Adaptions to resolve all current Errors and Warnings in R CMD check

[mnlang]

Hi Alex,

I came across your interesting package as I was looking for a lightweight implementation of generic functions for various distributions for one of my projects. Since the CRAN version seems to be missing some important updates compared to the Github version according to issue #64, I tried to fix the current warnings and errors w/i R CMD check.

In summary, this PR makes the following adaptions:

1. A few fixes in the Rd files (update Rd files using roxygen2, remove \donttest environment, export missing S3 method kurtosis).
2. Change first argument w/i S3 method mean from d to x.
3. Update geom_*s and stat_*s for coloring AUC in a PDF plot.

Remark that for me devtools::check() still fails because ggplot2 is not in Imports but only in Suggests. But since R CMD check works, this seems to be a problem with devtools.

Further comments:

• Add 2: This is a necessary fix, but not an ideal one, since now the S3 methods mean() and e.g. variance() have different arguments x and d. This could be fixed by changing all first arguments in the methods to x, which would also solve the issue #58 of masking stats::quantile().

• Add 3: I tried to build on the original implementation. As some of these functions still could be improved, it might be an idea to build an interface to ggdist instead of providing own geom_*s.

If desired, I can of course revise parts accordingly. Thank you!

[alexpghayes]

Thanks for this! Looks good to me!

[mnlang]

Thanks for your fast reply and for merging my PR. Would it be in your interest if I continue working on the second point?

I have looked a bit into the package, and it is really a very nice and clean project. Are there any plans from you to increase your work on the package again, or is it still as discussed in PR #57 with @paulnorthrop that you unfortunately don't have much time for it at the moment?

[alexpghayes]

Adds 2 and 3 both seem great!

I personally don't use distributions3 for any coding projects, and so I can't put a ton of time towards it. However I'd be more than happy (and will have some time) to try and get a new release out this winter, and would be happy to give collaborators write access to the repo if they'd like to do more active development.

[mnlang]

Okay, sounds good to me. I'm not using distributions3 for a project yet, but might do so in the future. I'll try to get some PRs done from time to time, and then we'll see where it goes! An updated CRAN version, with maybe even some extensions, would be great!