checkhelper

A package to help you deal with devtools::check() outputs and helps avoids problems with CRAN submissions

Complete documentation in the {pkgdown} site: https://thinkr-open.github.io/checkhelper/

Installation

You can install the last version of checkhelper from r-universe with:

install.packages('checkhelper', repos = 'https://thinkr-open.r-universe.dev')

Or from GitHub:

remotes::install_github("thinkr-open/checkhelper")

Examples

  • Check your current package under development and get all the globals missing: no visible global variable and no visible global function
  • Detect exported functions with missing or empty @return / @noRd tags

Directly in your package in development

  • Use checkhelper::find_missing_tags() on your package in development to find which functions are exported but missing @export roxygen2 tag.
    • CRAN policy asks for every exported function to have a value (named @export when using {roxygen2}).
    • This also checks that not exported functions dont have roxygen title, or have @noRd in case you faced Please add \value to .Rd files CRAN message for documented but not exported functions.
  • You can directly use checkhelper::print_globals() on your package instead of devtools::check(). This is a wrapper around rcmdcheck::rcmdcheck(). This will run the checks and directly list the potential “globalVariables” to add in a globals.R file.
checkhelper::find_missing_tags()

checkhelper::print_globals(quiet = TRUE)

Reproducible example with a fake package in tempdir

  • Create a fake package with
    • a function having global variables
    • a function with @export but no @return
    • a function with title but without @export and thus missing @noRd 
library(checkhelper)

# Create fake package ----
pkg_path <- tempfile(pattern = "pkg.")
dir.create(pkg_path)

# Create fake package
usethis::create_package(pkg_path, open = FALSE)
#> ✔ Setting active project to '/tmp/RtmpT2YnX3/pkg.26be9459ed409'
#> ✔ Creating 'R/'
#> ✔ Writing 'DESCRIPTION'
#> ✔ Writing 'NAMESPACE'
#> ✔ Setting active project to '<no active project>'

# Create function no visible global variables and missing documented functions
cat("
#' Function
#' @importFrom dplyr filter
#' @export
my_fun <- function() {
data %>%
filter(col == 3) %>%
mutate(new_col = 1) %>%
ggplot() +
  aes(x, y, colour = new_col) +
  geom_point()
}

#' Function not exported but with doc
my_not_exported_doc <- function() {
  message('Not exported but with title, should have @noRd')
}
", file = file.path(pkg_path, "R", "function.R"))

attachment::att_amend_desc(path = pkg_path)
#> Updating pkg.26be9459ed409 documentation
#> ℹ Loading pkg.26be9459ed409Writing ']8;;file:///tmp/RtmpT2YnX3/pkg.26be9459ed409/NAMESPACENAMESPACE]8;;'Writing ']8;;file:///tmp/RtmpT2YnX3/pkg.26be9459ed409/NAMESPACENAMESPACE]8;;'Writing ']8;;ide:run:pkgload::dev_help('my_fun')my_fun.Rd]8;;'Writing ']8;;ide:run:pkgload::dev_help('my_not_exported_doc')my_not_exported_doc.Rd]8;;'ℹ Loading pkg.26be9459ed409[+] 1 package(s) added: dplyr.

# Files of the package
fs::dir_tree(pkg_path, recurse = TRUE)
  • Find missing @return and find missing @noRd for not exported function with documentation
find_missing_tags(pkg_path)
#> ℹ Loading pkg.26be9459ed409
#> Missing or empty return value for exported functions: my_fun
#> 
#> 
#> 
#> Doc available but need to choose between `@export` or `@noRd`: my_not_exported_doc
#> 
#> 
#> 
#> ℹ Loading pkg.26be9459ed409
#> # A tibble: 2 × 11
#>      id filename   topic has_e…¹ has_r…² retur…³ has_n…⁴ rdnam…⁵ not_e…⁶ test_…⁷
#>   <int> <chr>      <chr> <lgl>   <lgl>   <chr>   <lgl>   <chr>   <lgl>   <chr>  
#> 1     1 function.R my_f… TRUE    FALSE   ""      FALSE   my_fun  FALSE   not_ok 
#> 2     2 function.R my_n… FALSE   FALSE   ""      FALSE   my_not… FALSE   ok     
#> # … with 1 more variable: test_has_export_or_has_nord <chr>, and abbreviated
#> #   variable names ¹​has_export, ²​has_return, ³​return_value, ⁴​has_nord,
#> #   ⁵​rdname_value, ⁶​not_empty_return_value, ⁷​test_has_export_and_return
  • Get global variables
globals <- get_no_visible(pkg_path, quiet = TRUE)
globals
#> $globalVariables
#> # A tibble: 4 × 6
#>   notes                                    fun   is_fu…¹ is_gl…² varia…³ propo…⁴
#>   <chr>                                    <chr> <lgl>   <lgl>   <chr>   <chr>  
#> 1 my_fun: no visible binding for global v… my_f… FALSE   TRUE    data    " impo…
#> 2 my_fun: no visible binding for global v… my_f… FALSE   TRUE    x        <NA>  
#> 3 my_fun: no visible binding for global v… my_f… FALSE   TRUE    y        <NA>  
#> 4 my_fun: no visible binding for global v… my_f… FALSE   TRUE    new_col  <NA>  
#> # … with abbreviated variable names ¹​is_function, ²​is_global_variable,
#> #   ³​variable, ⁴​proposed
#> 
#> $functions
#> # A tibble: 5 × 6
#>   notes                                    fun   is_fu…¹ is_gl…² varia…³ propo…⁴
#>   <chr>                                    <chr> <lgl>   <lgl>   <chr>   <chr>  
#> 1 my_fun: no visible global function defi… my_f… TRUE    FALSE   %>%     <NA>   
#> 2 my_fun: no visible global function defi… my_f… TRUE    FALSE   mutate  <NA>   
#> 3 my_fun: no visible global function defi… my_f… TRUE    FALSE   ggplot  <NA>   
#> 4 my_fun: no visible global function defi… my_f… TRUE    FALSE   aes     <NA>   
#> 5 my_fun: no visible global function defi… my_f… TRUE    FALSE   geom_p… <NA>   
#> # … with abbreviated variable names ¹​is_function, ²​is_global_variable,
#> #   ³​variable, ⁴​proposed
  • Print globals to copy-paste
print_globals(globals)
#> --- Functions to add in NAMESPACE (with @importFrom ?) ---
#> 
#> my_fun: %>%, aes, geom_point, ggplot, mutate
#> 
#> --- Potential GlobalVariables ---
#> -- code to copy to your globals.R file --
#> 
#> globalVariables(unique(c(
#> # my_fun: 
#> "data", "new_col", "x", "y"
#> )))
  • Store the output of print_globals() in package using usethis::use_r("globals"). Note that you can also transform all these variables with .data[[variable]]

Experimental: Check as CRAN with CRAN global variables

Use the exploration of CRAN scripts by the RConsortium to check a package as CRAN does it with their env. variables. See https://github.com/RConsortium/r-repositories-wg/issues/17 for more details.

check_dir <- tempfile("example")
# Check the current directory
check_as_cran(check_dir = check_dir)
# Open directory with all outputs
utils::browseURL(check_dir)

Please note that this project is released with a Contributor Code of Conduct. By participating in this project you agree to abide by its terms.