Rework score() #344
Rework score() #344
Conversation
…d validates all inputs
…heck_data_columns that check input is data.table with right columns
…e more helpers for input checking
…e reason, but plots are otherwise still the same.
…ostly left untouched.
…, fix linting issues, update documentation, remove unnecessary comments
…to being a named list. Add documentation.
In an ideal world these changes would be there own PR as it adds a lot of noise to this one. |
seabbs
left a comment
seabbs
left a comment
There was a problem hiding this comment.
Because of how much is loaded intoo this PR its a real nightmare to review. I would suggest addressing the things mentioned in earlier PRs in those PRs and saving moving files around etc for their own PRs.
I will circle back and address more fully. Current review comments flag issues linting and issues with the setup of score and the naming of validate
| score <- function(data, | ||
| metrics = NULL, | ||
| ...) { | ||
| score <- function(data, ...) { |
There was a problem hiding this comment.
why is metrics not an arg for score?
There was a problem hiding this comment.
I wasn't quite sure what the default should be. No default (but then again the default is in the methods)? NULL? I was hoping it would be more clear that way when looking at the documentation for the methods.
|
Merging after reviewing and discussing sesssions |
3 participants
This PR
check_forecasts()by a new function,validate()which assigns the class on an input without a class and also serves as the validator function for thescoringutils_binary,scoringutils_point,scoringutils_quantileandscoringutils_sampleclassesscore(), but as of now leavesscoringutils_quantile()untouched. This one is scary. And maybe can be it's own PR in the spirit of my (already failed) attempt to partition things into smaller PRs.check_,assert_andtest_)