ENH: Graphical Representation of Execution Layer Structure

[flongford]

At the moment the 'Execution Layer' model view displays redundant information. It would also be useful for the user to have a clearer understanding of how each layer and base data source fits together in a visual form.

This could be achieved by embedding an enable/kiva panel displaying each data source as a flow diagram-like object with arrows pointing to input/output variables being passed between them.

To begin with the representation could be static, though future adaptions could make this panel dynamic and therefore augment the current text-based UI structure.

In order to get this functional, some enhancements need to be made to the code:

• [ ] Method created to extract information on all input and output variables being passed from a stack of ExecutionLayers (probably located in VariableNamesRegistry)
• [ ] Graph-like object created that contains information on variables being passed between each DataSource and ExecutionLayer
• [ ] View created that takes in a graph-like object and returns a network display
• [x] Container class created to store view + object (#304: ProcessView)
• [x] Space in UI for container class view to be selected by user (#304)

[flongford]

Creating a directed graph object of the workflow using networkx would make this representation straight forward to display.

[flongford]

If we look to redesign the current prototype, we should probably use the ETS package graphcanvas https://github.com/enthought/graphcanvas