Roadmap: Integrate HaploBlocker into Database objects

[josiahseaman]

Followup #19.
https://docs.google.com/document/d/1o_IWThHWK4qyyXk3-X4rZ2KJNGTbf-pPCRhzVmnQ1Xg/edit#
Much of haplonetwork.py needs to be rewritten to use Graph.models objects. Strand support could come at the same time or as a separate step.

Haplonetwork.py should use Graph.Node.

• upstream/downstream compared to Paths
• Should allow copy number variation in a Node to work inside Haploblocker

What does downstream mean in a graph context?