Open nikosbosse opened 1 month ago
I think the issue that needs addressing here is mainly around the validation of forecast objects and verbosity of that validation?
One solution which I think @Bisaloo is getting at is have fewer user facing functions -> fewer functions that need to validate -> fewer suppressed messages etc.
The other thing that comes to mind is the current desire for score
to warn but not fail when an underlying metric fails (as this could be very annoying to users). Potentially we are just happy with it being annoying to users and so should drop? The downside is it may be somewhat unclear to users what has failed
What other issues are there? I think this could do with some detailed feedback?
Originally posted by @Bisaloo in https://github.com/epiforecasts/scoringutils/pull/791#pullrequestreview-2003401135