Normalization extensions for VRS 2.x

As VRS grows to encompass more complex use cases, additional normalization constraints need to be imposed to promote consistency in computed identifiers. In VRS 1.x the primary normalization concern was correcting ambiguous sequence insertions/deletions in repeating regions. However, in 2.x we have additional concerns:

- [x] Preferentially representing a variant using RLE when possible
- [ ] Haplotype member ordering conventions

We should implement these additional normalization rules in VRS-Python.