feat: handle clab topo "kinds"

[carlmontanari]

we dont, and we need to. so... title?

[hellt]

This would also mean supporting defaults for a full resolution tree
defaults -> kinds -> node

In the long run this is definitely should be leveraging clab functions to not duplicate a considerable amount of code.

[carlmontanari]

copy

yep, I think from c9s perspective its just literally copying the stuff into the sub topology, so not really anything for c9s to do. I guess at the most complicated end we could parse the kinds and only put the relevant kind in the sub topo but for pass 1 I dont think thats even needed.

[carlmontanari]

I believe this is sorted now as we include kinds/defaults from parent topo into sub topos, so ima close for now!