PR review request – higher order interpolations

[hexane360]

Hi, similar to #33164, I'm finding the review process a bit unclear. I have a PR #33970 which implements a feature requested in the (admittedly stale) issue #3928. I'm not quite sure the code is mergeable at this point but would appreciate if someone could take a look and/or give their feedback.

The code is a pretty straightforward re-implementation of scipy's version (and the original source paper they based spline calculations on). I can also attest that no AI was used to draft the PR.

I appreciate the work done on jax, and realize it's a huge burden to review out-of-the-blue PRs like this. In the future I'll start a discussion before opening the PR; that said, the functionality is critical for parity with scipy and the implementation problems in #3928 have been resolved by scipy's improved boundary condition handling.

[jakevdp]

Hi - I think I ignored your PR when I saw it because the first line said it was a draft. If it's ready for review, you should make that clear in the description!

Sorry for only skimming it, but we have a lot of traffic lately (especially with the uptick in garbage AI PRs) and I don't always have time to read the description deeply – so when I see "draft" in the first sentence I move along.

I'll take a look now.