FIX: improve CfRadial roundtrips and export after sweep mapping

[syedhamidali]

• Closes Bug in Either map_over_sweeps or in to_cfradial1 #254 cfradial1 to cfradial2 failed #322
• Tests added
• Changes are documented in history.md

Summary

This PR improves the CfRadial1/CfRadial2 transform and export paths in a few places where the current behavior was inconsistent or too fragile.

The main fixes are:

• make to_cfradial2() work on root-only datasets returned by xr.open_dataset(..., engine="cfradial1") by reopening the original source when available
• preserve the georeferencing_correction group during CfRadial transforms even when it is empty
• make CfRadial1 export robust after map_over_sweeps() workflows that use dataset-wide operations like where(), which can broadcast sweep metadata into 2D arrays

What changed

• xradar/transform/cfradial.py

  • to_cfradial2() now detects the normalized root-only dataset returned by engine="cfradial1"
  • if encoding["source"] is available, it reopens the original CfRadial1 file through open_cfradial1_datatree(...)
  • if not, it raises a clearer error
  • georeferencing_correction is now preserved as a structural group even when it contains no variables

• xradar/io/export/cfradial1.py

  • normalize sweep metadata variables such as sweep_number, sweep_fixed_angle, and sweep_mode back to scalar form before CfRadial1 export
  • this fixes failures after map_over_sweeps() operations that apply where() and broadcast metadata across (time/azimuth, range)

• tests/transform/test_cfradial.py

  • added regression tests for:
    • root-only engine="cfradial1" datasets passed into to_cfradial2()
    • missing encoding["source"] on those datasets
    • CfRadial1 export after map_over_sweeps(... where(...))

• docs/history.md

  • documented the fix in the development history

Motivation

This addresses two related issues:

1. to_cfradial2() behaved inconsistently depending on how the source dataset was opened
2. to_cfradial1() / io.to_cfradial1() could fail after map_over_sweeps() even though the transformed tree looked structurally valid

In particular, the exporter previously assumed sweep metadata stayed scalar, which is not true after some dataset-wide masking operations.

[codecov]

Codecov Report

❌ Patch coverage is 96.66667% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 93.94%. Comparing base (e8e905a) to head (47d9dcd).
⚠️ Report is 1 commits behind head on main.

Files with missing lines	Patch %	Lines
xradar/io/export/cfradial1.py	95.23%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #350      +/-   ##
==========================================
+ Coverage   93.85%   93.94%   +0.08%     
==========================================
  Files          28       28              
  Lines        6166     6191      +25     
==========================================
+ Hits         5787     5816      +29     
+ Misses        379      375       -4     

Flag	Coverage Δ
notebooktests	0.00% <0.00%> (ø)
unittests	93.94% <96.66%> (+0.08%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ 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.

[kmuehlbauer]

Thanks @syedhamidali, this seems reasonable to do. One comment here.

[kmuehlbauer]

@syedhamidali We can move on with this, please fix the merge conflicts and then we can get this in 👍

[kmuehlbauer]

Need to fix history.md here. Otherwise LGTM.