Workaround PyTensor bug in vectorize of logp graph

[ferrine]

fix #7414

Description

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--7415.org.readthedocs.build/en/7415/

[ferrine]

what is this error in float64 tests, is it related?

[ricardoV94]

what is this error in float64 tests, is it related?

That's weird, values seem transposed

[ferrine]

feels to me this is failing check_pymc_draws_match_reference

[ferrine]

Another thing, I'm not able to reproduce the bug

(/home/ferres/dev/pymc/.venv) @c1 ➜ pymc git:(fix/vectorize-graph) PYTENSOR_FLAGS=floatX=float64 pytest tests/distributions/test_continuous.py::TestPolyaGamma::test_distribution
=========================================================================================== test session starts ============================================================================================
platform linux -- Python 3.11.9, pytest-8.2.2, pluggy-1.5.0
rootdir: /home/ferres/dev/pymc
configfile: pyproject.toml
collected 1 item

tests/distributions/test_continuous.py .                                                                                                                                                             [100%]

============================================================================================= warnings summary =============================================================================================
.venv/lib/python3.11/site-packages/pytensor/configdefaults.py:375
  /home/ferres/dev/pymc/.venv/lib/python3.11/site-packages/pytensor/configdefaults.py:375: DeprecationWarning: Use shutil.which instead of find_executable
    newp = find_executable(param)

-- Docs: https://docs.pytest.org/en/stable/how-to/capture-warnings.html
======================================================================================= 1 passed, 1 warning in 0.36s =======================================================================================

Should we merge instead and open a new issue to resolve that?

[ricardoV94]

If it's not failing on main we shouldn't merge. Did you retrigger the CI to see if it was a (very weird) fluke?

[ferrine]

same error, retriggered

[ferrine]

attempted to rebase on latest main

[ferrine]

I have not seen these transforms are even used by PolyaGamma

[ricardoV94]

I'll have to dig, is rather surprising, but somehow related

[zoj613]

I have not seen these transforms are even used by PolyaGamma

Can you try pinning polyagamma to 1.3.6 just to see if the tests fail. I just released 1.3.7 which includes numpy 2.0 support. I suspect this change might be the root cause.

Looking at this output:

E           AssertionError: 
E           Arrays are not almost equal to 6 decimals
E           
E           Mismatched elements: 14 / 15 (93.3%)
E           Max absolute difference: 0.25700269
E           Max relative difference: 7.27861386
E            x: array([0.175503, 0.304112, 0.32734 , 0.285465, 0.15268 , 0.208075,
E                  0.324094, 0.261004, 0.256576, 0.173434, 0.115312, 0.034482,
E                  0.121655, 0.047109, 0.358831])
E            y: array([0.358831, 0.047109, 0.121655, 0.034482, 0.115312, 0.173434,
E                  0.256576, 0.261004, 0.324094, 0.208075, 0.15268 , 0.285465,
E                  0.32734 , 0.304112, 0.175503])

It appears that the linked PR leads to values being generated in a reverse order than in v1.3.6....or at at least that's my immediate guess since i'm not quite familiar with the code that is being tested in the CI.

[ferrine]

I can confirm that updating the polyagamma to 1.3.7 I reproduce the error

[ferrine]

since this test is a blocker, I created an issue #7417 and a temporary solution #7418 to unblock any related work

[ferrine]

just rebased on the updated CI

[zoj613]

I've pulled 1.3.7 from Pypi and re-released it as v2.0.0 so you likely no longer need to pin the package. see : https://github.com/zoj613/polyagamma/actions/runs/10006421637

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.18%. Comparing base (8fd4f1c) to head (944478e).
Report is 94 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #7415   +/-   ##
=======================================
  Coverage   92.18%   92.18%           
=======================================
  Files         103      103           
  Lines       17259    17259           
=======================================
  Hits        15910    15910           
  Misses       1349     1349           

Files with missing lines	Coverage Δ
pymc/logprob/transforms.py	95.46% <100.00%> (ø)