Dear Professor Peixoto,
Could you please provide the code for generating Figure 9 in your paper <Inferring the mesoscale structure of layered, edge-valued and time-varying networks> ? That could be very helpful as a nice example for how to generate similar figures for my paper 
I have used following code to model a time varying (5 time point) weighted edge network.
state = gt.minimize_blockmodel_dl(
g,
state = gt.LayeredBlockState,
state_args=dict(
deg_corr=True,
overlap=True,
ec=g.ep.layer,
layers=True,
recs=[g.ep.weight],
rec_types=["real-normal"]
)
)
[bv,r1,r2,r3] = state.get_overlap_blocks()
for v in g.vertices():
community_ids = bv[v]
print(f"node {int(v)} is in: {community_ids}")
And I got the print for each node’s membership:
node 0 is in: array([ 555, 1631, 9441], dtype=int32)
node 1 is in: array([ 555, 1631, 9441], dtype=int32)
node 2 is in: array([ 555, 1631, 9441], dtype=int32)
node 3 is in: array([ 555, 1631, 9441], dtype=int32)
node 4 is in: array([ 555, 1631, 9441], dtype=int32)
node 5 is in: array([ 555, 1631, 9441], dtype=int32)
…
Could I know how to draw each layer’s figure in Figure 9 ? Should I do sth like this ?:
# only draw layer 0
layer_index = 0
sub = gt.GraphView(g, efilt=lambda e: g.ep.layer[e] == layer_index)
g = sub
state.draw(
edge_color=g.ep.weight.copy("double"),
edge_gradient=[],
edge_pen_width=gt.prop_to_size(g.ep.weight, 1, 5),
eorder=g.ep.weight,
output=f"layer_{layer_index}.svg"
)
Could I know how can I know which node’s membership belongs to which time point (layer) ? Should I do sth like this ?:
from collections import defaultdict
membership_per_layer = defaultdict(set)
edge_blocks_map = state.get_edge_blocks()
layer_prop = g.ep["layer"]
for e in g.edges():
L = layer_prop[e]
edge_halfedges = edge_blocks_map[e]
node0 = int(e.source())
node1 = int(e.target())
half0_groups = edge_halfedges[0]
half1_groups = edge_halfedges[1]
if isinstance(half0_groups, int):
membership_per_layer[(node0, L)].add(half0_groups)
else:
for gr in half0_groups:
membership_per_layer[(node0, L)].add(gr)
if isinstance(half1_groups, int):
membership_per_layer[(node1, L)].add(half1_groups)
else:
for gr in half1_groups:
membership_per_layer[(node1, L)].add(gr)
Thank you so much for your kind help !