Skip to content

chore: delete dead PYAUTO_SMALL_DATASETS/PYAUTO_FAST_PLOTS from build profiles#29

Merged
Jammy2211 merged 1 commit into
mainfrom
feature/autofit-dead-env-knobs
Jul 23, 2026
Merged

chore: delete dead PYAUTO_SMALL_DATASETS/PYAUTO_FAST_PLOTS from build profiles#29
Jammy2211 merged 1 commit into
mainfrom
feature/autofit-dead-env-knobs

Conversation

@Jammy2211

Copy link
Copy Markdown
Contributor

Summary

PYAUTO_SMALL_DATASETS and PYAUTO_FAST_PLOTS are dead config here — nothing that reads them is ever imported in an autofit-only workspace. This deletes them from the defaults: block of this repo's build profile(s).

  • PYAUTO_SMALL_DATASETS is read only in PyAutoArray (util/dataset_util.py, structures/grids/uniform_2d.py, mask/mask_2d.py, operators/over_sampling/over_sample_util.py, operators/convolver.py), PyAutoGalaxy and PyAutoLens. Zero occurrences in PyAutoFit or PyAutoNerves — not even a docstring.
  • PYAUTO_FAST_PLOTS is read in autoarray/plot/utils.py and autogalaxy/util/plot_utils.py. In PyAutoFit it appears only in prose (autofit/non_linear/quick_update.py:71 and a test docstring), never in an os.environ lookup.
  • PyAutoFit does not import autoarray, and no script in this repo imports autoarray / autogalaxy / autolens — so the reading code never loads.
  • PyAutoFit's single tight_layout() is in autofit/non_linear/live_viewer.py:102, inside the interactive plt.ion() desktop viewer, which never runs headless in CI. Wiring PYAUTO_FAST_PLOTS up there would buy nothing, so deletion is the right call rather than implementation.

The stale comments gave the copy-paste origin away: an autofit config claiming to "reduce MGE gaussians" and skip "critical curve/caustic overlays", both lensing concepts.

This does not violate the release-profile doctrine ("every var this profile cares about gets an EXPLICIT value, not left absent") — autofit cares about neither, so absence is correct rather than a fall-through hazard. autofit_workspace_test/config/build/env_vars_release.yaml already omitted both and is untouched.

Follow-up to #321 (the env_vars key-order sweep), which made the profiles diffable and surfaced this.

Scripts Changed

None — configuration only. No file under scripts/ is touched.

  • config/build/env_vars*.yaml — removed the two dead defaults: entries. No overrides: entry referenced either key, so the change is contained to defaults:.

Test Plan

  • Resolved-env delta is exactly the two keys. Across 227 script×profile environments in the three affected repos, resolved from an empty base via resolve_clean() (PyAutoHands/autohands/validate_env_profiles.py): only PYAUTO_SMALL_DATASETS and PYAUTO_FAST_PLOTS disappear. Zero other key added, removed or changed; overrides: and args_default byte-identical in every profile.
  • validate_env_profiles.py verdict unchanged from main.
  • git diff --stat contains nothing but config/build/env_vars*.yaml.
  • Re-grepped: no .py in autofit_workspace / autofit_workspace_test / HowToFit references either var.

Refs PyAutoLabs/autofit_workspace#111.

Generated by the PyAutoLabs agent workflow.

… profiles

Neither knob is ever read in an autofit-only workspace. PYAUTO_SMALL_DATASETS
is read only in PyAutoArray/PyAutoGalaxy/PyAutoLens; PYAUTO_FAST_PLOTS only in
PyAutoArray and PyAutoGalaxy. PyAutoFit does not import autoarray and no script
here imports autoarray/autogalaxy/autolens, so the reading code never loads.

The stale comments gave the copy-paste away: an autofit config claiming to
'reduce MGE gaussians' and skip 'critical curve/caustic overlays'.

Refs PyAutoLabs/autofit_workspace#111
@Jammy2211 Jammy2211 added the pending-release PR queued for the next release build label Jul 23, 2026
@Jammy2211
Jammy2211 merged commit 11e77cc into main Jul 23, 2026
10 checks passed
@Jammy2211
Jammy2211 deleted the feature/autofit-dead-env-knobs branch July 23, 2026 14:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

pending-release PR queued for the next release build

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant