This is partly a small QoL improvement, but having a way of regenerating this data programatically is also a prerequisite for making it possible to run regression tests without having to store the results in the git repo (see #351).
Are we happy that this makes sense regardless of what solution we come up with for #351? @dalonsoa @tsmbland