chircollab / chircollab20

Home for the 2020 Chicago R Collab event
https://chircollab.github.io/
10 stars 3 forks source link

Add package management with renv to one of your existing projects #24

Open jdblischak opened 4 years ago

jdblischak commented 4 years ago

Have you ever struggled to run someone else's code (or even your own after returning to it after some time) because the code doesn't run with the versions of the package you currently have installed? A package management tool like renv can help with this. It records the exact versions of the R packages that you originally used, so that in the future you or anyone else can recreate the exact computing environment to re-run or update the code.

I told the author of renv that we were looking for open source R packages to contribute to, and he said the most helpful thing would be to try out renv on various projects and report back any unexpected behavior or pain points (https://github.com/rstudio/renv/issues/359#issuecomment-594076913).

Thus if you have an existing project that you'd like to add package management support to, I'd be happy to help mentor you through the process of setting up and using renv.

ataustin commented 4 years ago

Love this idea -- and I would also be interested in making a comparison of different package management strategies for R (renv, packrat, switchr, etc).

jdblischak commented 4 years ago

a comparison of different package management strategies for R

@ataustin That'd be very useful! Like you, I mainly use conda, so I'm interested in learning about the other options.

A detailed comparison could also be useful for workflowr. I've long wanted to provide a uniform interface to package managers so that workflowr users can choose the "backend" they'd like to use. For example, can all package managers be broken down into steps such as "snapshot" and "restore"? See https://github.com/jdblischak/workflowr/issues/128 for discussion.

rpodcast commented 4 years ago

I use renv for almost every project at the day job, a big fan! Happy to share my findings.

ataustin commented 4 years ago

Excellent, let's discuss! I'm also a conda user day-to-day, but that's mostly because my team needs multiple languages. If I'm sticking to R, I'd love to learn more about renv @rpodcast