Handling case in WIS where quantile levels don't form valid intervals

[nikosbosse]

Description

This PR closes #926.

As discussed in #926, this PR lets wis() error by default if not all quantile levels form valid prediction intervals. I realised that there was a na.rm argument already (it was just set to TRUE). I think FALSE is a better default.
If the quantile levels aren't symmetric, the function errors immediately.

(Note that there is a second way in which na.rm can be relevant: if a specific quantile is missing for a forecast. With na.rm = TRUE NA values will be dropped and otherwise the output is NA.)

Checklist

• My PR is based on a package issue and I have explicitly linked it.
• I have included the target issue or issues in the PR title as follows: issue-number: PR title
• I have tested my changes locally.
• I have added or updated unit tests where necessary.
• I have updated the documentation if required.
• I have built the package locally and run rebuilt docs using roxygen2.
• My code follows the established coding standards and I have run lintr::lint_package() to check for style issues introduced by my changes.
• I have added a news item linked to this PR.
• I have reviewed CI checks for this PR and addressed them as far as I am able.

[seabbs]

Kätzchen spielt im Hof
Miau, miau, ruft es leise
Mutters Strickzeug tanzt

[seabbs]

Note that there is a second way in which na.rm can be relevant: if a specific quantile is missing for a forecast. With na.rm = TRUE NA values will be dropped and otherwise the output is NA.)

This was the motivation for the previous setting wasn't it because historically hubs allowed a mixed set of quantiles to be submitted and we wanted to support that?

[seabbs]

Alte Katze schläft
Erinnerung an Spiele
Strickzeug längst verblasst

[nikosbosse]

Kätzchen spielt im Hof
Miau, miau, ruft es leise
Mutters Strickzeug tanzt

I couldn't agree more.

This was the motivation for the previous setting wasn't it because historically hubs allowed a mixed set of quantiles to be submitted and we wanted to support that?

This is really more missing forecast values. If you had mixed sets of quantiles, then you'd get a warning in score(), but otherwise it should be fine.

@elray1 @nickreich are there any issues you'd expect from this PR?
TLDR: wis() errors if you supply quantile_levels like c(0.1, 0.4, 0.7) that don't form intervals. And if you have quantile_levels = c(0.1, 0.2, 0.8, 0.9) but predicted = c(2, 4, 8, NA) then it will result in NA by default.

[seabbs]

Nice looks good. A few linting issues flagged and maybe more verbosity in the news. Also, I have found a great use of AI. See if you can guess what the prompt was.

[seabbs]

This is really more missing forecast values. If you had mixed sets of quantiles, then you'd get a warning in score(), but otherwise it should be fine.

Okay makes sense.

[nikosbosse]

Also, I have found a great use of AI. See if you can guess what the prompt was.

1950's German folk poetry? :D

[seabbs]

haiku from a 19th century german 9 year old and a haiku from the same 9 year old when they are 95

[seabbs]

Nice. LGTM