[FEATURE] Possible To Make Syntax More R-Like?

[TheCedarPrince]

Hi @stefan-m-lenz,

Hope you are well! I was working again with JuliaConnectoR recently and was wondering: is it possible to make the syntax using JuliaConnectoR even more R-Like? Meaning, for example, given this current syntax:

library(JuliaConnectoR)

pkg <- juliaImport("Pkg")

pkg$activate("STUDY", shared = TRUE) 

Could it somehow be reformed to look like this:

library(JuliaConnectoR)

juliaImport("Pkg")

Pkg::activate("STUDY", shared = TRUE) 

Or even:

library(JuliaConnectoR)

Pkg <- juliaImport("Pkg")

Pkg::activate("STUDY", shared = TRUE) 

This was some feedback I had gotten from some collaborators when using the package within R. Otherwise, thanks for the great work!

~ tcp :deciduous_tree:

[stefan-m-lenz]

Hi @TheCedarPrince , Good to hear that you still find the package to be useful!

In the first version of the package I actually wanted to make the syntax more Julia-like and I had implemented the following approach:

juliaImport("Pkg")
Pkg.activate("STUDY")

To achieve this, I had to use the attach function. When I was trying it to publish it on CRAN, it was not going to be accepted like that. The reason was that functions in packages should not alter the search path. I had to switch to the current syntax to get the package accepted on CRAN. If one were to implement the suggested syntax with the ::, I'm pretty sure that similar problems would arise when trying to publish this on CRAN because this would also require to alter the search path.

[TheCedarPrince]

Oh interesting. I was confusing Julia metaprogramming syntax with R metaprogramming as I thought you could just "invent" your own syntax within R to mimic the standard :: R operator. Since that may not be possible, another avenue of thought is could it be possible to import specific Julia functions? Like this for example:

library(JuliaConnectoR)

activate <- juliaImportFunction("import Pkg: activate")

activate("STUDY", shared = TRUE) 

I know some masking could occur in this scenario but was thinking the juliaImportFunction could bind the function Pkg.activate as an object to activate.

But yea, in terms of feedback from collaborators, they are fine using Julia as a backend for R, but there was some concern about syntax like pkg$activate. I was thinking the above could be a potential compromise to address that concern.

What do you think?

[stefan-m-lenz]

Pkg$activate is simply an R function. You can assign it to another R variable as any R function if you want:

> library(JuliaConnectoR)
> Pkg <- juliaImport("Pkg")
> activate <- Pkg$activate
> activate("STUDY", shared = TRUE) 

Or like this, after a call to juliaImport or juliaEval("import Pkg"):

> activate <- juliaFun("Pkg.activate")
> activate("STUDY", shared = TRUE) 

So there's no need for an extra function for that.

[TheCedarPrince]

OH! I didn't realize you could do this:

activate <- juliaFun("Pkg.activate")
activate("STUDY", shared = TRUE) 

That entirely solves what I was imagining! Thank you! I'll go ahead and close this issue. :)