Build pre-filtered networks

[greebie]

Create a network output that includes sufficient analytic data to create attractive graphs in sigma.js.

[greebie]

Two approaches --

1. collect data from AUT output and use Python libraries to calculate algorithms.
2. create generic network output and form it as json.

[ianmilligan1]

While we now output GEXF as of #137, we are still interested in creating "sufficient analytic data to create attractive graphs in sigma.js." Some thoughts from @greebie:

The sparksql api has some pretty good tools for basic analysis etc. that could help us form the networks out of the box (eg. have a preset network look without having to touch the Yifan Hu algorithms, filtering etc.)

[greebie]

Re-named to remove references to GEXF, which is resolved.
What I would like to have here is some preset analytics to filter network graphs efficiently.

The goal is to have a nice attractive network out of the box and avoid the "Borg cube" in Gephi.

Looking at GraphStream and other scripts, the best we can do is O(n^2) which would not work for large networks. I could try to produce a script that checks for size first.

[ianmilligan1]

This is being handled outside of AUT with borg-reducer