Only reload affected areas when proofreading

[philippotto]

Rough measurements show that ~80% of bucket reloads can be avoided that way for an average proofread operation.

URL of deployed dev instance (used for testing):

• https://smartreload.webknossos.xyz

Steps to test:

• open l4dense_motta_et_al_dev_v2
• select mapping 80
• do proofreading with
  • merge and split (without skeletons)
  • load agglomerate skeletons and split/merge/min-cut these
• only the affected areas should reload

Issues:

• fixes During Proofreading, only reload "affected" buckets after split/merge #7022

---

(Please delete unneeded items, merge only when none are left open)

• Updated changelog
• Updated migration guide if applicable
• Updated documentation if applicable
• Adapted wk-libs python client if relevant API parts change
• Removed dev-only changes like prints and application.conf edits
• Considered common edge cases
• Needs datastore update after deployment

[daniel-wer]

Very elegant solution! 🤩

I didn't encounter any issues during testing. I found it hard to assess how much of the latency is due to the added containment check - did you measure this somehow? In any case, I think it's definitely worth it to avoid reloading all the data! Much better UX.

[philippotto]

did you measure this somehow?

Yes, I measured this (forgot to include it in the PR description). It takes about 1ms to compute the Set, which seems like a very good investment to me :)