Who was the first to consider that categories were semi-simplicial sets (and in particular groupoids were simplicial sets)?

I think there was a concept of nerve of a covering in algebraic topology before (maybe Alexandroff).