ENH: add xarray-native open_datatree with engine= parameter

[aladinor]

Summary

Closes #329

Note: This is a draft. Full implementation depends on #333 (station coords + optional_groups) and #332 (NEXRAD chunk support) being merged first. After those land, this branch will be rebased to incorporate:

• optional_groups parameter flow-through in open_groups_as_dict / open_datatree
• Station coordinates as root coordinates (propagates automatically via _assign_root)
• NEXRAD chunk list/tuple input support in open_groups_as_dict

Adds open_datatree() and open_groups_as_dict() methods with supports_groups = True to three prototype backends (ODIM, CfRadial1, NEXRAD Level2), enabling:

import xradar as xd
import xarray as xr

# New unified xradar API
dtree = xd.open_datatree("file.h5", engine="odim")

# xarray native API
dtree = xr.open_datatree("file.h5", engine="odim")

# Legacy (emits FutureWarning)
dtree = xd.io.open_odim_datatree("file.h5")

Changes

• xradar/io/backends/common.py: Add shared _build_groups_dict() helper and _deprecation_warning() utility
• xradar/io/backends/odim.py: Add supports_groups=True, open_groups_as_dict(), open_datatree(); refactor open_odim_datatree() to delegate with deprecation warning
• xradar/io/backends/cfradial1.py: Same pattern (unique single-file slicing logic)
• xradar/io/backends/nexrad_level2.py: Same pattern (store-based with sweep resolution)
• xradar/io/__init__.py: Add engine registry and unified open_datatree(file, engine=) function
• xradar/__init__.py: Export open_datatree at package level
• requirements.txt: Exclude broken netCDF4 1.7.3/1.7.4
• tests/io/test_backend_datatree.py: 24 new tests (integration, deprecation, equivalence)
• examples/notebooks/Open-Datatree-Engine.ipynb: Demo notebook

Depends on

• ENH: move station coords to root and add optional_groups parameter #333 — station coords to root + optional_groups parameter (will require rebase)
• ADD: Support reading NEXRAD Level 2 chunks from real-time S3 bucket #332 — NEXRAD chunk list input (will require adding chunk support to open_groups_as_dict)

Test plan

• 24/24 new tests pass (pytest tests/io/test_backend_datatree.py)
• 381/381 existing passing tests still pass (0 regressions)
• ruff check clean
• black --check clean
• xd.open_datatree() works for all 3 engines
• xr.open_datatree() works for all 3 engines
• Deprecated functions emit FutureWarning and still produce correct results
• Rebase after ENH: move station coords to root and add optional_groups parameter #333 and ADD: Support reading NEXRAD Level 2 chunks from real-time S3 bucket #332 merge

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov]

Codecov Report

❌ Patch coverage is 98.00000% with 2 lines in your changes missing coverage. Please review.
✅ Project coverage is 93.57%. Comparing base (e59d9e6) to head (a9c9519).

Files with missing lines	Patch %	Lines
xradar/io/backends/nexrad_level2.py	94.73%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #335      +/-   ##
==========================================
+ Coverage   93.52%   93.57%   +0.04%     
==========================================
  Files          27       27              
  Lines        5717     5755      +38     
==========================================
+ Hits         5347     5385      +38     
  Misses        370      370              

Flag	Coverage Δ
notebooktests	?
unittests	93.57% <98.00%> (+0.04%)	⬆️

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.