Resolve mypy error in mcmc.py

[Ok3ks]

Description

Switching from list comprehension to np.fromiter converts the list to an NDArray which satisfies Mypy Type checks

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--7854.org.readthedocs.build/en/7854/

[welcome]

]
💖 Thanks for opening this pull request! 💖 The PyMC community really appreciates your time and effort to contribute to the project. Please make sure you have read our Contributing Guidelines and filled in our pull request template to the best of your ability.

[ricardoV94]

Still failing: https://github.com/pymc-devs/pymc/actions/runs/16338464408/job/46163429561?pr=7854#step:5:370

You can test locally by running python scripts/run_mypy.py --verbose

[Ok3ks]

I did and I was focused on only this pymc/sampling/mcmc.py. Is the PR to clear all mypy warnings that occur from running this syntax?

python scripts/run_mypy.py --verbose

[ricardoV94]

I did and I was focused on only this pymc/sampling/mcmc.py. Is the PR to clear all mypy warnings that occur from running this syntax?

python scripts/run_mypy.py --verbose

No, just mcmc, but as you can see it still fails on the CI. Does it pass locally for you? If so, you may not be installing the same dependencies (specially numpy)

[Ok3ks]

I've also noticed that the tests fail on the ubuntu runner. I'm unable to reproduce locally, as I use macOS

[ricardoV94]

I doubt that's the reason for the difference. Type hints are not OS-specific

[Ok3ks]

my environment has numpy 2.3.1, and I noticed requirements.txt has numpy >=1.25.0.

Is there a specific numpy version I should use?

[ricardoV94]

You can check the versions the CI is installing in the log. It seems to be picking numpy 2.2.6

https://github.com/pymc-devs/pymc/actions/runs/16338464408/job/46163429561?pr=7854#step:4:22

[Ok3ks]

I've been able to reproduce locally with numpy 2.2.6, is bumping the numpy version to 2.3.2 an option ( i understand, there's possible more to consider)

https://numpy.org/doc/stable/release/2.3.0-notes.html

[ricardoV94]

I've been able to reproduce locally with numpy 2.2.6, is bumping the numpy version to 2.3.2 an option ( i understand, there's possible more to consider)

https://numpy.org/doc/stable/release/2.3.0-notes.html

We can just wait for the CI to start picking that version of numpy and the issue will auto-fix itself then? Or do we still need this change?

[Ok3ks]

I've been able to reproduce locally with numpy 2.2.6, is bumping the numpy version to 2.3.2 an option ( i understand, there's possible more to consider)
numpy.org/doc/stable/release/2.3.0-notes.html

We can just wait for the CI to start picking that version of numpy and the issue will auto-fix itself then? Or do we still need this change?

I just checked. Yes, the change fixes the mypy error on numpy 2.3.1.

Without the change, the mypy error persists even on numpy 2.3.1

[ricardoV94]

In that case we can merge the changes but leave the mcmc.py as still failing

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.99%. Comparing base (8a436d8) to head (8838710).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #7854      +/-   ##
==========================================
+ Coverage   89.25%   92.99%   +3.73%     
==========================================
  Files         108      108              
  Lines       18327    18327              
==========================================
+ Hits        16358    17043     +685     
+ Misses       1969     1284     -685     

Files with missing lines	Coverage Δ
pymc/sampling/mcmc.py	91.37% <100.00%> (ø)

... and 22 files with indirect coverage changes

🚀 New features to boost your workflow:

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

[Ok3ks]

In that case we can merge the changes but leave the mcmc.py as still failing

alright,

In that case we can merge the changes but leave the mcmc.py as still failing

I have readded "sampling/mcmc.py" to the list of failing files, now waiting on CI to pass