Creating user-defined subgraphs

[Keavon]

• Double-clicking nodes (except those based on proto nodes) enters their subgraph, replacing what's currently viewed in the node graph
• A breadcrumb trail shows which node you're in relative to the document's root network
• The user can edit the nodes that make up a network-based document node
• The current "Output" node is generalized to support multiple graph exports, and an "Input" node is added to support (multiple) graph imports
• All the built-in nodes are given sane X/Y coordinates so users viewing those node internals aren't left with a mess
• Question for @TrueDoctor: what do we do with DocumentNodeImplementation::Extract nodes— just make them not double clickable as with DocumentNodeImplementation::ProtoNode?
• Do we need a way to edit the network's document node's type signature and Properties panel entries?

Future work (for another issue and PR):

• A way to edit a document node definition, not just a document node instance
  • This has no parent network, so unlike editing an instance, the currently-hooked-up imported/exported data has no effect on the wire colors inside the graph for generic data flow (it'll just be gray)
  • A way to define additional metadata and Properties panel layout for the definition
  • A way to save an entire definition in a nice compact and human-readable file format so we can replace the long document node definitions file with a folder each containing its own definition file in whatever format we choose (RON?)