renv: Project Environments for R

This article is originally published at https://www.rstudio.com/blog/

We’re excited to announce that renv is now available on CRAN! You can install renv with:

install.packages("renv")

renv is an R dependency manager. Use renv to make your projects more:

• Isolated: Each project gets its own library of R packages, so you can feel free to upgrade and change package versions in one project without worrying about breaking your other projects.

• Portable: Because renv captures the state of your R packages within a lockfile, you can more easily share and collaborate on projects with others, and ensure that everyone is working from a common base.

• Reproducible: Use renv::snapshot() to save the state of your R library to the lockfile renv.lock. You can later use renv::restore() to restore your R library exactly as specified in the lockfile.

If you’ve used Packrat before, this may all feel familiar. User feedback has made it clear that a number of the decisions we made during Packrat’s development ultimately made it frustrating to use, and led to a sub-optimal user experience. The goal then is for renv to be a robust, stable replacement for the Packrat package, with fewer surprises and better default behaviors. While we will continue maintaining Packrat, all new development will focus on renv.

In addition, we’ve built renv to work well with R projects using Python through reticulate. Using renv, you can also create project-local Python environments, and instruct reticulate to automatically bind to, manage, and use these environments.

Getting Started

The core essence of the renv workflow is fairly simple:

1. Use renv::init() to initialize a project. renv will discover the R packages used in your project, and install those packages into a private project library.

2. Work in your project as usual, installing and upgrading R packages as required as your project evolves.

3. Use renv::snapshot() to save the state of your project library. The project state will be serialized into a file called renv.lock.

4. Use renv::restore() to restore your project library from the state of your previously-created lockfile renv.lock.

In short: use renv::init() to initialize your project library, and use renv::snapshot() / renv::restore() to save and load the state of your library.

After your project has been initialized, you can work within the project as before, but without fear that installing or upgrading packages could affect other projects on your system.

Collaborating

When you want to share a project with other collaborators, you may want to ensure everyone is working with the same environment – otherwise, code in the project may unexpectedly fail to run because of changes in behavior between different versions of the packages in use. You can use renv to help make this possible.

When using renv, the packages used in your project will be recorded into a lockfile, renv.lock. Because renv.lock records the exact versions of R packages used within a project, if you share that file with your collaborators, they will be able to use renv::restore() to install exactly those packages into their own library. This implies the following workflow for collaboration:

1. Select a way to share your project sources. The most common way nowadays is to use a version control system with a hosted repository; e.g. Git with GitHub, but many other options are available.

2. Make sure your project is initialized with renv by calling renv::init().

3. Call renv::snapshot() as needed, to generate and update renv.lock.

4. Share your project sources, alongside the generated lockfile renv.lock.

After your collaborators have received your renv.lock lockfile – for example, by cloning the project repository – they can then also execute renv::init() to automatically install the packages declared in that lockfile into their own private project library. By doing this, they will now be able to work within your project using the exact same R packages that you were when renv.lock was generated.

Time Travel

On some occasions, you might find that you’ve made a change to renv.lock that you’d like to roll back. If you’re using Git for version control with your project (and we strongly encourage you to!), renv has a couple helper functions that make it easy to find and use previously-committed versions of the lockfile.

• Use renv::history() to view past versions of renv.lock that have been committed to your repository, and find the commit hash associated with that particular revision of renv.lock.

• Use renv::revert() to pull out an old version of renv.lock based on the previously-discovered commit, and then use renv::restore() to restore your library from that state.

If you have an alternate version control system you’d like to see us support, please let us know!

Integration with Python

renv also makes it easy to set up a project-local Python environment to use with your R projects. This can be especially useful if you’re using the reticulate package, or other packages depending on reticulate such as tensorflow or keras. Just call:

renv::use_python()

and a project-local Python environment will be set up and used by reticulate. When renv's Python integration is active, a couple extra features will activate:

1. renv will instruct reticulate to load your project-local version of Python by default, avoiding some of the challenges with finding and selecting an appropriate version of Python on the system.

2. Calling reticulate::py_install() will install packages into the project’s Python environment by default.

3. When renv::snapshot() is called, your project’s Python library will also be captured into requirements.txt (for virtual environments) / environment.yml (for Conda environments).

4. Similarly, renv::restore() will also attempt to restore your Python environment, as encoded in requirements.txt / environment.yml from a previous snapshot.

Packrat

If you’ve used Packrat before, you’re likely interested to learn what’s changed in renv. We’ll try to summarize the most poignant changes:

Project Initialization

packrat::init() would, by default, attempt to retrieve package sources from CRAN under the assumption that you might want to rebuild packages from sources in the future (e.g. in an offline environment). This assumption was rarely true, and still often was unhelpful as many packages are difficult to build from sources.

To alleviate this, renv::init() no longer downloads package sources, and also attempts to copy and reuse packages already installed in your R libraries. This makes initializing new projects a breeze – you no longer have to sit around and wait as your project’s multitude of dependencies get reinstalled; instead, the copies already available on your system will be copied and re-used.

Snapshots and Dependencies

packrat::snapshot() would, in addition to capturing the state of your project library, also attempt to discover the R packages used in your project by crawling your .R and .Rmd files for dependencies. Unfortunately, this system was fairly unreliable and caused a number of issues, especially when the machinery itself emitted warnings or errors that could not be easily diagnosed.

The dependency discovery machinery in renv has been rewritten from the ground up, and should now be much more reliable. However, if you discover that this still causes issues for you, you can disable this altogether by changing the type of snapshot performed in your project. Use renv::settings$snapshot.type("simple") to use “simple” snapshots in your project, where the state of your library is captured as-is without any extra filtering to limit which of your installed packages enter the lockfile.

Extra Tools

In addition, renv comes with a couple extra tools out-of-the-box to help with common development workflows:

• Install packages from a wide variety of sources with renv::install(). renv::install() understands a subset of the remotes specification, and so can be used for simple, dependency-free package installation in your projects. Currently, you can install packages from CRAN, GitHub, Gitlab, and Bitbucket. In addition, renv is also compatible with other tools commonly used to install packages, such as remotes and pak.

• Use renv::dependencies() to enumerate the R dependencies in your project. If necessary, use .renvignore files to tell renv which files and folders should not be scanned during dependency discovery.

Finally, if you have a Packrat project that you’d like to try porting to renv, you can use renv::migrate() to migrate the project infrastructure over to renv.

Learning More

Please check out the renv Getting started guide to learn more. If you are looking for strategies to manage reproducible environments, or don’t know if renv is the right fit, check out https://environments.rstudio.com. If you have questions or comments, please get in touch with us on the RStudio community forums.

Thanks for visiting r-craft.org
This article is originally published at https://www.rstudio.com/blog/
Please visit source website for post related comments.