Resource Allocation in Gen2

[Luxonis-Brandon]

Start with the why:

In Gen1, resources were easier to allocate because of the more static structure of the pipeline.

In Gen2, significantly more dynamic assignment of resources is required to enable the theoretically-infinite (Touring-like) nature of the Gen2 Pipeline Builder (#136)

Move to the how:

Convert many functions/core capabilities to allow dynamic loading, with managers that are aware of overall resource utilization.

Move to the what:

Implement a resource allocation system that allows the Gen2 pipeline builder to be scalably used.

[Luxonis-Brandon]

This is implemented and merged in Gen2, here: luxonis/depthai-python#74