A note for creating R-packages

This is the notes for myself on creating R packages in 6 steps, following Jeff Leeks’s guide and Hadley Wickham’s guide.

Step 1: Loading necessary packages

library("devtools")    
library("roxygen2")    

Step 2: Create package folder

create('packagename')    

This will create a folder with packagename in your working directory with the following files:

• DESCRIPTION
• NAMESPACE
• R (folder)
• Man (folder)
• packagename.Rproj (file)

Open packagename.Rproj with Rstudio, this will bring you to the package directory

Step 3: Add functions

In the R folder, write your desire functions in to files, one function per file.

Remeber to add the following lines (started with #’)on the top of each function.

#' Title    
#'    
#' Description..........    
#' ...................    
#'    
#' @param x1: \code{inputParameter1}    
#' @param x2: \code{inputParameter2}    
#'    
#' @return y1: what is it    
#'    
#' @keywords keywords    
#'    
#' @export    
#'    
#' @examples    
#' R code here showing how your function works    

try <- function(x1,x2){    
	y1 <- x1 + x2    
	return(y1)    
}    

@export is needed if this function is intended to be used by the users.

Step 4: Add manuals for functions

As the header lines for each function is written in the desired format, it can be automatically translate into manual page by devtools and roxygen2.

document("packagename")    

This writes the manual for each functions in to the man folder.

Step 5: Add informations into DESCRIPTION

The DESCRIPTION is automatically generated as:

Package: packagename    
Title: What the Package Does (one line, title case)    
Version: 0.0.0.9000    
Authors@R: person("First", "Last", email = "first.last@example.com", role = c("aut", "cre"))    
Description: What the package does (one paragraph).    
Depends: R (>= 3.2.2)    
License: What license is it under?    
LazyData: true    

Fill up the infomation without messing with the format, and add the following:

Imports:    
    dplyr,    
	stringi    
Suggests:    
	dplyr,    
	stringi    

This will tells R to install everything under Imports when this package is installed, whereas everything under suggests will not.

Step 6: Push to github

On github, create a repository naming it packagename

echo "# packagename" >> README.md    
git init    
git add *    
git commit -m "first commit"    
git remote add origin https://github.com/username/packagename.git    
git push -u origin master    

Now, the package can be install using the command:

library(devtools)    
install_github('packagename',username='username')    

Rcpp add on

So in the project Directory, the folder tree should be like:

.
├── DESCRIPTION
├── NAMESPACE
├── R
│   └── function1.R
│   └── function2.R
├── README.md
├── man
│   └── function1.Rd
│   └── function2.Rd
└── packagename.Rproj

Do the following:

mkdir src    
mkdir -p inst/include    

Put the .cpp files in the src and .hpp files in inst/include.
And do:

echo "PKG_CPPFLAGS += -I../inst/include/    
PKG_CXXFLAGS = $(CXX1XSTD)" >> src/Makevars    

For every cpp functions that needed to be export into R space,
Add the following before defining the function, no comments are allowed between this phrase and the funciton:

// [[Rcpp::export]]    
void function(int x, int y){    
}