Ivan Carvalho

Sounds good to me. Also, feel free to make this test Linux only. I am expecting some differences across operating systems. Let’s keep it simple, we just want to catch...

I think we can have `weighted_betweenness_centrality` as a method. The code will probably look completely different because we exploit the fact that `betweenness_centrality` is unweighted to speed the shortest path...

Our type annotations are added manually, so I think you just found a mismatch in the annotations! I might bundle the fix of this with #1242 for a 0.15.2 release....

From the CLA that everyone signs to contribute: ``` Grant of Copyright License. Subject to the terms and conditions of this Agreement, You hereby grant to IBM and to recipients...

I think with that regard nearly all of it will be IBM copyright. For code forked from NetworkX we kept the MIT license with their name.

Also, according to https://cibuildwheel.pypa.io/en/stable/options/#linux-image i686 is going to be deleted on `manylinux_2_28`. That is very soon. So I think downgrading i686 to Tier 4 is fairly reasonable.

~@mtreinish just a reminder to make `python3.10-arm64 ubuntu-24.04-arm MSRV ` the job blocking for the CI check if we merge this. Right now it checks `python3.10-x64 ubuntu-latest MSRV`~ we are...

Test run: https://github.com/IvanIsCoding/rustworkx/actions/runs/14986825229 (excluding macOS, but no changes were made to macOS tiers)

> Overall I'm fine with these changes they all make sense. I had a few inline suggestions, comments, and musings. But nothing major. I am wondering if we want to...

> The PR itself looks straightforward, but I wanted to understand how you're leveraging this. You mention views, but they're not typically defined via the output schema. Can you elaborate...