph)
len(i) # 20
# while the graph has less vertices
g.vcount() # 19
# if we plot this with the original colors, indeed we get a mess:
igraph.plot(g, vertex_color = colors_original)

But the colors in the g.vs['color'] vertex attribute are still correct, they show the clusters, only the deleted vertex is missing (from the dark blue cluster):

igraph.plot(g,
            vertex_color = g.vs['color']) # this is automatic, I am 
                                          # writing here explicitely only to be clear