[graph-tool] Re: Inference for weighted, hierarchical stochastic block models