Prototype data architecture

[fjoanis-legion]

Based on https://hackmd.io/uIb_hEm4QXG9Z4ZwQut15Q, make this architecture in a workable prototype to prove that this architecture is working and is the right solution going forward.

• [x] Build graph is processed recursively (Only the root node is sent to compilation and the whole graph is built.)
• [x] Worker takes element in a queue and compiles them
• [x] Test runtime / build dependencies build variation (i.e. Resource A has a runtime dependency on Resource B that has a build dependency on Resource C)
• [x] Invalidate a node and recompile the minimum amount of data
• [ ] Support multiple compilers versions / Map between GitHub commit ID and compiler to use
• [x] Run compilers as .dll
• [x] Support queries dependencies (Navmesh depends on all collision primitive and is invalidated on add/remove/update of collision position/shape)
• [ ] AWS Data execution & Broker to choose what runs locally vs cloud
• [x] Compiler definition similar to Bevy systems (Make it dead simple to create new compilers)