Fix nbytes test

[JamesMcClung]

The nbytes check was passing because it was making sure the decoded dataset had the same number of bytes as... itself. Turns out the decoded and original datasets don't actually have the same number of bytes, but their data does, so that's what it checks now.

Also, use pytest fixtures everywhere.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

🚀 New features to boost your workflow:

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

[Copilot]

Pull Request Overview

This PR fixes the nbytes test by comparing the summed bytes of dataset data variables between raw and decoded datasets, and it converts tests to leverage pytest fixtures.

• Replace direct dataset opening with pytest fixtures
• Introduce _get_nbytes helper function for comparing dataset data sizes
• Update several tests to use new fixtures

Comments suppressed due to low confidence (2)

tests/test_xarray_adios2.py:120

• Consider adding a docstring for _get_nbytes to explain why summing data_vars is used instead of relying on the built-in nbytes property.

def _get_nbytes(ds: xr.Dataset) -> int:

tests/test_xarray_adios2.py:120

• [nitpick] Consider renaming _get_nbytes to something like get_dataset_data_nbytes to clarify that it sums only the bytes of data_vars, avoiding potential confusion with the dataset's nbytes attribute.

def _get_nbytes(ds: xr.Dataset) -> int:

[germasch]

👍 -- Thanks!

I'm thinking to use Copilot PR review regularly, not sure it would have caught my original bug -- it probably would have.