Support autoscaling/v2 API alongside autoscaling/v2beta1

[voelzmo]

What this PR does / why we need it:
Adds conversions to use autoscaling/v2beta1 in all calls to kube-apiserver when autoscaling/v2 is not available. Internally, the HVPA continues to operate on autoscaling/v2beta1 objects.

This is essentially what g/g does as well: we're using autoscaling/v2 for all k8s versions >=1.23, which introduced autoscaling/v2.

Which issue(s) this PR fixes:
fixes #104

Special notes for your reviewer:

• Supersedes Update to k8s.io/api/autoscaling/v2 #112 and Controller uses autoscaling v2 instead of v2beta1 when necessary #114
• HPA conversion logic is taken from k8s implementation, using the existing schema conversions in k8s 1.25.4

Release note:

HVPA supports k8s versions >= 1.25 by switching to `k8s.io/autoscaling/v2` when necessary for all API calls.

[voelzmo]

/invite @andrerun

[andrerun]

Looks good. Check the question about the v2 code syncing multiple fields where the v2beta1 code is only syncing MinReplicas. Also whether you actually want to update cache before validating the incoming object instance.

The rest are all cosmetic nits - just check if you want to make changes based on any of them.