You are very welcome to help out in the development of outsider!

If you have any ideas for future features then please create an issue. Otherwise, if you have the guile, time and inspriation dig deep into the code and fix/create, then please fork and send a pull request!

Outsider structure

The outsider universe is split between four different entities:

  • outsider.base: low-level interface between Docker and modules.
  • outsider: the user-friendly functions for running and installing modules.
  • outsider.devtools: toolkit for making it easier to develop modules.
  • “Test suites”: pipelines connecting modules to test holistic functionality.

If you wish to raise an issue or create a pull-request, please first determine which part of the outsider structure is most relevant. For example, if the problem seems to be Docker-related it is likely outsider.base. If it seems to be module discoverability it is likely outsider. (If in doubt select outsider as default.)

Areas for possible contribution

More modules!

If you think that you can code up a module that does not yet exist in outsider module form – then please it give a go!

A whole other package and website exists for helping uses contribute modules.


Any help with documentation to do with the development of modules would be greatly appreciated. In particular, more examples of module designs/structures along with descriptions. Additionally, any more tips and tricks or contributions to the FAQs would be equally welcomed.


The idea of outsider was to allow incorporation of code outside of R into R. But some of this integration can come at a computational cost: code that may have otherwise been run outside of an R analysis on a powerful desktop or a HPC facility could be run on a desktop instead. One solution so far has been to enable SSH’ing to remote servers. Any ideas on improving this or documentation on setting up hosting services (AWS, Google Cloud, Azure … etc.) that can interact with outsider would be greatly welcomed.

New test suite

outsider aims to make it easier to string pipelines together in R that make use of multiple external programs. To ensure that this is continually functional, as the project and its packages develop, a GitHub repo called “outsider-testsuites” is used to run a series of pipelines (“suites”) that run domain-specific pipelines. For example, “suite_1” runs a biological analysis for inferring an evolutionary tree of pineapple-like plants.

If you have created a series of outsider modules for a specific domain and think they would make for a given test suite, then please add a new pipeline to the “outsider-testsuites”!

How to contribute to an R package

To contribute you will need a GitHub account and to have knowledge of R (plus, depending on where you contribute, knowledge of bash and Docker). You can then create a fork of the repo in your own GitHub account and download the repository to your local machine. devtools is recommended.

devtools::install_github('[your account]/outsider')

All new functions must be tested. For every new file in R/, a new test file must be created in tests/testthat/. To test the package and make sure it meets CRAN guidelines use devtools.


For help, refer to Hadley Wickham’s book, R packages.

Style guide

outsider is developed for submission to ROpenSci. This means the package and its code should meet ROpenSci style and standards. For example, function names should be all lowercase, separated by underscores and the last word should, ideally, be a verb.

# e.g.
species_ids_retrieve()  # good
sppIDs()                # not great
sp.IDS_part2()          # really bad
sigNXTprt.p()           # awful

It is best to make functions small, with specific names. Feel free to break up code into multiple separate files (e.g. tools, helper functions, stages …). For more details and better explanations refer to the ROpenSci guide.

API tokens

The repository search features can make use of private tokens which make searching faster and more reliable. In the case of GitLab, however, an “access token” is a requirement for using the API.

A token is set up by generating one by logging into your code-sharing service and then saving the generated token to an R environment file.

To create a token for your code-sharing site, visit either for GitHub or for GitLab. The generated tokens should allow searching of all public repositories. Then add the token to your R environment using usethis::edit_r_environ(). The .Renviron file should contain the lines:

GITHUB_PAT=[your private token]
GITLAB_PAT=[your private token]

Code of Conduct

Please note that the ‘outsider’ project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.