How should we represent/record standard dataset?

Created by: asogaard

In the interest of reproducibility, benchmarking, and such it could be useful to have simple "cards" that unambiguously define various benchmark datasets. This could be in the form of a simple structure like:

datasets/
|-- upgrade-sensitivity-mc.yml
|-- deepcore-doublebang-mc.yml
|-- deepcore-oscillations-2021-data.yml
`-- (...)

with each YAML file containing little more than path(s) to I3 files on Cobalt.

What do y'all think?