Publish JJUS to npm about json-stat

[badosa]

As issue #14 is probably not trivial, I decided to publish JJUS on npm as it is.

npm install jsonstat-utils

tbrowser has been removed from the npm version.

[badosa]

I never thought JJUS as a module but more as a bundle of different high level helper functions, some of them (tbrowser) they even require a browser and a DOM.

Give me some time to work on it.

Would it make sense to you to also add some of the functionality of JJUS to jsonstat-conv (Command line tools for converting JSON-stat documents)? Things like:

jsonstat2jsonstat (to support join())
arrobj2jsonstat
csv2jsonstat
...

[jfsiii]

I never thought JJUS as a module but more as a bundle of different high level helper functions, some of them (tbrowser) they even require a browser and a DOM.

There are many "high-level helper function" packages on npm. npm isn't just for node modules. Even before bower became deprecated
npm was a great way to publish modules.

Putting JJUS on npm allows us to install and import it like all the rest of our dependencies.

Would it make sense to you to also add some of the functionality of JJUS to jsonstat-conv (Command line tools for converting JSON-stat documents)? Things like:

jsonstat2jsonstat (to support join())
arrobj2jsonstat
csv2jsonstat

That sounds like a nice addition. Especially in light of #6

[badosa]

a bundle of different high level helper functions

What I also meant by this is that JJUS was a way for me to offer in a bundle some unrelated JSON-stat code that I had been using for demos and examples and that I thought it might be useful to share. What actually triggered the publication of JJUS was the discovery that not up-to-date code from old examples of the table browser had been copied and was used in several sites.

The fact that this code was originally developed for demos and examples explains why there wasn't a focus on performance: its main goal was to show that a particular functionality was doable in JSON-stat. That was specially true of join(): this method was a way to try to start a discussion on what sort of merges made sense. Which explains issue #14.

It probably makes sense to face issue #14 first before publishing JJUS on npm.