- Confirm env creation, teardowns, etc.
- Use
lmod
as another layer of (compiler/software) control (e.g., via a "pyrty
module") - Any speed-up from cloning conda envs? Trade-off of larger mem on disk when cloning
- Fix
optparse
-ing: Ensureoptparse
is installed, if args;opt$<arg>
without explicitly adding to code - Backend script and env registries?
- Add a
from_template
class method toPyRFunc
for (1) returning dataframes, (2) saving dataframes, (3) building on tidyverse, and so on pyrty
-specific envfile- More informative error messages when R script errors. E.g., check if all libraries loaded in Rscript are installed
- Create a docstring for function in call signature
- Refactor env creation: different managers, different inputs (e.g.
mamba
using an envfile or using config or using list of packages) - Clean up signatures for
PyRFunc
funcs that return instances. Tackle after refactoring env creation. - Log env and script creation
- Use
Arrow
for big-data "`df"-capturing - Reduce
subprocess
calls for apply-style functions