Lazy loading of Config Editor tabs

Minor release to address some performance problems with the config editor on Azure ML notebooks platform.
The control uses lots of ipywidgets and the AML notebook front-end has some perf issues with lots of widgets.
The improvement is to not load all controls/widgets for the settings editor - only load each tab as it is selected.

The release also includes a minor change to the entities classes - to allow them to be JSON serializable by adding a
JSON Encoder class to entities.Entity.

json.dumps(my_entity_list, cls=entities.Entity.JSONEncoder)