Skip to content

gallery_runner_missing_tiers.md

Latest commit

 

History

History
27 lines (24 loc) · 1.13 KB

gallery_runner_missing_tiers.md

File metadata and controls

27 lines (24 loc) · 1.13 KB
Raw

Gallery runner: add visualization_upper + decide the modeling_visualization_jit tier

Type: feature Target: workspaces Repos:

  • autolens_workspace_test Difficulty: small Autonomy: supervised Priority: low Status: formalised Filed: 2026-07-16 (backfilled from git)

First real findings from the Eyes conductor's live survey (PyAutoBrain#117 Phase 2, eyes survey autolens_workspace_test):

  1. scripts/imaging/visualization_upper.py and scripts/interferometer/visualization_upper.py produce full figure sets (12+17 png) but are missing from scripts/gallery/gallery_run.sh's DEFAULT_SCRIPTS — add them (time them first; slot into fast or slow tier).
  2. The modeling_visualization_jit* scripts (imaging ×3, interferometer, point_source) are never-rendered in the survey. Decide: include them in a runner tier (probably --all) or exclude them from the producer-token match story — they are JIT smoke tests whose figures may still be worth eyeballing once.
  3. While there: consider eyes survey output as the runner's source of truth for gaps going forward (the survey already diffs producers vs rendered trees).