We're already creating dynamic HighCharts for our webpages--these have a few javascript dependencies and a chart-generation script. We'd like to start creating PDFs of these charts using the new HighCharts/Node/PhantomJS suite HighCharts has rigged up (see press release). Our image-gen node server would run on a separate box/vm. How can we do this without having to maintain two separate codebases for the same chart? I'm not too familiar with Node yet, so I'm not sure how requesting the scripts with a web request would work. I'm guessing a lot of HighCharts users that want to start generating some of their charts as images server-side will run into a similar issue with managing two related codesets...
In essence, we already have have this:
Webserver -> JSON (data) + foo.js + bar.js + foo.html -> webpage with dynamic charts.
We'd like to build:
Web-server -> JSON (data) + separate Node Server + foo.js + bar.js -> images available via web request
Obviously some redundancy. How can we manage the dependencies?
We decided to pass the entire highcharts config object (e.g. Highcharts.chart(configObj)) to the node server as a URL-encoded string. We had to put a few rendering functions over on the node server, but it wasn't too bad. We also stuck some of the rendering functions in the string config object. Not the most beautiful result, but it worked.