Generalize score generation and optimization

this should not need the whole batch container, it should work on subset:
https://github.com/bedapub/designit/blob/fb873c1c7c85f186d3fcae26156f0224d114250f/R/score_plates.R#L80

list of lists, should work with group_by:
https://github.com/bedapub/designit/blob/fb873c1c7c85f186d3fcae26156f0224d114250f/R/score_plates.R#L253

TODO:
- [ ] decide how to proceed
1. to act on a sample subset `bc |> group_by() |> optimize(, scoring_fun = mk_scoring(batch_var))`
2. or do it like `bc |> group_by() |> mk_scoring("osat") |> optimize()`
- [ ] batch container NEEDS grouping (write a method to group_by the samplesheet and add a key to the bc)
- [ ] we NEED a new function to work on a subset of samplesheet -> `grid_2d_score_generator()`
- [ ] how to name it?
- [ ] ~scoring function might collide with the shuffling~ shuffling needs to act on the subset as well
- [ ] what would this break? - https://github.com/bedapub/designit/blob/fb873c1c7c85f186d3fcae26156f0224d114250f/R/osat.R#L142
- [ ] remove the "group_by" variable from the batch_vars?
- [ ] shuffle function should also respect this

<1.> `bc |> optimize(scoring = "osat") |> group_by("plate") |> optimize(scoring = "2d")`
<2.> `bc |> add_scoring_fun("osat", "plate") |> optimize() |> group_by("plate") |> add_scoring_fun("2d") |> optimize()`

this should be generalizable to any grouping by batch, etc.

even more general:
`bc |> add_scoring_fun("osat", "plate") |> add_shuffling_fun("row_swap") |> optimize()`