Reloading Python modules with R reticulate package

Question

I would like to know if there is a way to reload modules imported into R via the reticulate package, similar to the autoreload extension for IPython. For example, suppose I am developing my own module locally, I might import it via

library(reticulate)
import_from_path("mypackage", "/path/to/package")

However, if I make code changes to mypackage the only way I can think to reflect them is to restart R. This is consistent with normal Python behavior where modules cannot be reloaded. One workaround would be to source a file from the package in which I am making changes, i.e.

source_python("/path/to/package/file.py")

However, if file.py is importing other files in the local package that are being changed than changes to those files will not be reflected.

If you use devtools you can use `load_all()` to load your package in and reload when changes are made. — Dason, Aug 22 '18 at 15:39
My understanding is that `load_all` is just for R packages, whereas here I am importing a Python package into R. — Abiel, Aug 23 '18 at 02:00

score 2 · Answer 1 · answered Nov 22 '19 at 19:12

2

It's possible by calling Python commands to reload:

For Python 2.7

> builtins <- import_builtins()
> builtins$reload(your_module)

For Python 3

> importlib <- import("importlib")
> importlib$reload(your_module)

answered Nov 22 '19 at 19:12

Louis

766
7
10

Reloading Python modules with R reticulate package

1 Answers1