Jetpack

🔥 A friendly package manager for R

• Designed for reproducibility (thanks to Packrat, no more global installs!)
• Lightweight (adds just three files to your project)
• Secure by default
• Works from both R and the command line

Inspired by Yarn, Bundler, and Pipenv

Installation

Install Jetpack

install.packages("jetpack")

How It Works

Jetpack creates a DESCRIPTION file to store your project dependencies. It stores the specific version of each package in packrat.lock. This makes it possible to have a reproducible environment. You can edit dependencies in the DESCRIPTION file directly, but Jetpack provides functions to help with this.

Getting Started

Open a project and run:

jetpack::init()

Commands

Install

Install packages for a project

jetpack::install()

This ensures all the right versions are installed locally. As dependencies change, collaborators should run this command to stay synced.

Be sure to prefix commands with jetpack::. Jetpack isn’t installed in your virtual environment, so library(jetpack) won’t work.

Add

Add a package

jetpack::add("randomForest")

Add multiple packages

jetpack::add(c("randomForest", "DBI"))

Add a specific version

jetpack::add("DBI@1.0.0")

Add from GitHub or another remote source

jetpack::add("plyr", remote="hadley/plyr")

Supports these remotes

Bioconductor packages do not currently work

Add from a specific tag, branch, or commit

jetpack::add("plyr", remote="hadley/plyr@v1.8.4")

Add from a local source

jetpack::add("plyr", remote="local::/path/to/plyr")

The local directory must have the same name as the package

Update

Update a package

jetpack::update("randomForest")

For local packages, run this anytime the package code is changed

Update multiple packages

jetpack::update(c("randomForest", "DBI"))

Remove

Remove a package

jetpack::remove("randomForest")

Remove multiple packages

jetpack::remove(c("randomForest", "DBI"))

Remove remotes as well

jetpack::remove("plyr", remote="hadley/plyr")

Check

Check that all dependencies are installed

jetpack::check()

Info

Get info for a package

jetpack::info("stringr")

Get info for a specific version

jetpack::info("stringr@1.0.0")

Search

Search for packages

jetpack::search("xgboost")

Works with title, description, authors, maintainers, and more

Source Control

Be sure to commit the files Jetpack generates to source control.

Deployment

Server

Install Jetpack on the server and run:

jetpack::install(deployment=TRUE)

Docker

Create an init.R with:

install.packages("jetpack")
jetpack::install(deployment=TRUE)

And add it into your Dockerfile:

FROM r-base

RUN apt-get update && apt-get install -qq -y --no-install-recommends \
  libxml2-dev libssl-dev libcurl4-openssl-dev libssh2-1-dev

RUN mkdir -p /app
WORKDIR /app

COPY init.R DESCRIPTION packrat.lock ./
RUN Rscript init.R

COPY . .

CMD Rscript app.R

Heroku

There’s ongoing work to get Packrat working with the R buildpack.

In the meantime, you can use Docker Deploys on Heroku.

Command Line

Jetpack can also be run from the command line. To install the CLI, run:

jetpack::cli()

On Windows, add C:\ProgramData\jetpack\bin to your PATH. See instructions for how to do this.

All the Jetpack commands are now available

jetpack init
jetpack install
jetpack add randomForest
jetpack add DBI@1.0.0
jetpack add plyr --remote=hadley/plyr
jetpack update randomForest
jetpack remove DBI
jetpack check

You can also use it to manage global packages

jetpack global add randomForest
jetpack global update DBI
jetpack global remove plyr
jetpack global list

Or get info about packages

jetpack info stringr
jetpack info stringr@1.0.0
jetpack search xgboost
jetpack search "neural network"

For the full list of commands, use:

jetpack help

Upgrading

To upgrade, rerun the installation instructions.

0.3.0

Jetpack 0.3.0 greatly reduces the number of files in your projects. To upgrade a project:

1. Move packrat/packrat.lock to packrat.lock
2. Delete the packrat directory
3. Delete .Rbuildignore and .gitignore if they only contain Packrat references
4. Replace all Jetpack and Packrat code in your .Rprofile with:

if (requireNamespace("jetpack", quietly=TRUE)) {
  jetpack::load()
} else {
  message("Install Jetpack to use a virtual environment for this project")
}

5. Open R and run:

jetpack::install()

History

View the changelog

Contributing

Everyone is encouraged to help improve this project. Here are a few ways you can help:

• Report bugs
• Fix bugs and submit pull requests
• Write, clarify, or fix documentation
• Suggest or add new features