If you’re new to package development, this could be helpful to read.
Starting a new package/one-time commands.
devtools::create('generalconference')
devtools::install()
usethis::use_mit_license("Bryan Whiting")
usethis::use_github()
usethis::use_github_links()
# specify packages you want to use
usethis::use_package('dplyr')
usethis::use_package('rvest')
# create files:
usethis::use_testthat()
usethis::use_r('newscript')
usethis::use_test()
# builds data-raw/ folder
use_this::use_data_raw()
# build functions in R and save, builds a test file
usethis::use_r(name="new_func")
usethis::use_test()
# Add new package to DESCRIPTION as necessary
usethis::use_package('xxxx'):
# Once function is written, load it. You'll run `load_all()` multiple times.
devtools::load_all()
devtools::test()
devtools::check() # checks package_pkgdown.yml with documents
# Add new documentation
usethis::use_vignette('introduction') # add a vignette
# (optional, one-off steps) Build individual files
devtools::run_examples() # builds examples and vignettes
devtools::build_vignettes() #
pkgdown::build_articles() #
pkgdown::build_reference() # edit reference in _pkgdown.yml reference: section
# Prepare the package
devtools::document() # generates NAMESPACE from documentation. Exports functions.
covr::report() # run the coverage test
devtools::test() # run unit tests
devtools::check() # check the package
devtools::build() # build the package
pkgdown::build_site() # Build the r package documentationThe data are nested to minimize redundancy, but they can easily be unnested.
Example of nested data:
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
mtcars %>%
select(mpg, disp, am, vs) %>%
tidyr::nest(data = c(vs, c(mpg, disp)))
#> # A tibble: 2 × 2
#> am data
#> <dbl> <list>
#> 1 1 <tibble [13 × 3]>
#> 2 0 <tibble [19 × 3]>