fix flaky test, remove high tolerance

[bwengals]

Description

The multi-output GP test was flaky. Turns out that it was lucky that it was passing occasionally. It was testing whether two GP implementations were correct when they were subtly different. pm.gp.Latent.conditional was conditioning on the original gp f, and pm.gp.Marginal.conditional was conditioning on the observed data y. The fix is to condition pm.gp.Latent.conditional on y also in the test.

Related Issue

• Closes Flaky tests/gp/test_gp.py::TestMarginalVsLatent::testLatentMultioutput #7531
• 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--7567.org.readthedocs.build/en/7567/

[ricardoV94]

Failed in windows seems to need a bit more tolerance

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.82%. Comparing base (f551718) to head (27058f3).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #7567      +/-   ##
==========================================
- Coverage   92.83%   92.82%   -0.01%     
==========================================
  Files         106      106              
  Lines       17669    17669              
==========================================
- Hits        16403    16402       -1     
- Misses       1266     1267       +1     

see 1 file with indirect coverage changes

[bwengals]

oh windows. increased rtol a bit

[ricardoV94]

Thanks @bwengals