Get distribution sample with similar trend

what if I don’t need trend prediction, but just to arrange my sample’s values in the right order, so it would be proportionally similar as in the initial data? is there such a function in pymc3?