Bump pymc dependency

[ricardoV94]

Closes #726

---

📚 Documentation preview 📚: https://pymc-marketing--1269.org.readthedocs.build/en/1269/

[juanitorduz]

Thanks for this one!

[codecov]

Codecov Report

Attention: Patch coverage is 54.54545% with 5 lines in your changes missing coverage. Please review.

Project coverage is 36.10%. Comparing base (39028fe) to head (e69031b).

Files with missing lines	Patch %	Lines
pymc_marketing/clv/distributions.py	54.54%	5 Missing ⚠️

❗ There is a different number of reports uploaded between BASE (39028fe) and HEAD (e69031b). Click for more details.

HEAD has 8 uploads less than BASE

Flag	BASE (39028fe)	HEAD (e69031b)
	11	3

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #1269       +/-   ##
===========================================
- Coverage   95.34%   36.10%   -59.24%     
===========================================
  Files          47       47               
  Lines        4963     4913       -50     
===========================================
- Hits         4732     1774     -2958     
- Misses        231     3139     +2908     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[juanitorduz]

There are just two failing tests (yay!)

FAILED tests/clv/models/test_gamma_gamma.py::TestGammaGammaModel::test_spend[True]
FAILED tests/test_model_builder.py::test_fit_random_seed_reproducibility[rng]

[wd60622]

Closes #726

[ricardoV94]

FAILED tests/clv/models/test_gamma_gamma.py::TestGammaGammaModel::test_spend[True]

AFAICT it's a seed problem. standard deviation is similar to old version, and I could get it to fail then if I changed the seed. Making tolerance higher

[ricardoV94]

The second failure seems to be a regression in PyMC regarding seeding with generators: pymc-devs/pymc#7612

[ricardoV94]

Thanks for this one!

Very specific appreciation noted :D

[juanitorduz]

The second failure seems to be a regression in PyMC regarding seeding with generators: pymc-devs/pymc#7612

So, do we need to wait for this to be fixed in PyMC before merging this one? Or shall we skip this test (in the meantime)?

[ricardoV94]

This is buggy behavior so I would wait until the next release with the fix

[PabloRoque]

Agua de Mayo

[ColtAllen]

This is buggy behavior so I would wait until the next release with the fix

The next pymc-marketing release, or next pymc release?

[ricardoV94]

Agua de Mayo

What does that mean? 🤔

[ricardoV94]

This is buggy behavior so I would wait until the next release with the fix

The next pymc-marketing release, or next pymc release?

Next pymc release

[PabloRoque]

What does that mean? 🤔

Agua de Mayo may be translated as something is godsend.

Context: I got a new Apple Silicon machine last week. Making blas work has been painful. I made it work following this release of pytensor you mentioned somewhere else.

However that version is not supported in the pymc-marketing library. So indeed this PR is godsend to me.

[aseyboldt]

@wd60622 Is there anything holding this back?

[wd60622]

@wd60622 Is there anything holding this back?

@ricardoV94 mentioned non deterministic behavior with pymc version