MRG, BUG: Fix interactive update of mapping

[larsoner]

On main:

Peek.2021-03-03.12-05.mp4

On this PR:

Peek.2021-03-03.12-05_2.mp4

I don't know a good wait to test this on CIs unfortunately since it has to do with actual rendering.

[GuillaumeFavelier]

we could check that layered_mesh._rng == brain._cmap_range?

[larsoner]

we could check that layered_mesh._rng == brain._cmap_range?

Great idea, pushed. Will merge and backport once CIs come back happy

[larsoner]

Okay I fought through a bit of a nightmare with control flow with LUT updates. Now:

1. by use of a context manager we hopefully avoid one param change causing multiple .update calls (avoid time-based update limiting in the process)
2. make it so programmatic brain.update_lut(...) calls are reflected in the widgets
3. fix rng[0] == rng[1] not displaying properly -- now it creates what is essentially a binary mask at rng[0] which is what I'd expect

This also makes interaction smoother for me, which is nice. Ready for review/merge from my end @GuillaumeFavelier, hopefully CIs come back happy with this one...

[larsoner]

Actually I think I realized a simpler pattern, will implement it tomorrow

[larsoner]

Okay @GuillaumeFavelier should be good to go here

[GuillaumeFavelier]

LGTM, it is indeed very snappy now 👍

[GuillaumeFavelier]

Thanks @larsoner !