From what I understood, we check whether "snake case category" from step (1) is already known to us (in the cache) so that we need no further processing. So tha

From what I understood, we check whether "snake case category" from step (1) is already known to us (in the cache) so that we need no further processing. So that step (2) and further don't apply for categories that were already produced earlier.