Naming inconsistency for evaluation measure

[HeidiSeibold]

It is "predictive_accuracy" in

evals <- listOMLRunEvaluations(task.id = tsk.id, evaluation.measure = "predictive_accuracy")

but "predictive.accuracy" in

names(evals)

is that intentional?

[giuseppec]

Hm. I see this is not user friendly. Reason is because R guys have their own naming convention (words in column names and function arguments are separated with dots).
Maybe allowing both is better better:

evals <- listOMLRunEvaluations(task.id = tsk.id, evaluation.measure = "predictive.accuracy")
evals <- listOMLRunEvaluations(task.id = tsk.id, evaluation.measure = "predictive_accuracy")

???

[berndbischl]

i think it was a bad decision to rename stuff, from the OML naming scheme, to something else.
but that is not easy to change now.

I REALLY dont want to complicate things further by doing this :

Maybe allowing both is better better:

document this properly, and MAYBE later we remove the renaming completely.
but this will break some existing stuff.

[berndbischl]

@giuseppec

can you please create a clean issue, and document which OML things we rename where?
then we can make an informed decision.
also stuff becomes more transpararent for all users.