Fix marimo progress bar issues

[ricardoV94]

Fixes display of nutpie progress bar in marimo, reported in https://discourse.pymc.io/t/is-progress-bar-in-marimo-broken-for-everyone-with-v6-0-0-or-its-just-me/17715

And another issue I found while debugging: speed/elapsed kept drifting fro finished bars (more obvious in sequential sampling, but also present in parallel)

Before:

Screencast.From.2026-05-16.15-18-33.mp4

After:

Screencast.From.2026-05-16.15-02-40.mp4

[read-the-docs-community]

Documentation build overview

📚 pymc | 🛠️ Build #32735807 | 📁 Comparing 7f1a194 against latest (a65e8b6)

  🔍 Preview build  

2 files changed

± glossary.html
± contributing/jupyter_style.html

[codecov]

Codecov Report

❌ Patch coverage is 40.90909% with 13 lines in your changes missing coverage. Please review.
✅ Project coverage is 91.73%. Comparing base (169e901) to head (7f1a194).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
pymc/progress_bar/marimo_progress.py	0.00%	9 Missing ⚠️
pymc/progress_bar/progress.py	69.23%	4 Missing ⚠️

❌ Your patch check has failed because the patch coverage (40.90%) is below the target coverage (50.00%). You can increase the patch coverage or adjust the target coverage.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #8298      +/-   ##
==========================================
- Coverage   91.78%   91.73%   -0.06%     
==========================================
  Files         125      125              
  Lines       20428    20448      +20     
==========================================
+ Hits        18749    18757       +8     
- Misses       1679     1691      +12     

Files with missing lines	Coverage Δ
pymc/progress_bar/progress.py	83.17% <69.23%> (-0.99%)	⬇️
pymc/progress_bar/marimo_progress.py	16.57% <0.00%> (-0.77%)	⬇️

🚀 New features to boost your workflow:

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

[ricardoV94]

Need to change the windows CI runners I think

[williambdean]

Seems reasonable. So is this also for the PyMC sampler as well? Or just nutpie where then it is safe to do the import.

[ricardoV94]

Seems reasonable. So is this also for the PyMC sampler as well? Or just nutpie where then it is safe to do the import.

the nutpie fix is just for nutpie because it calls it from a thread. I don't think we do this anywhere ourselves. if we start doing so we can port the code over