The previous part teaches you on how to import / export data from module, and in this part, you’ll learn on how to create a module that mimics the package system in R and package system in other languages such as Python. One of the strengths of {box} is that you don’t need to create an R package just to store the R codes and make it accessible to other end-users. Because of this package, you can mimic the way on how Python packages are created. It’s modular, maintainable, and accessible.
Recall the not_func.r script. Keep it and this time, create another scripts named convert.r and hello_world.r.
Recall that #' @export is only necessary when you are exporting specific part of the assigned R code into the namespace. When #' @export is placed, it triggers the Roxygen2 functionality, where it keeps the public namespace into the private namespace.
Copy this code:
celsius_to_fahrenheit = function (celsius) {
fahrenheit = (celsius * 9/5) + 32
fahrenheit
}
fahrenheit_to_celsius = function (fahrenheit) {
celsius = (fahrenheit - 32) * 5/9
celsius
}This example is an example code from the official documentation of {box} package.
#' @export
hello = function (name) {
message('Hello, ', name, '!')
}
#' @export
bye = function (name) {
message('Goodbye ', name, '!')
}And access them like:
box::use(
./module/convert,
greet = ./module/hello_world
)
temperature = 212
glue::glue(
"{temperature} degrees Fahrenheit is equivalent to {convert$fahrenheit_to_celsius(temperature)} degrees Celsius"
)212 degrees Fahrenheit is equivalent to 100 degrees Celsiusgreet$hello("Amy")If you have __init__.r initialization file already, you can do the following:
box::use(
md = ./module
)
temperature = 212
glue::glue(
"{temperature} degrees Fahrenheit is equivalent to {md$convert$fahrenheit_to_celsius(temperature)} degrees Celsius"
)212 degrees Fahrenheit is equivalent to 100 degrees Celsiusmd$hello_world$hello("Amy")And by the way, you can apply aliases for the script accessed as a module, not limited to functions and any objects:
box::use(
./module[hw = hello_world],
./module[cv = convert],
./module[nf = not_func]
)Under the {./module}, let’s create a subfolder named statistics, where the small set of statistical functions are contained in separate scripts. Don’t forget the __init__.r initialization file.
Here is the structure:
statistics/
├── __init__.r
├── cor.r
├── corrr.r
└── time_series.rSee the source code of each script here: https://github.com/kisha126/modules-in-r/tree/main/module/statistics
Access all of the modules within statistics by:
box::use(
md_sts = ./module/statistics
)Example:
with(cars, md_sts$cor$custom_cor(speed, dist))[1] 0.8068949cor_pipe = md_sts$corrr$cor_pipe
mtcars |>
cor_pipe(mpg, hp, disp, qsec) mpg hp disp qsec
mpg 1.0000000 -0.7761684 -0.8475514 0.4186840
hp -0.7761684 1.0000000 0.7909486 -0.7082234
disp -0.8475514 0.7909486 1.0000000 -0.4336979
qsec 0.4186840 -0.7082234 -0.4336979 1.0000000mas = md_sts$time_series$moving_average(AirPassengers, win = 3)
plot(
seq_along(AirPassengers),
AirPassengers,
type = "l",
col = "blue",
lwd = 2,
xlab = "Time Index",
ylab = "Passengers",
main = "AirPassengers with Moving Averages"
)
lines(mas, col = "red")
The {statistics} submodule is accessible:
box::use(
./module/statistics
)If __init__.r file under {statistics} subfolder is initialized through:
#' @export
box::use(
./cor_nse,
./cor,
./moving_average
)I’ll explain about nested modules in the Chapter 3.4.
If you access {statistics} subfolder as a module, under {./module} folder like this:
box::use(
./module[st = statistics]
)The {st} module will not be loaded into the current (root) environment, but loaded under {./module} environment
