r-lib
Is There a Plan to Include Linters for Reproducibility Best-Practices?
So, just to repeat the title: is there a plan to include linter rules for best practices to achieve reproducible scripts? With reproducible R scripts being a problem (e.g., outlined by Trisovic et al. (2022)) i think this would be a great addition!
For example linter rules could check:
- Are required/desired package versions specified (e.g., using groundhog)?
- Usage of rang/renv/...
- Problematic hardcoded paths (like the existing absolute_path_linter,...)
Or are these already covered somewhere (if so, i was unable to find them and apologize)? I would be happy to contribute regarding this!
Sure, what sorts of static analyses do you have in mind? Please file more specific follow-ups for tracking (you can keep this issue open as the master, c.f. #884).
The path linters could go for some love IMO, e.g. #468, #1356.
Thank you very much for the pointers! I hope I find time for them soon!
I think this is too broad. Feel free to open more specific follow-up issues.