Skip to content

fix(config): configure Delaunay.areas_factor and dPIE ra/rs notation - #630

Merged
Jammy2211 merged 1 commit into
mainfrom
feature/workspace-config-cleanup
Sep 27, 2026
Merged

Jammy2211 merged 1 commit into
mainfrom
feature/workspace-config-cleanup

Conversation

@Jammy2211

@Jammy2211 Jammy2211 commented Sep 27, 2026 •

Copy link
Copy Markdown
Collaborator

Summary

Two library config gaps that workspaces were papering over:

  • config/priors/mesh/delaunay.yaml was a bare Delaunay:, so af.Model(ag.mesh.Delaunay) / al.mesh.Delaunay reported areas_factor Prior Missing. It is now a Constant at the class default 0.5 (the value autolens_workspace already sets).
  • config/notation.yaml had no label / label_format for the dPIE ra / rs parameters, so both fell back to the first-letter label r. Adds r_{\rm a} / r_{\rm s} and '{:.4f}', matching autolens_workspace.

Part of PyAutoLabs/PyAutoMind#441 (workspace-config-cleanup).

Merge order: merge this PR first — the companion PyAutoLabs/PyAutoLens#751's model-figure acceptance tests depend on this config. Workspace PRs (autofit_workspace, autogalaxy_workspace, autolens_workspace, autocti_workspace) follow in the same task and remove the now-redundant workspace overrides.

Heart YELLOW acknowledged by human (reasons: manifest drift: hub organism blurb (organs present) — 7 mismatches; manifest drift: organism-map blocks (generated) — 1 mismatch; manifest drift: workspace checkouts (manifest ↔ disk) — 1 mismatch; release validation stale: source moved since rehearsal (PyAutoFit, PyAutoArray, PyAutoGalaxy, PyAutoLens)).

API Changes

No Python API change. Config-default change: Delaunay models now get areas_factor = 0.5 as a fixed Constant from library config (previously it needed a workspace prior, or was reported as Missing). dPIE ra / rs now have their own LaTeX labels.
See full details below.

Test Plan

  • New: Delaunay model has no missing prior (test_autogalaxy/test_priors_config.py)
  • New: dPIE ra/rs labels are not the fallback (test_dual_pseudo_isothermal_mass.py) — both red before the config change
  • test_model_figure_semantics.py: real-class missing-state witness moved Delaunay.areas_factor → reg.CurvatureMask.coefficient
  • python -m pytest test_autogalaxy/ — 1243 passed
Full API Changes (for automation & release notes)

Changed Behaviour

  • autogalaxy/config/priors/mesh/delaunay.yaml — Delaunay.areas_factor configured as Constant value: 0.5; af.Model(ag.mesh.Delaunay) no longer reports a missing prior and has one fewer free parameter slot.
  • autogalaxy/config/notation.yaml — label.ra: r_{\rm a}, label.rs: r_{\rm s}, label_format.ra/rs: '{:.4f}'.

Migration

  • None required. Workspaces that set areas_factor / the dPIE labels themselves keep working (workspace config still overrides); those overrides become redundant.

Generated by the PyAutoLabs agent workflow.

🤖 Generated with Claude Code

priors/mesh/delaunay.yaml was a bare `Delaunay:`, so af.Model(ag.mesh.Delaunay)
(and al.mesh.Delaunay) reported "areas_factor  Prior Missing". Configure it as
a Constant at the class default (0.5), the value autolens_workspace already
sets.

notation.yaml had no label/label_format for the dPIE `ra` and `rs`
parameters, so both fell back to the first-letter label "r". Add r_{\rm a} /
r_{\rm s} and '{:.4f}', matching autolens_workspace.

Tests: Delaunay model has no missing prior; dPIE ra/rs labels are not the
fallback (both red before the config change). The model-figure semantics
test used Delaunay.areas_factor as its real-class witness of the `missing`
state; it now uses CurvatureMask.coefficient (still unconfigured), and the
Delaunay test pins areas_factor as a fixed Constant.

Refs PyAutoLabs/PyAutoMind#441

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@Jammy2211
Jammy2211 merged commit cbd89ce into main Sep 27, 2026
4 checks passed
@Jammy2211
Jammy2211 deleted the feature/workspace-config-cleanup branch September 27, 2026 14:20
@Jammy2211 Jammy2211 removed the pending-release PR queued for the next release build label Oct 7, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant