Create snippet or extend addin to allow specification of LANGUAGE for the reprex

[andresrcs]

My native language is spanish, so the locale on my PC is ES_us, but when posting a reprex to English forums that includes error messages I would like to use EN_us, that way I wouldn't have to manually translate the error messages, or remeber to manually change my locale beforehand, would you consider adding this funtionality? I think it would be pretty handy for non-English users.

[batpigandme]

Temporary workaround – you can use the withr package with_envvar() to set the language.
So, for example, if I run the following code as a reprex as I usually do, the message is in English.

dplyr::select(foo)
#> Error in dplyr::select(foo): object 'foo' not found

If I run:

withr::with_envvar(c("LANG" = "de_DE.UTF-8"), code = {reprex::reprex(dplyr::select(foo))})

I get

dplyr::select(foo)
#> Error in dplyr::select(foo): Objekt 'foo' nicht gefunden

^{Created on 2019-04-16 by the reprex package (v0.2.1)}

[andresrcs]

Thanks Mara!, this is a nice walkaround solution although I still think it would be more convenient to have this functionality on the reprex() function itself.

[jennybc]

Assuming that one attaches reprex in .Rprofile (I think heavy users do) and that target code is on the clipboard, @batpigandme's suggestion gets very slick:

withr::with_envvar(
  c("LANG" = "de_DE.UTF-8"),
  reprex()
)

I can't see a built-in solution being that much easier/shorter and it would create more code to document and maintain. This looks like a good thing to create an RStudio snippet for? Or maybe even incorporate this as a feature in the current gadget/addin? I think all of those are better ways to make this withr trick more available than to add another argument to reprex() itself.

Do you @andresrcs have any interest in noodling on those alternatives?

[jennybc]

Re: tidyverse dev day

An interesting project would be to explore exposing this method of setting the language in the existing addin/gadget.

reprex already imports withr. A minimum viable product, therefore, it's adding some sort of input for language and reworking the reprex() call if the user wants something other than their default. I don't have much experience with doing this in general, so I would anticipate some thought around ... Does the user have to provide this string or is there some sort of dropdown menu? Can we determine all of the valid choices and present them?

[andresrcs]

Do you @andresrcs have any interest in noodling on those alternatives?

I whish I had the skills needed, I just ended up creating my own reprex_en() variant in .Rprofile, this is more convenient for me since I mostly work on RStudio Server, so no clipboard available.

reprex_en <- function(...) {
  withr::with_envvar(
    c("LANG" = "en_US.UTF-8"),
    reprex::reprex(...)
  )
}

[jennybc]

I think that's a good solution! Esp. for an individual.

I will leave this open in case anyone want to tackle at our next tidyverse dev day.

[andresrcs]

Maybe something like this would be more generalizable

reprex_locale <- function(..., locale = "en_US.UTF-8") {
  withr::with_envvar(
    c("LANG" = locale),
    reprex::reprex(...)
  )
}

[jennybc]

This is related to locale in the Sys.getlocale() sense. But, from those docs, note that:

Note that the LANGUAGE environment variable has precedence over "LC_MESSAGES" in selecting the language for message translation on most R platforms.

And this thread appears to really be about this, more than true aspects of locale. Not sure what to make of this difference: "LANGUAGE" (from the docs) vs. "LANG" (from all the working examples here).

[hadley]

Details at https://cran.rstudio.com/doc/manuals/r-devel/R-admin.html#Localization-of-messages — I think setting LANGUAGE is most appropriate here

[andresrcs]

This make me think about this other use case, being English the "lingua franca" I also want my date labels in English when rendering a reprex.

reprex_locale <- function(..., locale = "en_US.UTF-8") {
    withr::with_envvar(
        c("LC_ALL" = locale),
        reprex::reprex(...)
    )
}

library(ggplot2)
df <- data.frame(stringsAsFactors = FALSE,
                 date = as.Date(c("2019-01-01", "2019-02-01")),
                 value = c(1, 2))
ggplot(df, aes(x = date, y = value)) +
    geom_col() +
    scale_x_date(date_breaks = "1 month", date_labels = "%b")