[Feature Request] sample weighted graph from fitted model
When you fit a model based on some property, i.e. edge weight:
state = gt.minimize_blockmodel_dl(gt_subgraph,state_args=dict(recs=[gt_subgraph.ep.weight],rec_types=["discrete-poisson"]))
if you later want to sample a graph from that state, the resulting graph will lack the properties of the original graph. For instance:
bs = state.sample_graph(canonical=True, multigraph=False, self_loops=False) print(bs.ep.weight.a) #error
The Request is then in particular to be able to sample weighted graphs from a state, but in general to preserve the graph properties when minimizing a model in order to sample a graph with those properties from the resulting state.