Change deprecated ggplot2 syntax to current recommendations

[bwiernik]

Summary:

ggplot2 recently added a linewidth argument to replace size for line geoms and elements. Currently, loading rstan or a package that loads it produces a deprecation warning about the size argument. This PR replaces size with linewidth for line geoms and elements in the rstan plotting functions. The PR also updates a few other instances of deprecated ggplot2 syntax with the recommended alternatives (replacing aes_string() with aes(.data$...), replacing "..density.." with after_stat(density)).

Intended Effect:

No visual changes to plot output. Removal of deprecation warnings on package load.

How to Verify:

Output is unchanged for the following plots with the current develop version and this PR:

library(brms)
library(rstan)
m <- brm(mpg ~ disp, data = mtcars)
stan_dens(m$fit)
stan_hist(m$fit)
stan_plot(m$fit)

Documentation:

https://github.com/tidyverse/ggplot2/blob/main/NEWS.md

Copyright and Licensing

Please list the copyright holder for the work you are submitting (this will be you or your assignee, such as a university or company):

Brenton M. Wiernik

By submitting this pull request, the copyright holder is agreeing to license the submitted work under the following licenses:

• Code: GPLv3 (http://opensource.org/licenses/GPL-3.0)
• Documentation: CC-BY 4.0 (https://creativecommons.org/licenses/by/4.0/)

[hsbadr]

I think there's no need for .data$ with aes(). This applies everywhere.

Suggested change

	ggplot2::aes(x = .data$iteration, y = .data$value, color = .data$chain)
	ggplot2::aes(x = iteration, y = value, color = chain)

[bwiernik]

No, the .data$ is necessary to avoid global variable name NOTEs in CRAN checks

[hsbadr]

@bwiernik Can you address the comment above and rebase this on the experimental branch?

[hsbadr]

Suggested change

	mapping = ggplot2::ggplot2::aes(x = .data$x, y = .data$y),
	mapping = ggplot2::aes(x = .data$x, y = .data$y),

[bgoodri]

I am guessing this is OK, although I batched a few suggestions.

[bwiernik]

I removed another aes_q() that had been added on the experimental branch. The .data pronoun inside aes() is necessary because aes(x = value) will trigger a Undeclared Global Variable Name NOTE in R CMD CHECK. data$value or data[['value']] is the current recommended syntax for situations like this.

From ?aes_string

All these functions are soft-deprecated. Please use tidy evaluation idioms instead. Regarding aes_string(), you can replace it with .data pronoun. For example, the following code can achieve the same mapping as aes_string(x_var, y_var).

x_var <- "foo"
y_var <- "bar"
aes(.data[[x_var]], .data[[y_var]])

See vignette("ggplot2-in-packages").

[jgabry]

Thanks @bwiernik!

[bgoodri]

Is this good to merge now?

[jgabry]

Were the linewidth changes merged already at some point? I don’t see them in the Files Changed view, but they are in the commit history. The Files Changed view now just shows aes and .data stuff, which looks good to go.

[bwiernik]

It looks like all of the commits but the most recent are showing as already merged.

[hsbadr]

Yes, I tested this in the experimental.