Skip to content
Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

Adding Python functionality to an R package

This guide demonstrates how to distribute Python code with your R package. The key ingredient is the reticulate package, which exposes Python code to an R session. To begin, ensure that Python and R are installed on your system. We will also make use of several R packages to make our lives easier. Install them by running the following command in R:

install.packages( c("devtools","usethis","reticulate") )

Create an R package

Create a skeleton for your package by running the following in R:

usethis::create_package( "~/projects/pypkg" )
usethis::use_package( "reticulate" )

replacing ~/project/pypkg with the desired name and path. The two commands instantiate an empty package called pypkg and ensure that reticulate is listed as its dependency. You may choose to add additional customization (such as authors, URL, etc.) by calling other usethis functions.

Add Python code

Inside your package create a inst/python/ subdirectory. By placing python/ directory inside inst/, you ensure that it will appear at the top level once the package is installed. Following the example above, let's create ~/projects/pypkg/inst/python/ and fill it with the following content:

import numpy as np

def pyadd(x,y):
    return x+y

def pymat():
    return np.arange(6).reshape(2,3)

The file defines two functions, the latter of which interacts with an external Python package numpy.

Expose Python code to R

Following the documentation for reticulate, create a new file ~/projects/pypkg/R/zzz.R with the following:

extra <- NULL

.onLoad <- function(libname, pkgname) {
    extra <<- reticulate::import_from_path("extra", path=system.file("python", package=pkgname))

As mentioned above, inst/python/ subdirectory will appear as python/ at the top level once the package is installed. The file above simply tells R to import the Python code we wrote above as extra module. For demonstration purposes, we use a global variable here. However, a better programming practice would be to write an accessor function that encapsulates the Python interface (see example).

Install the R package

That's it! You're done! Install your new package with

devtools::install( "~/projects/pypkg" )

and take it out for a spin:

library( pypkg )

## If using virtualenv to manage your external Python packages
reticulate::use_virtualenv( "/path/to/virtualenv" )

## If using conda instead of virtualenv
reticulate::use_condaenv( "/path/to/condaenv" )

# Module(extra)

# [1] 5

#      [,1] [,2] [,3]
# [1,]    0    1    2
# [2,]    3    4    5

Working example

A full working example is included in this repo. You can install it directly from GitHub by running the following commands in R:

if( !require(devtools) ) install.packages("devtools")


Adding Python functionality to an R package







No releases published


No packages published