Feature/unconvolved images by Jammy2211 · Pull Request #386 · Jammy2211/PyAutoLens

Jammy2211 · 2026-02-09T15:55:17Z

This pull request makes it so that images which have not been operated on (e.g. before PSF convolution) can be computed and extracted easily from inversions, with the end goal making it so that these images can be easily output to .fits files during autolens lens modling.

This pull request refactors the terminology and internal logic related to reconstructed data in the inversion modules, aiming for clearer distinctions between operated (PSF-convolved) and unoperated (unconvolved) data. The changes update property and variable names, unify implementation patterns, and improve documentation. No core algorithms are changed, but naming is clarified throughout the codebase to avoid confusion between different forms of reconstructed data.

Key changes include:

Core Functionality: Galaxy Image Handling

Added a new galaxy_image_dict property to FitImaging and FitInterferometer, which provides galaxy images before any operation (e.g., without PSF convolution), and updated galaxy_model_image_dict to clarify its use for operated/model images. [1] [2] [3]
Updated the methods for combining reconstructed data to use mapped_reconstructed_operated_data for model data and use_operated flags for correct image types in both imaging and interferometer fits. [1] [2] [3] [4] [5]

Plotting and Output

Changed plotter methods to use reconstructed_operated_data instead of reconstructed_image, ensuring outputs match the intended data form. [1] [2] [3] [4]

Configuration Updates

Split the configuration options for outputting galaxy images in plots.yaml into fits_galaxy_images (without PSF convolution) and fits_model_galaxy_images (with PSF convolution), and updated relevant test configs for clarity and consistency. [1] [2]
Added a new use_border_relocator option to the inversion configuration, allowing control over pixel relocation behavior.
Changed the default for fits_are_zoomed in the main config to false to avoid zooming FITS outputs by default.

Testing and Validation

Updated and expanded tests to cover the new galaxy_image_dict property, and adjusted expected values in several tests to reflect the improved logic and new default settings. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11]

These changes collectively improve clarity, flexibility, and correctness in how galaxy images are handled, output, and tested throughout the codebase.

Copilot

Pull request overview

This PR refactors inversion “reconstructed data” terminology to clearly separate operated (e.g., PSF-convolved / transformed) vs unoperated (pre-operation) outputs, and introduces explicit accessors/config for exporting unconvolved galaxy images for FITS output workflows.

Changes:

Add / update fit-level galaxy image dictionary accessors to distinguish unoperated galaxy images from model/operated images.
Rename and propagate reconstructed-data accessors in fits and plotters (e.g., mapped_reconstructed_operated_data, reconstructed_operated_data).
Update visualize configuration and adjust/extend tests to reflect the new naming and defaults.

Reviewed changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
test_autolens/lens/test_to_inversion.py	Updates test to new inversion reconstructed-data accessor name.
test_autolens/interferometer/test_fit_interferometer.py	Updates expected values; still references old galaxy image dict property name.
test_autolens/imaging/test_simulate_and_fit_imaging.py	Passes new inversion settings option and updates expected likelihood/reconstruction assertions.
test_autolens/imaging/test_fit_imaging.py	Adds new `galaxy_image_dict` test and updates expected values.
test_autolens/config/visualize.yaml	Updates test visualization config to include new FITS galaxy image options.
test_autolens/config/general.yaml	Adds `use_border_relocator` to inversion config in test configuration.
test_autolens/analysis/test_result.py	Updates expected source-plane centre value.
test_autolens/analysis/test_analysis.py	Updates expected grid values for border relocator test.
autolens/interferometer/plot/fit_interferometer_plotters.py	Switches plotting calls to `reconstructed_operated_data`.
autolens/interferometer/fit_interferometer.py	Switches model data to operated reconstructed data; introduces `galaxy_image_dict` property.
autolens/imaging/plot/fit_imaging_plotters.py	Switches pixelization plotting to `reconstructed_operated_data`.
autolens/imaging/fit_imaging.py	Uses operated reconstructed data in `model_data`; adds `galaxy_image_dict`; updates linear-object dict flag naming.
autolens/config/visualize/plots.yaml	Splits FITS galaxy-image outputs and changes default `fits_are_zoomed` to `false`.

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

Copilot · 2026-02-09T16:03:02Z

test_autolens/interferometer/test_fit_interferometer.py

+        -0.14416215690290285, 1.0e-4
    )

    # Normal light + Linear Light PRofiles + Pixelization + Regularizaiton


Typo in comment: Regularizaiton should be Regularization.

Suggested change

# Normal light + Linear Light PRofiles + Pixelization + Regularizaiton

# Normal light + Linear Light PRofiles + Pixelization + Regularization

Copilot · 2026-02-09T16:03:02Z

autolens/interferometer/fit_interferometer.py

    @property
-    def galaxy_model_image_dict(self) -> Dict[ag.Galaxy, np.ndarray]:
+    def galaxy_image_dict(self) -> Dict[ag.Galaxy, np.ndarray]:
        """
        A dictionary which associates every galaxy in the tracer with its `image`.



