Optimize get_structure_mask for structures with many children

[sdiebolt]

get_structure_mask can be optimized for structures with many children: see #119.

Thanks so much for creating and maintaining this awesome package! I'm not yet used to contributions on GitHub, feel free to let me know if I can improve anything.

[vigji]

@FedeClaudi @adamltyson tests are running fine locally on my machine, but there seems to be something funky in GitHub actions going on - would you agree to merge and fix tests in a separate PR?

[FedeClaudi]

I'm happy with it

[adamltyson]

TBH I'd rather we at least tried to fix the tests within the PR, unless it turns out to be a separate problem. Otherwise I fear we'll never get round to it.

[sdiebolt]

It seems like the fail comes from the dependency resolver: it should not have anything to do with the changes from this PR.

You might also want to try running the tests again: as the dependencies didn't change since last commit, it might either be a one-time bug from pip, or a problem introduced by a new version of one of the dependencies (in which case you might want to try to be more strict in requirements.txt?).

If the dependency resolver still fails after a rerun, I can try to reproduce it on my machine in the following days!

[sdiebolt]

Still there after a rerun... I just tried running pip install .[dev] on my machine with Python 3.8.12 (as defined in the GitHub Actions workflow) and there is indeed a lot of backtracking. I think this could be solved very simply by setting stricter version limits in requirements.txt!

Changing requirements.txt is best left to someone who is more aware of the exact use of each of those dependencies in BG-AtlasAPI, but do tell me if I can help. 🙂

[vigji]

Yes it's definitively not on you @sdiebolt to fix! I'll try to give it a shot in the next days

[adamltyson]

Having looked at this again, sorry @sdiebolt, you're right, the tests failing is nothing at all to do with this PR, we should merge this now, and fix the test failure separately.

[sdiebolt]

No problem, that is indeed a weird test failure!

[adamltyson]

@vigji are we good to merge? I haven't tested this myself yet.

[vigji]

yes, all tests are running locally! Although we'll break tests for main I guess