Add experimental XarraySource

[SanketMeghale]

Summary

This draft PR adds an experimental native XarraySource for xarray-backed datasets.

Current MVP scope:

• add lumen.sources.xarray.XarraySource
• expose Dataset.data_vars as logical tables
• support coordinate-aware filtering for 1D coordinates
• flatten query results to pandas DataFrames for downstream compatibility
• add docs and example config/demo for NetCDF/Zarr-backed datasets
• add focused test coverage for schema, filtering, metadata, URI loading, and example rendering

This is intended for early design feedback rather than final review.

later improvements

• aggregation in the source API
• SQL interoperability beyond DataFrame conversion
• multi-file dataset support

Validation

• python -m pytest lumen/tests/sources/test_xarray.py -q

Refs #1748

[codecov]

Codecov Report

❌ Patch coverage is 92.78937% with 38 lines in your changes missing coverage. Please review.
✅ Project coverage is 67.94%. Comparing base (4b4ccb2) to head (3e3a312).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
lumen/sources/xarray.py	89.58%	30 Missing ⚠️
lumen/tests/sources/test_xarray.py	97.81%	5 Missing ⚠️
lumen/sources/__init__.py	70.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1751      +/-   ##
==========================================
+ Coverage   67.43%   67.94%   +0.51%     
==========================================
  Files         168      170       +2     
  Lines       27947    28498     +551     
==========================================
+ Hits        18845    19364     +519     
- Misses       9102     9134      +32     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[ahuang11]

I think to make this work with Lumen (AI part), we'll probably need to base it off BaseSQLSource and https://github.com/alxmrs/xarray-sql, so that AI can write the transforms.

Separately, I think there's a lot of AI things that are undesired, e.g.

    This first draft is intentionally conservative:
    - accepts an in-memory xarray.Dataset
    - exposes Dataset.data_vars as logical tables
    - supports coordinate-based filtering via .sel(...)
    - returns pandas DataFrames for downstream compatibility

This should be part of your PR description, not docstring.

[SanketMeghale]

Thanks for the detailed feedback!

That makes sense, aligning this with BaseSQLSource and xarray-sql would definitely make it more flexible and allow AI-generated transforms in a cleaner way. I’ll explore restructuring the implementation in that direction.

Also noted on keeping the AI behavior minimal and predictable , I’ll avoid adding unnecessary complexity there.

And thanks for pointing out the docstring vs PR description , I’ll move that explanation to the PR description and keep the docstrings focused on code-level details.

Appreciate the guidance!

[ahuang11]

Superseded by #1741