Fix hi l1a tests broken by upgrade to cdflib 1.3.9

[tmplummer]

Change Summary

Overview

Hi L1A processing was broken by upgrading to cdflib 1.3.9. The problem was actually that Hi L1A code was not passing check_schema=False to the CDF attributes manager for epoch and other coordinate variables. This PR fixes that.

Testing: No testing changes because existing tests failed once I updated my cdflib.

Closes: #3006

[Copilot]

Pull request overview

Fixes IMAP-Hi L1A dataset finalization to be compatible with cdflib 1.3.9 by bypassing schema validation when fetching CDF metadata for certain coordinate variables, preventing attribute/schema enforcement issues during CDF writing.

Changes:

• Pass check_schema=False when retrieving CDF variable attributes for epoch in histogram and memdump dataset finalizers.
• Pass check_schema=False when retrieving CDF variable attributes for the histogram angle coordinate.

Comments suppressed due to low confidence (1)

imap_processing/hi/hi_l1a.py:415

• For the angle coordinate, hi_hist_angle attributes in imap_hi_variable_attrs.yaml include a dtype (uint16). Currently that dtype stays in attrs (so it may be written out as a CDF attribute) and np.arange(2, 360, 4) will default to int64. Consider copying the attrs, popping dtype, and explicitly creating/casting the coordinate data to that dtype before assigning attrs, to keep metadata clean and ensure the coordinate type matches the spec.

    attrs = attr_mgr.get_variable_attributes("hi_hist_angle", check_schema=False)
    dataset.coords.update(
        {
            "angle": xr.DataArray(
                np.arange(2, 360, 4),

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.