Add Poisson and NegativeBinomial to pymc.dims

[williambdean]

Summary

Add Poisson and NegativeBinomial to pymc.dims to unblock the Bass model dimension migration (pymc-labs/pymc-marketing#2598). These are the last two distributions needed before the Bass model can migrate from create_dim_handler to the xdist=True / pymc.dims pattern used by the rest of pymc-marketing.

Changes

• pymc/dims/distributions/scalar.py: Add Poisson (wraps ptxr.poisson, accepts mu) and NegativeBinomial (wraps ptxr.nbinom, accepts mu/alpha or n/p parameterization) as DimDistribution subclasses

• tests/dims/distributions/test_scalar.py: Add test_poisson and test_negative_binomial following the existing pattern — create a model with the dims distribution + reference model with the regular distribution, assert equivalent random and logp graphs

Background

These are the first discrete count distributions in pymc.dims. Poisson is the default likelihood for the Bass model, and NegativeBinomial is the alternative. DiracDelta is out of scope (can be handled by supporting non-Prior configurations in pymc-marketing).

The underlying PyTensor xtensor operations (ptxr.poisson, ptxr.nbinom) already existed — only the pymc.dims wrapper classes were missing.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 91.73%. Comparing base (88dc395) to head (cb0bf6b).
⚠️ Report is 3 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8305      +/-   ##
==========================================
- Coverage   91.77%   91.73%   -0.05%     
==========================================
  Files         125      125              
  Lines       20437    20471      +34     
==========================================
+ Hits        18757    18779      +22     
- Misses       1680     1692      +12     

Files with missing lines	Coverage Δ
pymc/dims/distributions/scalar.py	96.85% <100.00%> (+0.24%)	⬆️

... and 2 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[read-the-docs-community]

Documentation build overview

📚 pymc | 🛠️ Build #32775369 | 📁 Comparing cb0bf6b against latest (88dc395)

  🔍 Preview build  

2 files changed

± glossary.html
± _modules/pymc/dims/distributions/scalar.html

[ricardoV94]

LGTM but I would make the alternative parameters keyword only

[ricardoV94]

thanks