ENH: <Automatic Sampling Diagnostics Summary Utility 'ENH: ' prefix>

[tanzeelgcuf]

Before

# Current workflow (manual & repetitive)

import arviz as az

idata = pm.sample(...)

az.summary(idata)
az.plot_trace(idata)
az.plot_energy(idata)

# Manually inspect:
# - R-hat values
# - ESS
# - Divergences
# - Tree depth saturation
# Then interpret results and decide next steps

After

pm.diagnostics.summary(
    trace=idata,
    model=model,
    warn=True
)

Context for the issue:

Interpreting sampling diagnostics is one of the biggest pain points for new PyMC users and a repetitive task for advanced users. While ArviZ already provides excellent low-level tools, PyMC currently lacks a high-level, opinionated diagnostics summary that answers the question:

“Is my sampling result healthy, and what should I do next?”

This feature would:

Improve onboarding and usability

Encourage best practices

Reduce misinterpretation of posterior samples

Build on existing ArviZ diagnostics without introducing new algorithms

Avoid breaking changes

The proposal intentionally focuses on aggregation, interpretation, and presentation, keeping the scope small and maintainable.

I’m happy to discuss API design, implementation location, and testing strategy based on maintainer feedback.

[welcome]

]
🎉 Welcome to PyMC! 🎉 We're really excited to have your input into the project! 💖

If you haven't done so already, please make sure you check out our Contributing Guidelines and Code of Conduct.

[aloctavodia]

Hi @tanzeelgcuf, thanks for the interest. I think this https://github.com/arviz-devs/arviz-plots is a better place to discuss this.