At least one column in the data corresponds to the name of a metric that will be computed by scoringutils. Please check `available_metrics()` - does not flag which variable

[seabbs]

As in the title this warning from check_forecasts does not flag which variable which is not a great user experience.

[Bisaloo]

Does that look okay?

library(scoringutils)
#> Note: scoringutils 1.0.0 introduces a lot of breaking changes and we apologise for any inconvenience. If you prefer the old interface, please download version 0.1.8 using remotes::install_github("epiforecasts/scoringutils@v0.1.8")

check <- example_quantile |> 
  dplyr::mutate(bias = runif(1)) |> 
  na.omit() |> 
  check_forecasts()
#> Warning in check_forecasts(na.omit(dplyr::mutate(example_quantile, bias = runif(1)))): The following warnings were produced when checking inputs:
#> 1.  At least one column in the data (bias) corresponds to the name of a metric that will be computed by scoringutils. Please check `available_metrics()`

^{Created on 2022-06-20 by the reprex package (v2.0.1.9000)}

If not, could you please post an example of the error message you'd like to see? I'll implement it.

[nikosbosse]

I think this loos great! I would suggest to put it in quotation marks, i.e. ("bias"). But really happy either way.