Mag L2 rotation - propogate FILLVALS

[maxinelasp]

Change Summary

This moves the change from #2901 into MAG-specific code.

[Copilot]

Pull request overview

This PR updates MAG L2/L1D frame rotation to preserve missing-data sentinels (NaN/FILLVAL) during rotations, preventing “nearly-but-not-quite” fill values from being introduced into rotated science vectors.

Changes:

• Apply a post-rotation mask in MAG L2 and L1D rotate_frame() to propagate NaN/FILLVAL from the pre-rotation vectors into the rotated output.
• Add unit tests for both L2 and L1D to verify FILLVAL/NaN preservation behavior during rotation.
• Adjust the MAG L2 test fixture vector dtype to floating-point to support NaN/FILLVAL handling.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
imap_processing/mag/l2/mag_l2_data.py	Propagates NaN/FILLVAL into rotated vectors after frame_transform() in L2 rotation.
imap_processing/mag/l1d/mag_l1d_data.py	Propagates NaN/FILLVAL into rotated MAGO/MAGI vectors after frame_transform() in L1D rotation.
imap_processing/tests/mag/test_mag_l2.py	Adds regression test ensuring L2 rotate_frame() preserves NaN/FILLVAL positions.
imap_processing/tests/mag/test_mag_l1d.py	Adds regression test ensuring L1D rotate_frame() preserves NaN/FILLVAL positions for both MAGO and MAGI vectors.
imap_processing/tests/mag/conftest.py	Updates the L2 norm_dataset vectors to floating dtype (but comment/dtype intent needs alignment).

---

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

[Copilot]

Same as above: this does redundant full-array scans and recomputes the isnan|==FILLVAL mask twice. Computing the mask once and reusing it (and potentially extracting a small helper to apply the mask for both MAGO and MAGI arrays) would reduce work on large datasets and avoid duplicated logic.

[subagonsouth]

A couple of thoughts to consider.

[subagonsouth]

Good to go from my perspective.