JavaScript in R

[jeroen]

The V8 package (a fully standalone JavaScript engine in R) is pretty stable by now. If you have some JavaScript scripts or libraries that you would like to expose in R, I can help writing bindings. I am also interested to learn if you use JavaScript in R in some other way (shiny or htmlwidgets or so); perhaps we could support server side validation/preprocessing.

Simple examples: viz.js or sql.js or cheerio. Also try the js package, which implements bindings to uglify-js and jshint:

devtools::install_github("jeroenooms/js")

Basic examples:

library(js)
code <- "function test(x, y){ x = x || 1; y = y || 1; return x*y;}"
cat(uglify_optimize(code))
cat(uglify_reformat(code, beautify = TRUE, indent_level = 2))
jshint(uglify_optimize(code))

[noamross]

+1 on this. I'm interested in the relationship between this and #3 (API wrappers), and possibly Rcpp, in that I'd like to discuss what design patterns should be used for javascript-wrapping packages. Related questions:

• What essential tasks does it make the most sense to do in JS, and/or what tasks are already well developed in JS that should be wrapped for R?
• If we need disk I/O, should it run through R, should we wrap node/io.js, or is it coming down the pipe for V8?
• Best way to integrate JS and R package development workflow (versioning/NPM/browserifiing/folder conventions/etc).
• Can we create templates or other tools for easily making JS-wrapping packages?

Relatedly, I'm interested in something that uses PhantomJS or something like it to scrape pages as rendered in browsers, a la quickscrape, which I tried to wrap but haven't made much progress on.

[srvanderplas]

I'm interested in something that uses PhantomJS or something like it to scrape pages as rendered in browsers, a la quickscrape, which I tried to wrap but haven't made much progress on.

I missed this initially, but I would be very interested in this as well