rbokeh has been removed from CRAN about rbokeh

[bryevdv]

@lcougnaud rbokeh is currently unmaintained. Several years ago, there was a good chunk work done to put rbokeh on more maintainable footing, with automatically generated low-level bindings to BokehJS, with "idiomatic R" interface built on top of that stable foundation. Unfortunately that work was never pushed out to a release (and in fact, may be lost except in online diff of this closed PR #217, cc @hafen)

Bokeh itself has come very far since the last official release of RBokeh, with an enormous amount of new features. I would love to see someone update and re-vitalized RBokeh. Unfortunately I am not actually an R user, so that's not a role I can contemplate taking on. It's just (well) outside what I can realistically accomplish given my other commitments. But if you (or someone you know) would is looking to be more involved in OSS and would like to take on the task of picking up and maintaining RBokeh, I can commit to offering technical quidance and/or mentoring.

Edit: and to answer this question

Is there a plan to include it back to CRAN?

I don't actually know how or by what mechanism it was published on CRAN, and similarly don't know how or why it was removed. Perhaps @hafen has insights there.

[hafen]

@lcougnaud sorry for the inconvenience of this getting removed from CRAN. It was removed for a simple (but apparently serious) violation of CRAN policy.

This creates and leaves behind a file test.png in /tmp during R CMD check --as-cran, in the \donttest part of the widget2png example (which is run by example()).

This is a serious violation of the CRAN policy.

I'm working on a fix now.

[lcougnaud]

Thanks @hafen & @bryevdv for the quick feedback!

@bryevdv I unfortunately don't have the bandwidth to help further for the maintenance of the rbokeh package.

@hafen for this specific issue, could this be solved by:

\donttest{
file <- tempfile(fileext = ".png")
figure(tools = NULL) %>%
  ly_points(1:10) %>%
  widget2png(file)
}

[bryevdv]

Serious they say!

[hafen]

Yeah I know 🙄

[hafen]

As simple as the problem is, addressing it will likely open a can of worms due to changing CRAN policies and package dependencies since the last submission. @lcougnaud sorry I let this get by me.

[hafen]

I submitted to CRAN yesterday. We'll see how quickly that goes.

[philibe]

I've done a POC with R Shiny and bokeh in pure R Shiny with session$registerDataObj() (without rbokeh and shiny.router) :) :

• https://github.com/philibe/RshinyBokehReticulatePOC

(I didn't know that rbokeh exists.)