FR: Pass fct_relabel fun parameter to rlang::as_function

[alistaire47]

It would be handy to pass the fun parameter of fct_relabel to rlang::as_function so it can accept formula (~tolower(.x)) or quosure (quo(tolower(.x)) style functions or string function names ("tolower") instead of just raw functions.

A trivial reprex (though realistically regex is the obvious use-case):

library(tidyverse)
library(forcats)

# Right now this is the only option for an anonymous function:
iris %>% 
    tbl_df() %>% 
    mutate(Species = fct_relabel(Species, 
                                 function(x){substr(x, 1, 3)}))
#> # A tibble: 150 x 5
#>    Sepal.Length Sepal.Width Petal.Length Petal.Width Species
#>           <dbl>       <dbl>        <dbl>       <dbl>  <fctr>
#>  1          5.1         3.5          1.4         0.2     set
#>  2          4.9         3.0          1.4         0.2     set
#>  3          4.7         3.2          1.3         0.2     set
#>  4          4.6         3.1          1.5         0.2     set
#>  5          5.0         3.6          1.4         0.2     set
#>  6          5.4         3.9          1.7         0.4     set
#>  7          4.6         3.4          1.4         0.3     set
#>  8          5.0         3.4          1.5         0.2     set
#>  9          4.4         2.9          1.4         0.2     set
#> 10          4.9         3.1          1.5         0.1     set
#> # ... with 140 more rows

# ...but this syntax would be convenient and consistent with the rest of the tidyverse:
iris %>% 
    tbl_df() %>% 
    mutate(Species = fct_relabel(Species, 
                                 ~substr(.x, 1, 3)))
#> Error in mutate_impl(.data, dots): Evaluation error: Expected function, got formula.

# ...and it would work:
iris %>% 
    tbl_df() %>% 
    mutate(Species = fct_relabel(Species, 
                                 rlang::as_function(~substr(.x, 1, 3))))
#> # A tibble: 150 x 5
#>    Sepal.Length Sepal.Width Petal.Length Petal.Width Species
#>           <dbl>       <dbl>        <dbl>       <dbl>  <fctr>
#>  1          5.1         3.5          1.4         0.2     set
#>  2          4.9         3.0          1.4         0.2     set
#>  3          4.7         3.2          1.3         0.2     set
#>  4          4.6         3.1          1.5         0.2     set
#>  5          5.0         3.6          1.4         0.2     set
#>  6          5.4         3.9          1.7         0.4     set
#>  7          4.6         3.4          1.4         0.3     set
#>  8          5.0         3.4          1.5         0.2     set
#>  9          4.4         2.9          1.4         0.2     set
#> 10          4.9         3.1          1.5         0.1     set
#> # ... with 140 more rows

The tradeoff for the convenience is that it requires adding rlang as a dependency. If that's ok, I'd be happy to assemble a PR if you like.

[hadley]

A PR would be awesome!

(And we need to add rlang for #110, so it's not a big deal)