[bugfix] Fix duplicated nodes in maximal_independent_set output

[tpoterba]

Fixes #5535

[danking]

this needs to distinct before the if too. Neither case should produce duplicates

[tpoterba]

won't this require keying (and therefore shuffling) then?

disagree

[catoverdrive]

@danking this PR is showing up as approved on the CI, but not as approved here.

[tpoterba]

github API troubles?

[cseed]

Requesting changes to get the CI unstuck.

[cseed]

This fixes the CI: #5757. I will dismiss my review after that goes in.

#5757 merged

[danking]

won't this require keying (and therefore shuffling) then?

@tpoterba yeah but that's the correct answer no?

[danking]

Heh, this is all my fault: https://github.com/hail-is/hail/pull/5078/files

[tpoterba]

yeah, the keyed=False path is for performance and I'm OK leaving that with duplicated nodes

[danking]

@tpoterba if you're good with the verbiage I added to the notes, then I'll approve.

[tpoterba]

ya