FitInterferometer defines galaxy_image_dict but the previous galaxy_model_image_dict property appears to have been removed. The test suite (e.g. test_fit_interferometer.py::test___galaxy_model_image_dict) still references galaxy_model_image_dict, and this is also likely part of the public API. Please either reintroduce galaxy_model_image_dict (e.g. as an alias to galaxy_image_dict for backwards compatibility) or update all call sites/tests and document the breaking change.

Copilot · 2026-02-09T16:03:02Z

test_autolens/interferometer/test_fit_interferometer.py

    assert (fit.galaxy_model_image_dict[g0_no_light].native == np.zeros((7, 7))).all()

    assert fit.galaxy_model_image_dict[galaxy_pix_0][0] == pytest.approx(
-        -0.15950206, 1.0e-4
+        -0.14416215690290285, 1.0e-4
    )


This test still calls fit.galaxy_model_image_dict, but FitInterferometer now exposes galaxy_image_dict (and no longer defines galaxy_model_image_dict). Update the test to use the correct property (or re-add a backwards-compatible alias in FitInterferometer).

Copilot · 2026-02-09T16:03:03Z

test_autolens/analysis/test_analysis.py


    grid = fit.inversion.linear_obj_list[0].source_plane_data_grid.over_sampled

-    assert grid[2] == pytest.approx([-82.89544515, 52.7491249], 1.0e-4)
+    assert grid[2] == pytest.approx([-82.991148773, 52.81254921], 1.0e-4)


Both assertions in this test currently expect (effectively) the same grid[2] value for use_border_relocator=False and use_border_relocator=True, so the test no longer verifies that the flag changes behavior. Consider asserting that the two grids differ (or checking a specific relocation effect) so the setting is actually exercised.

Copilot · 2026-02-09T16:03:03Z

test_autolens/imaging/test_fit_imaging.py

+    )
+    assert (fit.galaxy_image_dict[g2] == np.zeros(9)).all()
+
+    # Normal light + Linear Light PRofiles + Pixelization + Regularizaiton


Typo in comment: PRofiles should be Profiles.

Suggested change

# Normal light + Linear Light PRofiles + Pixelization + Regularizaiton

# Normal light + Linear Light Profiles + Pixelization + Regularizaiton

Copilot · 2026-02-09T16:03:03Z

test_autolens/imaging/test_fit_imaging.py

+    )
+    assert (fit.galaxy_image_dict[g2] == np.zeros(9)).all()
+
+    # Normal light + Linear Light PRofiles + Pixelization + Regularizaiton


Typo in comment: Regularizaiton should be Regularization.

Suggested change

# Normal light + Linear Light PRofiles + Pixelization + Regularizaiton

# Normal light + Linear Light PRofiles + Pixelization + Regularization

Copilot · 2026-02-09T16:03:04Z

test_autolens/interferometer/test_fit_interferometer.py

+        -0.14416215690290285, 1.0e-4
    )

    # Normal light + Linear Light PRofiles + Pixelization + Regularizaiton


Typo in comment: PRofiles should be Profiles.

Suggested change

# Normal light + Linear Light PRofiles + Pixelization + Regularizaiton

# Normal light + Linear Light Profiles + Pixelization + Regularizaiton

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

…utoLens into feature/unconvolved_images

Jammy2211 added 5 commits February 8, 2026 11:18

fix unitt ests failed due to border swap

c2098de

galaxy_image_dict

009ebd4

fixes to generating unconvolved images

c087867

fix config files for tests

f876b92

removed reconstructwed_imAGE api for reconstructed_operated_data

8445a0b

Jammy2211 requested a review from Copilot February 9, 2026 15:55

Copilot started reviewing on behalf of Jammy2211 February 9, 2026 15:57 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Jammy2211 and others added 5 commits February 9, 2026 16:13

Update test_autolens/interferometer/test_fit_interferometer.py

93a5b0e

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

fix unit tests due to galaxy dict rename

8afda4b

Galaxy images ENum in aggregator

f6442cd

Merge branch 'feature/unconvolved_images' of github.com:Jammy2211/PyA…

0aa4140

…utoLens into feature/unconvolved_images

fixes

983af2b

Jammy2211 merged commit 983af2b into main Feb 9, 2026
0 of 8 checks passed

Jammy2211 deleted the feature/unconvolved_images branch February 13, 2026 13:43

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Feature/unconvolved images#386

Feature/unconvolved images#386
Jammy2211 merged 10 commits intomainfrom
feature/unconvolved_images

Jammy2211 commented Feb 9, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Copilot AI Feb 9, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Comments

	# Normal light + Linear Light PRofiles + Pixelization + Regularizaiton
	# Normal light + Linear Light PRofiles + Pixelization + Regularization

Conversation

Jammy2211 commented Feb 9, 2026

Core Functionality: Galaxy Image Handling

Plotting and Output

Configuration Updates

Testing and Validation

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Feb 9, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Comments