Constraints for multivariate prior in Bayesian update scheme

It’s not, we need to wrap a Scipy Op that computes the CDF for the truncation probability