Skip to content

feat: keep dataset-only image_with_points.png; after-fit figure → fit_image_with_points.png #781

Closed

Description

@Jammy2211

Overview

Follow-up to #779. The point visualizer writes the dataset-only figure (observed image + observed positions) before the fit and the fit figure (+ model positions + critical curves) after it, but both under image_with_points.png, so the dataset-only version is overwritten and never reaches PyAutoEyes. This renames the after-fit figure to fit_image_with_points.png so both coexist, and shows both in autolens_visualization / PyAutoEyes.

Plan

  • PyAutoLens: after-fit figure → fit_image_with_points.png; dataset-only stays image_with_points.png; fit plots.yaml key renamed to fit_image_with_points (unreleased, so safe).
  • autolens_workspace: docstrings/comments in the six point-source scripts describe both figures.
  • autolens_visualization: both point producers keep both figures; gallery 66 → 68.
  • PyAutoEyes: dashboard_refresh after the viz PR merges (no Eyes PR).
  • Order: library → workspace → visualization → Eyes refresh.

Tier: judge — merge mode: human /prm

Detailed implementation plan

Affected Repositories

  • PyAutoLens (primary)
  • autolens_workspace
  • autolens_visualization

Branch Survey

Repository Current Branch Dirty?
lens/PyAutoLens main clean
lens/autolens_workspace main untracked root dataset.fits (not ours)
lens/autolens_visualization main clean

Suggested branch: feature/point-image-fit-filename

Implementation Steps

  1. autolens/point/plot/fit_point_plots.py plot_image_with_points: filename image_with_points when fit is None, else fit_image_with_points. The fit version is retitled "… Fit Image With Points", and the docstring covers dataset-only use before modeling.
  2. autolens/config/visualize/plots.yaml: fit_point_dataset.image_with_points → fit_point_dataset.fit_image_with_points. point_dataset.image_with_points is unchanged.
  3. autolens/point/model/plotter.py: read the renamed fit key.
  4. Tests: before-fit writes only image_with_points.png; after-fit writes fit_image_with_points.png; both coexist after before_fit + visualize; the dataset-only call draws no Model scatter and no lines.
  5. autolens_workspace scripts/point_source/{modeling,features/fluxes,features/time_delays,features/extra_galaxies/modeling,features/scaling_relation/modeling}.py: describe both figures. Regenerate notebooks and smoke-test.
  6. autolens_visualization: producers extra_pngs=("image_with_points.png", "fit_image_with_points.png"), updated docstrings. Re-run producers + gallery_build (+ --check).

Key Files

  • autolens/point/plot/fit_point_plots.py, autolens/point/model/plotter.py, autolens/config/visualize/plots.yaml
  • autolens_visualization/scripts/point_source_{image,source}_chi_squared/visualization.py

Original Prompt

Click to expand starting prompt

ok, can I check we also have a dataset-only version of this which does not include model points or critical curves? It should be an option (E.g. for display before lens modeling). This model image is good thiugh, albeit the filename should prob be model_image_with_points.png

Decision (human, 2026-10-21): after-fit name fit_image_with_points.png; keep both figures in Eyes.

Activity

  1. Jammy2211 commented on Oct 21, 2026

    @Jammy2211
    CollaboratorAuthor

    Library PR Created: #782 (843 passed, 1 xfailed). Workspace: PyAutoLabs/autolens_workspace#592. Visualization: PyAutoLabs/autolens_visualization#7. Heart YELLOW was acknowledged by the human. Judge tier: human /prm, library first; then the Eyes refresh.

  2. Jammy2211 commented on Oct 21, 2026

    @Jammy2211
    CollaboratorAuthor

    Shipped — 2026-10-21

    Merged by human /prm, library first:

    • PyAutoLens#782: the after-fit figure is now saved as fit_image_with_points.png, so the dataset-only image_with_points.png (observed positions only, for display before lens modeling) is kept. Fit plots.yaml key renamed to fit_image_with_points. Pending release (with feat: overlay point-source positions on a dataset image #780).
    • autolens_workspace#592: point-source script docs describe both figures; notebooks regenerated.
    • autolens_visualization#7: both point domains keep both figures (gallery 66 → 68). The first lint run failed ruff format; a formatting-only commit, human-approved, made it green.
    • PyAutoEyes: dashboard_refresh run 39591165754 (commit a4cc7da). The board lists image_with_points.png and fit_image_with_points.png for both point domains.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions