GEOPY-2866: Research auto-scaling strategy for cross-gradient regularization terms

[domfournier]

GEOPY-2866 - Research auto-scaling strategy for cross-gradient regularization terms

[Copilot]

Pull request overview

Adds an optional “iterative rescaling” mechanism for joint cross-gradient inversions, wiring it through options/UI and adjusting expected run-test targets to reflect the updated regularization behavior.

Changes:

• Add iterative_rescaling parameter to JointCrossGradientOptions (default enabled) and expose it in the UI schema.
• Modify cross-gradient multiplier scaling logic and (when enabled) add a SimPEG directive to rescale cross-gradient terms during iterations.
• Update joint cross-gradient run-test target metrics to match the new inversion behavior.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 3 comments.

File	Description
tests/run_tests/driver_joint_cross_gradient_test.py	Updates expected inversion metrics (targets) for the joint cross-gradient run test.
simpeg_drivers/joint/joint_cross_gradient/options.py	Introduces the iterative_rescaling option (and documents it).
simpeg_drivers/joint/joint_cross_gradient/driver.py	Adjusts cross-gradient scaling and appends an iterative scaling directive when enabled.
simpeg_drivers-assets/uijson/joint_cross_gradient_inversion.ui.json	Exposes iterative_rescaling as a UI-configurable parameter.

Comments suppressed due to low confidence (1)

simpeg_drivers/joint/joint_cross_gradient/options.py:51

• iterative_rescaling defaults to True, which changes the cross-gradient scaling behavior for existing projects/options files that don't specify this new field: previously the multiplier always included a base_length**4 dimensional scaling, but with the new default it will not. If this is intended, consider calling it out explicitly in release notes; otherwise set the default to preserve legacy behavior (e.g., default False and let the UI enable it explicitly).

    cross_gradient_weight_a_b: float = 1.0
    cross_gradient_weight_c_a: float | None = None
    cross_gradient_weight_c_b: float | None = None

    iterative_rescaling: bool = True
    models: JointModelOptions = JointModelOptions()

---

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

[codecov]

Codecov Report

❌ Patch coverage is 76.92308% with 3 lines in your changes missing coverage. Please review.
✅ Project coverage is 90.40%. Comparing base (0edfe9d) to head (0078ea1).

Files with missing lines	Patch %	Lines
...impeg_drivers/joint/joint_cross_gradient/driver.py	75.00%	1 Missing and 2 partials ⚠️

Additional details and impacted files

@@                Coverage Diff                 @@
##           release/GA_4.8     #411      +/-   ##
==================================================
- Coverage           90.43%   90.40%   -0.04%     
==================================================
  Files                 110      110              
  Lines                6532     6542      +10     
  Branches              826      830       +4     
==================================================
+ Hits                 5907     5914       +7     
- Misses                413      414       +1     
- Partials              212      214       +2     

Files with missing lines	Coverage Δ
...mpeg_drivers/joint/joint_cross_gradient/options.py	100.00% <100.00%> (ø)
...impeg_drivers/joint/joint_cross_gradient/driver.py	88.23% <75.00%> (-4.63%)	⬇️

🚀 New features to boost your workflow:

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