Add check function

[nikosbosse]

Adds a function to check the input for eval_forecasts().
Returns a list with different checks

[nikosbosse]

sorry, mixed this up with a few other tweaks. Relevant file is check_forecasts.R

[codecov]

Codecov Report

Merging #154 (1009e7b) into master (29ac025) will increase coverage by 6.12%.
The diff coverage is 60.29%.

@@            Coverage Diff             @@
##           master     #154      +/-   ##
==========================================
+ Coverage   48.44%   54.56%   +6.12%     
==========================================
  Files          18       19       +1     
  Lines        1348     1468     +120     
==========================================
+ Hits          653      801     +148     
+ Misses        695      667      -28     

Impacted Files	Coverage Δ
R/eval_forecasts.R	67.27% <ø> (+1.20%)	⬆️
R/pairwise-comparisons.R	42.36% <ø> (ø)
R/plot.R	20.29% <ø> (+20.29%)	⬆️
R/utils.R	71.42% <ø> (+7.93%)	⬆️
R/utils_data_handling.R	79.24% <ø> (+6.60%)	⬆️
R/check_forecasts.R	57.14% <57.14%> (ø)
R/eval_forecasts_quantile.R	93.75% <100.00%> (-0.31%)	⬇️
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 29ac025...1009e7b. Read the comment docs.

[nikosbosse]

tests still missing at the moment (although examples are run).

[Bisaloo]

There are some inconsistency in using = vs <- so it might be worth running styler.

[seabbs]

This looks good Nikos. Mostly just need to resolve @Bisaloo points.

One comment: If the end goal is to have eval_forecasts be a S3 method itself then at some point the input needs to be assigned a scoringutils class to make that happen. One way you could do this is to assign a class here for each data type as well as your check type and then return the data and checks in a list.

This as a workflow would then be something like this:

data |> 
   check_forecasts() |>
   eval_forecasts() 

or 

checked_data <- check_forecasts(data)
print(checked_data)
eval_forecasts(checked_data)

[nikosbosse]

My idea re S3 classes would be this:

• first we add the check function without changing the workflow
• then we do a workflow change (if we agree we want that) together with all the other breaking changes that will come soon.

At the moment you can run check_forecasts() independently and I think for now that's good and definitely better than not having it. I was a bit on the fence whether it should also return the data with all the NA values removed (that's essentially the only thing it does to alter the data).

[nikosbosse]

Thank you very much for your help and review @Bisaloo, @seabbs 🎉 !