11

For conda yml files one can define python dependencies from github using pip as

name: sample_env
channels:
dependencies:
   - python=2.7
   - pip:
     - git+https://github.com/my/repository

Is there a way to do this similarly (e.g. with devtools) for R packages using something like 

name: sample_env
channels:
dependencies:
   - R=3.6
   - devtools:
     - git+https://github.com/my/repository
Jonas
  • 1,639
  • 1
  • 18
  • 29
  • It kind of is though I was hoping to get an answer that doesn't require creating an own repository... – Jonas Jul 16 '19 at 15:25
  • It's only a suggestion; if you say that the answers listed in the linked question don't solve yours, I can revoke it. Personally, I doubt there's a way other than building conda packages from skeleton. – hoefling Jul 16 '19 at 15:41
  • no I think it's helpful to have the link here. I should have put it there in the first place. And I think the same, just hoping to be proven wrong by somebody who knows more than I do;) – Jonas Jul 16 '19 at 15:59
  • I'm reproposing closing as duplicate. There isn't anything like what OP asks, and there likely never will be in the foreseeable future. If one prefers not to host their package builds in an Anaconda Cloud user channel, then consider writing up [a Conda Forge recipe](https://github.com/conda-forge/staged-recipes) (although it is preferable that R packages source from CRAN and not GitHub). – merv May 12 '21 at 21:35

0 Answers0