Add Hurdle distributions

[tomicapretto]

What is this PR about?

Implements four hurdle distributions. These are like zero-inflated distributions, but the process that generates the zeroes is independent of the process that generates the non-zero values (in zero-inflated distributions the zeros can come from the two distinct processes).

These distributions are added in distributions/mixture.py. To do so, I had to move a couple of things around

• DiracDelta is part of distributions/distribution.py
• Zero-inflated distributions are now part of distributions/mixture.py
• These new distributions go into distributions/mixture.py as well
• Moved tests for DiracDelta and zero-inflated distributions to their respective places

There's a hurdle.ipynb notebook. My goal is not to merge it into the repository (it's in a very bad place). But I think it can be useful if you want to see how these distributions work.

My end goal is to add hurdle families to Bambi and I think it would be better to have these distributions implemented and tested in PyMC rather than as something that only lives within Bambi.

Checklist

• Explain important implementation details 👆
• Make sure that the pre-commit linting/style checks pass.
• Link relevant issues (preferably in nice commit messages)
• Are the changes covered by tests and docstrings?
• Fill out the short summary sections 👇

New features

• Added hurdle distributions (HurdlePoisson, HurdleNegativeBinomial, HurdleGamma, and HurdleLogNormal).

Maintenance

• Move DiracDelta to distributions/distribution.py
• Move zero-inflated families to distributions/mixture.py
• Remove pm.Constant

Edit: @ricardoV94

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[ricardoV94]

Looks good functionality-wise.

In terms of organization, since there are two things going on, can I ask you to clean the git history and separate the move stuff around and implement hurdle models into two distinct commits?

[ricardoV94]

Good time as any to remove it

[tomicapretto]

Done :)

[codecov]

Codecov Report

Merging #6688 (be27167) into main (fa82fef) will increase coverage by 2.49%.
The diff coverage is 93.61%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #6688      +/-   ##
==========================================
+ Coverage   89.50%   91.99%   +2.49%     
==========================================
  Files          94       95       +1     
  Lines       15942    16162     +220     
==========================================
+ Hits        14269    14869     +600     
+ Misses       1673     1293     -380     

Impacted Files	Coverage Δ
pymc/distributions/discrete.py	99.11% <ø> (+0.37%)	⬆️
pymc/distributions/mixture.py	94.16% <91.17%> (-1.24%)	⬇️
pymc/distributions/__init__.py	100.00% <100.00%> (ø)
pymc/distributions/distribution.py	96.63% <100.00%> (+0.33%)	⬆️
pymc/logprob/basic.py	99.15% <100.00%> (ø)

... and 13 files with indirect coverage changes

[ricardoV94]

We should remove the notebooks (right?)

There seem to be some conflict lines <<<< in the intermediate commits. It probably makes things easier for you to 1) remove Constant, 2) Move things around, 3) Implement the new stuff, in that order.

Let me know if you want me to do it instead.

Everything else 👌

[tomicapretto]

There seem to be some conflict lines <<<< in the intermediate commits

I thought they were completely removed. I expected them to be caught by the linting tools. It seems I was wrong.

Let me know if you want me to do it instead.

I would appreciate it a lot :D

[ricardoV94]

And its correct to remove the notebooks?

[tomicapretto]

And its correct to remove the notebooks?

Yes, my bad. It was something I created to see how it worked. But it's not needed.

[welcome]

Congrats on merging your first pull request! 🎉 We here at PyMC are proud of you! 💖 Thank you so much for your contribution 🎁