Adding functionality to represent non linear spectral dispersion by A-Derks · Pull Request #713 · DKISTDC/dkist

A-Derks · 2026-04-30T19:21:04Z

Background:
DKIST instruments (Cryo SP, DL-NIRSP, ViSP) produce FITS headers with CTYPE values of AWAV-GRA, indicating a non-linear pixel-to-wavelength transform based on the grating/grism equation described in Greisen et al. (2006, FITS WCS Paper 3, sections 5.1.2 and 5.2). The PV{n}_m keywords encode the physical grating parameters.

The gwcs library (via gwcs#745) provides two models for evaluating the FITS grism transform: RefractedAngleSineModel, which computes the pixel-dependent refracted-angle sine from the detector geometry and grating parameters, and WavelengthFromGrismEquation, which converts the incident-angle sine and refracted-angle sine into wavelength using the FITS grism formalism. WavelengthFromGratingEquation remains the classic two-input grating-equation model and is not used by this PR. This PR provides the builder function that wires the FITS grism models together using the parameters extracted from the FITS header.

What changed:

dkist/wcs/models.py
- build_grating_spectral_transform — the core implementation. Given the full set of grating/grism parameters, it:
  1. Computes the constant sin(alpha_in) directly from incident_angle and wraps it in a Const1D
  2. Constructs a RefractedAngleSineModel (from gwcs.spectroscopy) with the detector geometry and grating parameters baked in
  3. Constructs a WavelengthFromGrismEquation (from gwcs.spectroscopy) with the grism parameters
  4. Assembles the compound model: Mapping((0, 0)) | (Const1D(alpha_in) & RefractedAngleSineModel(...)) | WavelengthFromGratingEquation(...)

All four optional grism parameters (refractive_index, refractive_index_derivative, out_of_plane_angle, camera_angle) default to their identity values, so the transform degrades to a standard grating equation when they are absent from the header.

RefractedAngleSineModel and WavelengthFromGrismEquation are defined in gwcs.spectroscopy (added via gwcs#745), which also provides the ASDF converter that serializes them. This PR imports both directly from gwcs rather than defining or serializing a local model in dkist.

Notes:

The CI test failures are due to the gwcs dependency: RefractedAngleSineModel and WavelengthFromGratingEquation do not exist in any released gwcs version yet. Tests will pass once gwcs#745 merges and the dependency is pinned or updated.
The CodSpeed regression (-6.9% on test_plot_dataset) is unrelated to the spectral transform code — it affects a plotting benchmark and is likely noise or a pre-existing issue.

eigenbrot · 2026-05-04T15:57:18Z

 ]


+def _grating_equation_constants(


Type-hints, please

eigenbrot · 2026-05-04T16:13:52Z

+    )
+    alpha_in = m.Const1D(amplitude=adjusted_incident_angle_sine)
+    wavelength_from_grating = WavelengthFromGratingEquation(
+        groove_density=wavelength_parameter / spectral_order,


If I'm reading the math right, I think groove_density = wavelength_parameter / spectral_order only in the case where refractive_index_derivative = 0 and out_of_plan_angle = 0. Let's just set groove_density to the input grating_density, which is the same thing.

But here's the part I don't get. If groove density is just G, then where is the denominator in Greisen eq 70? I'm looking at step 3 on page 756: "Compute the wavelength \lambda from \Gamma via Eq. 70" That equation has the two numerator parts, an adjusted sin \alpha and a sin \gamma, which we pass in the return of this function. The denominator in eq 70 is what is called wavelength_parameter above. Are we supposed to scale the sin(angles) before passing them to the model?

eigenbrot · 2026-05-04T17:03:05Z

+    adjusted_incident_angle_sine = (
+        refractive_index - (refractive_index_derivative * reference_wavelength).decompose()
+    ) * np.sin(alpha)
+    wavelength_parameter = grism_parameter - refractive_index_derivative * np.sin(alpha)


Why is this variable called wavelength_parameter? When reading through these functions this name caused me a lot of confusion because I thought it had something to do with reference_wavelength or just wavelength in general. The units are 1 / m, which doesn't feel like a wavelength to me.

Based on my comment down in generate_grating_spectral_transform I'd argue we don't actually need this variable outside of this function. Given that, we could sidestep any naming issue by just removing this variable altogether and writing out angle_slope with the other variables directly.

I agree with Arthur, but if you do want to name these parts of the grism equation (Greisen eq 70), then I would stick with names that imply units. You already do that with adjusted_incident_angle_sine, so you could name this part with something that includes ...inverse_wavelength.

eigenbrot · 2026-05-04T17:08:33Z

+        adjusted_incident_angle_sine,
+        wavelength_parameter,
+        reference_refracted_angle,
+        theta,


I don't think we need to return theta from this function. It's just a variable rename of the input parameter camera_angle, which (I think) just muddies the path of variables in generate_grating_spectral_transform. I.e., it makes it seem like theta is meaningfully different from camera_angle, which it isn't.

True, a .to(u.rad) has happened, but the return value is still a Quantity and so, e.g., np.sin will work correctly in any case.

@eigenbrot, see my comment down on line 84.

eigenbrot · 2026-05-04T19:45:18Z

+    (
+        adjusted_incident_angle_sine,
+        wavelength_parameter,
+        reference_refracted_angle,
+        theta,
+        angle_slope,
+    ) = _grating_equation_constants(
+        reference_wavelength=reference_wavelength,
+        grating_density=grating_density,
+        spectral_order=spectral_order,
+        incident_angle=incident_angle,
+        refractive_index=refractive_index,
+        refractive_index_derivative=refractive_index_derivative,
+        out_of_plane_angle=out_of_plane_angle,
+        camera_angle=camera_angle,
+    )


It appears this function (_grating_equation_constants) exists to save a little bit of code repetition when computing some of the inputs to _refracted_angle_sine_model, alpha_in, and WavelengthFromGratingEquation, but, in doing so, it separates the inputs to generate_grating_spectral_transform from their actual usage and adds another layer of variable names and calculations. I think this added layer causes confusion (see my comments above) and obfuscates the flow of information through generate_grating_spectral_transform.

My suggestion would be to remove _grating_equation_constants entirely and rewrite _refracted_angle_sine_model to accept grating_density, spectral_order, refractive_index, and reference_wavelength. adjusted_incident_angle_sine could be computed directly in the function body.

eigenbrot · 2026-05-04T20:14:32Z

+    ) * np.sin(alpha)
+    wavelength_parameter = grism_parameter - refractive_index_derivative * np.sin(alpha)
+    reference_refracted_angle = np.arcsin(
+        (grism_parameter * reference_wavelength).decompose() - refractive_index * np.sin(alpha)


Is the .decompose() call necessary here? It looks like the results of this calculation stay in Quantity land for the duration of their life.

eigenbrot · 2026-05-04T20:24:42Z

+    epsilon = out_of_plane_angle.to(u.rad)
+    theta = camera_angle.to(u.rad)
+
+    grism_parameter = (grating_density * spectral_order) / np.cos(epsilon)


Not to be That Guy, but Greisen already defines something called the "grism parameter" in Eq. 75, which is different than the definition here.

"grism constant" would be an OK alternative

eigenbrot · 2026-05-04T20:38:51Z

+    )
+
+    pixels = np.array([0, 100, 217, 300, 511], dtype=float)
+    expected = WCS(header).all_pix2world(pixels, 0)[0] * u.m


As mentioned on the related dkist-inventory PR, a disadvantage of all_pix2world is that the results are unit-less and require us to apply the assumed-correct unit to its output. (It also has this weird signature and output that require non-obvious 0 and [0]).

Instead, I’d suggest using WCS(header).spectral.pixel_to_world(pixels), which will have proper units based on the CUNIT key. Then expected.to_value(u.nm) will work directly without needing to assume anything in the meantime.

Cadair

Sorry to keep bouncing parts of this code between here and dkist-inventory!

The thing this PR is missing is the asdf converter, schema and corresponding entries in the manifest, tests etc.

This is a thing, so this should be the last thing to do, but it will be needed to successfully test a roundtrip to ASDF file for this (which there should be a test of in dkist-inventory). We do have a helper script for this process which is here.

You should (for this PR) be able to run it like this:

$ pipx run tools/update_schema.py --manifest dkist-wcs refracted-angle-sine-model

~~However, it seems like this is bugged, so we'll have to fix it first.~~ It's now fixed enough it runs, it should get you started.

Cadair · 2026-05-05T09:07:28Z

+def _refracted_angle_sine_model(
+        reference_pixel, dispersion, reference_refracted_angle, theta, angle_slope
+):
+    @custom_model


Can you refactor this into class form, I think that with it like this it will be hard to add the required asdf schemas.

Cadair · 2026-05-05T09:10:02Z

    assert u.allclose(shift1.offset, 0)


+def test_generate_grating_spectral_transform():


Can you also add tests for just the new model in isolation? I also suspect this test should probably be in dkist-inventory and not here.

Added in there

krislarsnso · 2026-05-05T16:34:28Z

+    epsilon = out_of_plane_angle.to(u.rad)
+    theta = camera_angle.to(u.rad)
+
+    grism_parameter = (grating_density * spectral_order) / np.cos(epsilon)


"grism constant" would be an OK alternative

krislarsnso · 2026-05-05T16:36:51Z

+    adjusted_incident_angle_sine = (
+        refractive_index - (refractive_index_derivative * reference_wavelength).decompose()
+    ) * np.sin(alpha)
+    wavelength_parameter = grism_parameter - refractive_index_derivative * np.sin(alpha)


I agree with Arthur, but if you do want to name these parts of the grism equation (Greisen eq 70), then I would stick with names that imply units. You already do that with adjusted_incident_angle_sine, so you could name this part with something that includes ...inverse_wavelength.

krislarsnso · 2026-05-05T16:50:17Z

+    reference_refracted_angle = np.arcsin(
+        (grism_parameter * reference_wavelength).decompose() - refractive_index * np.sin(alpha)
+    )
+    angle_slope = wavelength_parameter / (


I would like to rename this. If we use the name "grism parameter" for Γ (i.e. rename Gm/cosϵ) then this is the slope of the grism parameter wrt wavelength and wrt the distance along the detector (ξ in the paper) scaled. Γ itself is dimensionless. I think grism_parameter_slope would be better. Or maybe grism_parameter_per_wavelength, which would make its use in _refracted_angle_sine_model more clear.

krislarsnso · 2026-05-05T16:56:52Z

+):
+    alpha = incident_angle.to(u.rad)
+    epsilon = out_of_plane_angle.to(u.rad)
+    theta = camera_angle.to(u.rad)


If these are already Quantities, then I don't think we need to convert them to radians. In this function, we only ever do trig on them, not use them directly in algebra, so the conversion isn't necessary.

krislarsnso · 2026-05-05T17:20:40Z

+        adjusted_incident_angle_sine,
+        wavelength_parameter,
+        reference_refracted_angle,
+        theta,


@eigenbrot, see my comment down on line 84.

krislarsnso · 2026-05-05T17:22:04Z

+        output_angle = (
+            np.arctan(-np.tan(theta) + wavelength_offset * angle_slope)
+            + reference_refracted_angle
+            + theta


Here is where theta aka camera_angle does need to be in radians. If it isn't passed through, the conversion should happen here.

krislarsnso · 2026-05-05T17:24:40Z

+    def refracted_angle_sine(pixel):
+        wavelength_offset = ((pixel - reference_pixel) * u.pix) * dispersion
+        output_angle = (
+            np.arctan(-np.tan(theta) + wavelength_offset * angle_slope)


Renaming angle_slope to something like grism_parameter_per_wavelength (see comment here) then makes this line clearer about what it is doing.

krislarsnso · 2026-05-05T19:46:11Z

+        name="Spectral",
+    )
+
+    return m.Mapping((0, 0)) | (alpha_in & refracted_angle_sine) | wavelength_from_grating


+1 to same name convention for both

Cadair · 2026-06-03T13:20:01Z

 ]


+class _refracted_angle_sine_model:


This should be refactored to be a subclass of Model like some of the other classes in this file, and not use @custom_model, like this example: https://docs.astropy.org/en/stable/modeling/new-model.html#a-full-example-of-a-linemodel

… will be in gwcs (PR pending)

Copilot

Pull request overview

This PR adds support for representing FITS grating/grism spectral WCS (“-GRA/-GRI”) as an Astropy modeling transform, enabling non-linear spectral dispersion handling in dkist.wcs.

Changes:

Added a grating/grism spectral transform builder (build_grating_spectral_transform) plus a public compatibility wrapper (generate_grating_spectral_transform).
Introduced an ASDF-stable refracted_angle_sine_model custom model to preserve a stable import path for serialization.
Added a regression test that compares the generated transform against astropy.wcs.WCS(...).spectral.pixel_to_world().

Reviewed changes

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

File	Description
`dkist/wcs/models.py`	Adds the grating/grism spectral transform builder + ASDF-stable refracted-angle model; wires in gwcs spectroscopy model.
`dkist/wcs/tests/test_models.py`	Adds a test validating the new spectral transform matches Astropy WCS for a representative header.
`changelog/713.feature.rst`	Documents the addition of `generate_grating_spectral_transform()` as a compatibility wrapper.

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

codspeed-hq · 2026-06-15T17:34:57Z

Merging this PR will degrade performance by 6.94%

❌ 1 regressed benchmark
✅ 15 untouched benchmarks

Warning

Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	`BASE`	`HEAD`	Efficiency
❌	Simulation	`test_plot_dataset[axes0]`	3.1 s	3.4 s	-6.94%

Tip

Investigate this regression by commenting @codspeedbot fix this regression on this PR, or directly use the CodSpeed MCP with your agent.

_{Comparing non_linear_spectral_dispersion (5913d79) with main (5d50445)¹}

No successful run was found on main (1a7c1db) during the generation of this report, so 5d50445 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

…dated gwcs grating equation

Copilot

Pull request overview

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

… a thin refracted_angle_sine_model subclass in here

…r_wavelength. Theseare now computed in gwcs

Copilot

Pull request overview

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

…which inputs have default values, removed generate_grating_spectral_transform(), adn updated changelog

…KISTDC/dkist into non_linear_spectral_dispersion

…which inputs have default values, removed generate_grating_spectral_transform(), adn updated changelog

…string

eigenbrot

Approving, but the current PR description is out of date

…_pure_python.yml (#731) Bumps [OpenAstronomy/github-actions-workflows/.github/workflows/publish_pure_python.yml](https://github.com/openastronomy/github-actions-workflows) from 2.5.0 to 3.0.1. - [Release notes](https://github.com/openastronomy/github-actions-workflows/releases) - [Commits](OpenAstronomy/github-actions-workflows@a138926...e5af21e) --- updated-dependencies: - dependency-name: OpenAstronomy/github-actions-workflows/.github/workflows/publish_pure_python.yml dependency-version: 3.0.1 dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Drew Leonard <andy.j.leonard@gmail.com>

…ts to load_dataset, which silently dropped it and only honored it for Path input (#737) Co-authored-by: Drew Leonard <andy.j.leonard@gmail.com>

Bumps [actions/checkout](https://github.com/actions/checkout) from 6 to 7. - [Release notes](https://github.com/actions/checkout/releases) - [Commits](actions/checkout@v6...v7) --- updated-dependencies: - dependency-name: actions/checkout dependency-version: '7' dependency-type: direct:production update-type: version-update:semver-major ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Drew Leonard <andy.j.leonard@gmail.com>

Bumps [CodspeedHQ/action](https://github.com/codspeedhq/action) from 4.17.0 to 4.18.1. - [Release notes](https://github.com/codspeedhq/action/releases) - [Changelog](https://github.com/CodSpeedHQ/action/blob/main/CHANGELOG.md) - [Commits](CodSpeedHQ/action@9d332c4...a4a36bb) --- updated-dependencies: - dependency-name: CodspeedHQ/action dependency-version: 4.18.1 dependency-type: direct:production update-type: version-update:semver-minor ... Signed-off-by: dependabot[bot] <support@github.com> Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com> Co-authored-by: Drew Leonard <andy.j.leonard@gmail.com>

updates: - [github.com/zizmorcore/zizmor-pre-commit: v1.25.2 → v1.26.1](zizmorcore/zizmor-pre-commit@v1.25.2...v1.26.1) - [github.com/astral-sh/ruff-pre-commit: v0.15.15 → v0.15.20](astral-sh/ruff-pre-commit@v0.15.15...v0.15.20) - [github.com/PyCQA/isort: 8.0.1 → 9.0.0a3](PyCQA/isort@8.0.1...9.0.0a3) Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> Co-authored-by: Drew Leonard <andy.j.leonard@gmail.com>

…in GWCS pr)

# Conflicts: # .github/workflows/build_workshop.yml # .github/workflows/ci.yml # .github/workflows/codspeed-walltime.yml # .github/workflows/codspeed.yml # .github/workflows/dkist_downstreams.yml # .github/workflows/prepare-release.yml # .github/workflows/sub_package_update.yml # .pre-commit-config.yaml

adding non linear spectral dispersion code

8c43c12

A-Derks requested a review from Cadair April 30, 2026 19:21

A-Derks added 2 commits April 30, 2026 13:23

adding chagelog

cfd3d4e

linting

2e45f28

A-Derks requested review from eigenbrot, fraserwatson and krislarsnso May 1, 2026 17:52

eigenbrot requested changes May 4, 2026

View reviewed changes

Cadair reviewed May 5, 2026

View reviewed changes

krislarsnso requested changes May 5, 2026

View reviewed changes

A-Derks added 3 commits May 27, 2026 14:30

Merge branch 'main' into non_linear_spectral_dispersion

de99978

renaming variables, adding better docstrongs

e8cadfd

making new class to hold grating function

a3c8938

Cadair reviewed Jun 3, 2026

View reviewed changes

eigenbrot reviewed Jun 8, 2026

View reviewed changes

Comment thread dkist/wcs/models.py

A-Derks added 3 commits June 9, 2026 09:13

Merge branch 'main' into non_linear_spectral_dispersion

774fd73

moving generate_grating_spectral_transform stuff out of this repo. It…

5893f99

… will be in gwcs (PR pending)

moving alpha_in and alpha_out computations here

19a35c7

A-Derks marked this pull request as draft June 11, 2026 19:53

refactoring models stuff into smaller functions

86d9515

A-Derks requested a review from Copilot June 12, 2026 15:41

Copilot started reviewing on behalf of A-Derks June 12, 2026 15:41 View session

Copilot AI reviewed Jun 12, 2026

View reviewed changes

Comment thread dkist/wcs/models.py Outdated

Comment thread dkist/wcs/models.py Outdated

Comment thread dkist/wcs/models.py

A-Derks added 2 commits June 15, 2026 11:16

adding extra grism params

7f9ba37

Update grating spectral wrapper for new gwcs grating-mode inputs

3dd823f

Restore composed alpha_out grating transforms in dkist against the up…

543f6cb

…dated gwcs grating equation

A-Derks requested a review from Copilot June 16, 2026 19:30

Copilot started reviewing on behalf of A-Derks June 16, 2026 19:30 View session

Copilot AI reviewed Jun 16, 2026

View reviewed changes

Comment thread dkist/wcs/models.py Outdated

Comment thread dkist/wcs/models.py Outdated

Comment thread changelog/713.feature.rst Outdated

Merge branch 'main' into non_linear_spectral_dispersion

5913d79

A-Derks marked this pull request as ready for review June 16, 2026 20:18

A-Derks added 3 commits June 18, 2026 11:18

moved refracted_angle_sine_model guts to gwcs repo, now there is just…

f1aaae2

… a thin refracted_angle_sine_model subclass in here

Merge branch 'main' into non_linear_spectral_dispersion

e49ed4f

moved refracted_angle_sine_model guts to gwcs repo, now there is just…

375cf78

… a thin refracted_angle_sine_model subclass in here

krislarsnso approved these changes Jun 18, 2026

View reviewed changes

A-Derks and others added 2 commits June 30, 2026 09:03

removed computing of reference_refracted_angle and grism_parameter_pe…

368cb6b

…r_wavelength. Theseare now computed in gwcs

Merge branch 'main' into non_linear_spectral_dispersion

0540894

A-Derks requested review from Cadair, Copilot and eigenbrot June 30, 2026 15:24

Copilot started reviewing on behalf of A-Derks June 30, 2026 15:25 View session

Copilot AI reviewed Jun 30, 2026

View reviewed changes

Comment thread dkist/wcs/models.py Outdated

Comment thread dkist/wcs/models.py Outdated

Comment thread dkist/wcs/models.py Outdated

Comment thread dkist/wcs/tests/test_models.py Outdated

eigenbrot requested changes Jun 30, 2026

View reviewed changes

Comment thread changelog/713.feature.rst Outdated

Comment thread dkist/wcs/models.py Outdated

Comment thread dkist/wcs/models.py

Comment thread dkist/wcs/models.py Outdated

A-Derks added 5 commits June 30, 2026 12:39

removed unused refracted_angle_sine_model, added documentation about …

c4dc59d

…which inputs have default values, removed generate_grating_spectral_transform(), adn updated changelog

Merge branch 'main' into non_linear_spectral_dispersion

58e262c

Merge branch 'non_linear_spectral_dispersion' of https://github.com/D…

4633621

…KISTDC/dkist into non_linear_spectral_dispersion

removed unused refracted_angle_sine_model, added documentation about …

d566b1f

…which inputs have default values, removed generate_grating_spectral_transform(), adn updated changelog

removing unnecessary addition to build_grating_spectral_transform doc…

569b0a3

…string

eigenbrot approved these changes Jul 1, 2026

View reviewed changes

A-Derks and others added 9 commits July 6, 2026 15:31

Merge branch 'main' into non_linear_spectral_dispersion

b102d5d

Let us all be more tolerant of numbers very close to zero (#739)

2fc780b

Forward ignore_version_mismatch for string, iterable and Results inpu…

24929e6

…ts to load_dataset, which silently dropped it and only honored it for Path input (#737) Co-authored-by: Drew Leonard <andy.j.leonard@gmail.com>

WavelengthFromGratingEquation --> WavelengthFromGrismEquaion (change …

1f149ed

…in GWCS pr)

fraserwatson approved these changes Aug 31, 2026

View reviewed changes

		assert u.allclose(shift1.offset, 0)


		def test_generate_grating_spectral_transform():

Uh oh!

Conversation

A-Derks commented Apr 30, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Cadair left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

codspeed-hq Bot commented Jun 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will degrade performance by 6.94%

Performance Changes

Footnotes

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

A-Derks commented Apr 30, 2026 •

edited

Loading

Cadair left a comment •

edited

Loading

codspeed-hq Bot commented Jun 15, 2026 •

edited

Loading