Our copulas will give us U(0, 1) marginal random variates. These can be plugged into the inverse empirical cumulative distribution function for random variates from our desired marginal distribution.
Actually, the trimmed_hd_quantile_estimator module is not working well. Just use method="inverted_cdf" in the numpy.quantile function for transforming U(0, 1) random variates from copulas.