First draft of subgroup normalization

[bethac07]

Might be made unnecessary by cytomining/pycytominer#154, but that's TBD. @ErinWeisbart, can you test when you get to the profiles stage of the cell painting project you just started? LMK if the implementation isn't self-evident in what you'd do to change the metadata and/or the actual config file you run.

[ErinWeisbart]

Worked for me.

Minor suggestion:

• I initially set the subgroup column name to what was in my metadata.csv but it required Metadata_ prepended to it (as happens during annotation). Might be handy to add in a small check like if not subgroup_col.startswith("Metadata_"): subgroup_col = "Metadata_"+subgroup_col

Bigger request:

• I think the subgroup output name should be different to indicate that it underwent subgroup normalization instead of whole plate normalization (eg. _normalized_subgroup similar to how we output _normalized and _normalized_negcon)

[ErinWeisbart]

I went ahead and wrote the "Bigger Request" because I wanted it for the batch I was processing.
Now, it will always have default behavior but if you set subgroup: true in normalize it will also output subgroup_normalized files. Additionally, if you set subgroup: true for feature_select it will output the default as well as subgroup_normalized_feature_selected

I have tested it on the one batch I was processing but have not otherwise done extensive testing.

[ErinWeisbart]

We discussed today that we want to abstract the subgroup further and create a separate option normalize_subgroup similar to how normalize_negcon is separate from normalize.