tidycheckUsage

When using magrittr pipes or ggplot2 reading the objects that you forgot to enquo is a pain. These functions can be run prior to devtools::check to return a data.frame containing all the warnings created when check runs codetools::checkUsage. Using this information actionable functions append rlang syntax to fix the warnings.

1. Installation
2. Examples
3. Reports
4. Tidyverse Warning Resolution

Installation

remotes::install_github('yonicd/tidycheckUsage')

Examples

Installed Packages

library(tidycheckUsage)

install.packages(c('dplyr'))

library(dplyr,warn.conflicts = FALSE,quietly = TRUE)

knitr::kable(tidycheckUsagePackage('dplyr'))

file	line	object	col1	col2	path	fun	warning
		dupe				as.tbl_cube.data.frame	local variable ‘dupe’ assigned but may not be used
		index				do.rowwise_df	local variable ‘index’ assigned but may not be used
		cols				fmt_cols	local variable ‘cols’ assigned but may not be used
		measures				fmt_measures	local variable ‘measures’ assigned but may not be used
		args				fmt_pos_args	local variable ‘args’ assigned but may not be used
		type				order_by	local variable ‘type’ assigned but may not be used
		type				select_var	local variable ‘type’ assigned but may not be used
		first_type				select_vars	local variable ‘first_type’ assigned but may not be used
		actual_type				switch_rename	local variable ‘actual_type’ assigned but may not be used

Built Packages

system('git clone https://github.com/hrbrmstr/slackr.git')
devtools::install('slackr')

tidycheckUsagePackage('slackr')

file	line	object	col1	col2	path	fun	warning
slackr_bot.r	57	resp_ret	3	10	/Users/jonathans/projects/forks/slackr/R	slackr_bot	local variable ‘resp_ret’ assigned but may not be used
slackr_bot.r	57	resp_ret	3	10	/Users/jonathans/projects/forks/slackr/R	slackrBot	local variable ‘resp_ret’ assigned but may not be used
slackr.R	41	resp_ret	3	10	/Users/jonathans/projects/forks/slackr/R	slackr	local variable ‘resp_ret’ assigned but may not be used

Functions in Environment

myfun <- function(x){
  
  ret <- mtcars%>%
    mutate(mpg2=mpg*x)
  
  ret <- ret%>%
    dplyr::mutate(mpg3=mpg2^2)
  
}

tidycheckUsage(fun = myfun)

file	line	object	col1	col2	path	fun	warning
	3	%>%	16	18		myfun	no visible global function definition for ‘%>%’
	4	mpg	17	19		myfun	no visible binding for global variable ‘mpg’
	4	mutate	5	10		myfun	no visible global function definition for ‘mutate’
	6	%>%	13	15		myfun	no visible global function definition for ‘%>%’
	7	mpg2	24	27		myfun	no visible binding for global variable ‘mpg2’

HTML Report outputs

Forking covr::report a similar output is created for symbol usage in R scripts.

tidycheckUsage::usage_report(tidycheckUsagePackage('sinew'))

Summary statistics for symbol usage

Number of relevant symbols are counted then categorized by usage warnings:

• Valid symbols (no warnings or notes)
• Problem symbols that generate warnings in codetools
  • General: Neither Missing Global or Unused Local
  • Missing Global: symbols that generate a warning: 'no visible binding for global variable'
  • Unused Local: symbols that generate a warning: 'local variable is assigned but may not be used'

Within Script Investigations

Red indicates no visible binding for global variable

Orange indicates local variable is assigned but may not be used

Appending rlang

Now we can use this information to place rlang syntax instead of calling the object itself within a tidyverse call.

To help make this painless the output from tidycheckUsage is used to find and replace all those objects. This can be done with the tidy object and append_rlang, for either small functions or whole packages.

x <- function(){
  
  data <- tidyr::unite(mtcars, col = vs_am, c(vs,am))
  
  ggplot2::ggplot(data = data, ggplot2::aes(x=mpg,y=qsec,colour=vs_am)) + 
  ggplot2::geom_point()
}

(obj <- tidycheckUsage(x))
  file line object col1 col2 path fun      warning_type                                        warning
1         3     am   50   51        x no_global_binding    no visible binding for global variable ‘am’
2         3     vs   47   48        x no_global_binding    no visible binding for global variable ‘vs’
3         3  vs_am   38   42        x no_global_binding no visible binding for global variable ‘vs_am’
4         5    mpg   47   49        x no_global_binding   no visible binding for global variable ‘mpg’
5         5   qsec   53   56        x no_global_binding  no visible binding for global variable ‘qsec’
6         5  vs_am   65   69        x no_global_binding no visible binding for global variable ‘vs_am’

(x1 <- append_rlang(obj)) 
function () 
{
    data <- tidyr::unite(mtcars, col = !!rlang::sym('vs_am'), c(!!rlang::sym('vs'), !!rlang::sym('am')))
    ggplot2::ggplot(data = data, ggplot2::aes(x = !!rlang::sym('mpg'), y = !!rlang::sym('qsec'), 
        colour = !!rlang::sym('vs_am'))) + ggplot2::geom_point()
}
<environment: 0x10ceb96d8>

tidycheckUsage(x1)
NULL

x1()