REF: Migrate to ArviZ 1.0 (InferenceData -> DataTree)

[AAlexxis222]

Migrates pymc-extras to support ArviZ 1.0, replacing InferenceData with xarray.DataTree.

Changes:

• Imports: Replaced arviz.InferenceData with xarray.DataTree.
• Attributes: Refactored .posterior access to dictionary style (["posterior"]).
• API: Updated constructors (DataTree.from_dict) and replaced .extend() with .update().

Verification:

• ~325 tests passing (verified against PyMC PR #8019 rebase).
• Validated prior and pathfinder modules.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[aloctavodia]

In many places you are apllying formating issues. Are you sure these are needed?

[AAlexxis222]

Hi! Yes, I applied the formatting tools (black, isort, ruff) because the pre-commit CI job was failing with linting errors on the modified files.

I wanted to ensure the CI was green. However, if you prefer to keep the diff minimal, I can try to revert the style changes and only keep the logic migration (though pre-commit might fail again)

[aloctavodia]

That's fine, just checking.

[Copilot]

Pull request overview

This pull request migrates pymc-extras to support ArviZ 1.0 by replacing InferenceData with xarray.DataTree. The migration updates imports, changes attribute access patterns from dot notation to dictionary-style access, and updates API methods.

Changes:

• Replaced arviz.InferenceData imports with xarray.DataTree and updated arviz imports to arviz_base where needed
• Changed attribute access from .posterior to ["posterior"] throughout the codebase
• Replaced .extend() method calls with .update() and .add_groups() with direct dictionary assignment
• Updated dict_to_dataset parameter names from library to inference_library and default_dims to sample_dims
• Applied code formatting improvements (multi-line formatting, import cleanup)

Reviewed changes

Copilot reviewed 85 out of 87 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
pymc_extras/model/marginal/marginal_model.py	Updated imports and DataTree usage, but has parameter naming issue
pymc_extras/utils/prior.py	Updated type annotations and DataTree access patterns
pymc_extras/inference/laplace_approx/*.py	Updated InferenceData to DataTree, dict_to_dataset parameters
pymc_extras/inference/pathfinder/*.py	Updated DataTree usage and API calls
pymc_extras/inference/dadvi/dadvi.py	Updated DataTree construction
pymc_extras/statespace/core/statespace.py	Updated type annotations and DataTree access
pymc_extras/model_builder.py	Updated .extend() to .update() and .add_groups() to dict assignment
tests/*	Updated test fixtures to create DataTree objects and use dict-style access
Various formatting changes	Multi-line formatting improvements throughout

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[jessegrabowski]

I wanted to ensure the CI was green. However, if you prefer to keep the diff minimal, I can try to revert the style changes and only keep the logic migration (though pre-commit might fail again)

The CI needs to be green, but this PR shouldn't touch 87 files. You should be running the pre-commit hooks we have defined locally that should be sufficient.