Improve label customization classes and remove WIP markers

[Quantum-Kayak]

This PR includes edits developed with assistance from a language model to ensure clarity, correctness, and documentation consistency. Specific contributions include:

• Refactored all labeller class and method docstrings using NumPy docstring style for consistency and clarity.
• Removed all """WIP.""" placeholder strings and corrected improper or duplicate docstring blocks.
• Fixed a logic issue in MapLabeller.model_name_to_str which incorrectly referenced var_name_map instead of model_name_map.
• Updated class-level docstrings for all labeller subclasses to clearly explain their unique behavior.
• Improved the docstring for mix_labellers with cleaner formatting, accurate parameter descriptions, and runnable examples.
• Ensured that all method overrides match the base method signatures and preserve expected output behavior.

These changes aim to make the labeller system more accessible and maintainable, especially for contributors working on customized label behavior in ArviZ plots.

---

📚 Documentation preview 📚: https://arviz--2456.org.readthedocs.build/en/2456/

[Quantum-Kayak]

Hey! Just a heads up that I'm still working on this — cleaning up the namespace exports and making sure all CI passes. Final fixes are coming soon.

[Quantum-Kayak]

Appreciate your patience — this is my first contribution under this username.

[Quantum-Kayak]

Hi @arviz-devs, just circling back on this PR. Everything should be in good shape now:

• Local + most CI tests pass ✅
• Namespace cleanup complete ✅
• Docstrings and structure reviewed ✅

A few failing CI checks seem unrelated to this PR (e.g., external integration tests and ReadTheDocs build).

Would love a review when someone gets a moment — happy to revise if anything's needed!

Thanks so much 🙌

[Quantum-Kayak]

Hi @aloctavodia @ColCarroll — PR #2456 is ready for review when you have a moment!
All core tests are passing and docs are updated. Appreciate your time 🙏

A few external test failures seem unrelated to this PR (integration/doc builds).
Let me know if there's anything else I can adjust!

[OriolAbril]

Very quick review pass (sorry but we are a bit swamped and it takes a while for us to get to the PRs, we we'll get to it). Please do not remove existing content, only extend and correct the existing one.

Improved the docstring for mix_labellers with cleaner formatting, accurate parameter descriptions, and runnable examples.

This seems like the opposite of what the PR does. The docstring had executable examples that the PR removes, parameter descriptions are unchanged except for the return value description which was moved to the notes section which I think doesn't align with numpydoc recommendations.

Ensured that all method overrides match the base method signatures and preserve expected output behavior.

I wasn't aware there was any inconsistency so it would be nice to get specific details on this.

---

Note that the "external tests" also check formatting and run linters, so failiures there might not be unrelated

[Quantum-Kayak]

Thanks for the feedback! I’ll make the changes and restore the original examples/doc structure as requested. I’ll update you when the new commit is pushed.

[OriolAbril]

The comments from the last review have not been addressed. The changes that decrease the quality of the mix_labellers function are a particularly important blocker as the whole point of the function (which is literally 1 line of code) is to document and explain how mixing labeller classes works.

[tomicapretto]

I'm confused by the authorship of this PR. Is this created by @Quantum-Kayak or @shubh2294?