Load GPU cache without AYON containerized object set

[BigRoy]

Changelog Description

Load GPU cache with the AYON metadata imprinted directly onto the GPU cache shape node. This allows duplicating, deleting and instancing the GPU cache nodes as an artist usually whilst AYON would continue to be able to track it.

Additional review information

Because the gpuCache node already has an attribute called representation this meant that for this 'imprinting' to be allowed that the attributes were required to be prefixed with AYON_ to avoid clashes. Which means that listing loaded containers now requires looking for both the AYON_id and id attributes.

Testing notes:

1. Loading GPU caches should work
2. Managing GPU caches should work (incl. legacy GPU caches)
3. Check whether publishing e.g. layouts still work as intended, as they might be expecting an additional 'root' group node
4. Check whether publishing works with input links being detected correctly