perf(topology/sheaves/presheaf_of_functions): speedups (#16873)
Speed up`CommRing_yoneda` (which caused [a bors failure](https://github.com/leanprover-community/mathlib/actions/runs/3211416019/jobs/5249570763)) from >20s to <2.5s by filling in automatic fields. (merged in another PR)
Speed up `presheaf_to_Types` from 13.3s to 0.3s and `presheaf_to_Type` from 4s to 0.1s by filling in automatic fields.