Extract endianness from Bytes codec in V2 metadata conversion

[maxrjones]

This PR complements #766 by adding to_kerchunk support (whereas #766 added to_icechunk support). For a more robust long-term solution that would eliminate the need for separate metadata pipelines between kerchunk and icechunk, see issue #768.

@rsignell I tested your workflow works using this PR (see https://github.com/maxrjones/test-scripts/blob/main/test-virtualizarr-netcdf3/test-virtualizarr.py).

• Closes NetCDF3Parser workflow returning gibberish (while equivalent HDFParser workflow works fine) #750
• Tests added
• Tests passing
• Full type hint coverage
• Changes are documented in docs/releases.rst
• New functions/methods are listed in api.rst
• New functionality has documentation

[rsignell]

Yep, this works for my use case!

[TomNicholas]

Thank you! Let's just make sure this important bugfix is in the changelog.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 87.75%. Comparing base (9261537) to head (e52f226).

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #769      +/-   ##
==========================================
+ Coverage   87.71%   87.75%   +0.03%     
==========================================
  Files          35       35              
  Lines        1872     1878       +6     
==========================================
+ Hits         1642     1648       +6     
  Misses        230      230              

Files with missing lines	Coverage Δ
virtualizarr/utils.py	70.76% <100.00%> (+2.97%)	⬆️

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.