Currently many config parameters for various different dataset classes are grouped together in the toml file table data_set.
We should go ahead and create subtables and percolate the configurations down correctly (probably at the same time that do the major "data_set"|"dataSet" -> "dataset" rename). i.e.
[data_set]
[data_set.hsc_dataset]
param_1 = 0
param_2 = "thing"
[data_set.hyrax_random_dataset]
param_1 = "stuff"
param_2 = 42
etc...