Roll HSGPPeriodic into HSGP

[maresb]

Description

Not much here yet, but the intention is to eliminate HSGPPeriodic in favor of HSGP as a first step to implementing boundary conditions.

Related Issue

• [ ] Closes #
• [ ] Related to #

Checklist

• [ ] Checked that the pre-commit linting/style checks pass
• [ ] Included tests that prove the fix is effective or that the new feature works
• [ ] Added necessary documentation (docstrings and/or example notebooks)
• [ ] If you are a pro: each commit corresponds to a relevant logical change

Type of change

• [ ] New feature / enhancement
• [ ] Bug fix
• [ ] Documentation
• [ ] Maintenance
• [ ] Other (please specify):

---

📚 Documentation preview 📚: https://pymc--7470.org.readthedocs.build/en/7470/

[maresb]

Ping @bwengals @AlexAndorra

[maresb]

TODO: why is it sqrt_psd in HSGP and psd in HSGPPeriodic? Which is it, or is the psd just not square rooted in the periodic version?

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 92.17%. Comparing base (68a7578) to head (5362113).

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #7470      +/-   ##
==========================================
+ Coverage   87.49%   92.17%   +4.67%     
==========================================
  Files         103      103              
  Lines       17215    17220       +5     
==========================================
+ Hits        15063    15872     +809     
+ Misses       2152     1348     -804     

Files	Coverage Δ
pymc/gp/hsgp_approx.py	88.70% <100.00%> (+0.31%)	:arrow_up:

... and 8 files with indirect coverage changes

[bwengals]

TODO: why is it sqrt_psd in HSGP and psd in HSGPPeriodic? Which is it, or is the psd just not square rooted in the periodic version?

The goal was a direct implementation appendix B from this paper

[maresb]

A GP model with a periodic covariance function does no fit in the framework of the HSGP approximation covered in this study

I'm pretty sure it fits in with the manifold HSGP framework outlined in ... that paper. So I think that if we cite that as well we could remove the comment about it not being a real HSGP.

[maresb]

Oh!!! Sorry @bwengals, I wasn't reading that closely enough. They're not quite doing what I expected. I'm 99% sure it's not qualitatively different from what I had in mind, but I need to read up and figure exactly how what they're doing differs from the canonical manifolds-based approach.

[maresb]

I just did a rebase because I was accidentally working from an old version of the main branch. There are no new changes relative to the rebase.