I didn't have to do it in two `replace` calls
Now I have to check if the rule has a or several corresponding
categories in the initial and final regex, find whether they have the
same amount of elements, and if so create rules that allow elements from
each category to be mapped 1:1