render person graph

[endpnt]

Workflow:

1. cherrypy: (user adds selections via the UI to redis)
2. featext.py extract graph features from the selections (done)
3. featext.py: create graph with igraph (done)
4. graph2redis.py: take the graph and create the layout edges and vertices x,y postions (done)
5. graph2redis.py: save the layout x,y postions into redis (done)
6. cherrypy: read the layout postions from redis and render html5 / svg (done)

All the time consuming operations are de-coupled from the user interface. Some useful calculations can be done in redis when the user adds the selections, but this has to be balanced.

[endpnt]

initial version done in the igraph branch

walkthrough:

1. select persons in different documents
2. run ./featext.py
3. run ./graph2redis.py
4. navigate to /persons/graph