Translates a fitted Bayesian structural equation model from the
blavaan package into a UMG. The graph topology is identical to
that of the frequentist umg_from_lavaan() translation; what
differs is the interpretive reading, since every free parameter
carries a prior. The construction makes literal the article's claim
that the Bayesian and frequentist forms of a model are the same
diagram differing only in which vertices have been prior-closed.