Allow lazy-loading chunk section data when using Vector API

[SirYwell]

Overview

Description

Previously, the use of the vector api forced us to load get and set arrays eagerly. This also caused some overhead in e.g. heightmap recalculation because a set array was present even if no changes were applied to a chunk section.

By using a thin wrapper, we can delay loading the data (or load dummy data) to when it is actually needed. This reduces the mentioned overhead.

Performance difference isn't really noticeable from a user's perspective, as the overhead ran pretty much in parallel to the edit itself. Manually inspecting profiles shows that there is less work done though (especially visible when using //count). This might be more noticeable on busy servers or servers with only 2-4 cores.

### Submitter Checklist
- [x] Make sure you are opening from a topic branch (**/feature/fix/docs/ branch** (right side)) and not your main branch.
- [x] Ensure that the pull request title represents the desired changelog entry.
- [x] New public fields and methods are annotated with `@since TODO`.
- [x] I read and followed the [contribution guidelines](https://github.com/IntellectualSites/.github/blob/main/CONTRIBUTING.md).