[graph-tool] Problem on bipartite-networks with label

I am using [SBMtopicmodeling](https://github.com/martingerlach/hSBM_Topicmodel) to inference partition on bipartite networks After a recent graph-tool update I noticed there is a problem with **state_args** in minimize_nested_blockmodel_dl step to reproduce: given a trivial bi-partite network, with VertexProperty **kind** that imposes the bi-partition clabel = g.vp['kind'] state_args = {} state = gt.minimize_nested_blockmodel_dl(g, state_args=state_args) finds partitions but when I impose bi-partition clabel = model.g.vp['kind'] state_args = {'clabel': clabel, 'pclabel': clabel} state = gt.minimize_nested_blockmodel_dl(g, state_args=state_args) it fails to find any partition. What could be the problem?