Skip to content

Rectangular mesh split API update: RectangularAdapt* → RectangularBilinearAdapt* - #74

Merged
Jammy2211 merged 2 commits into
mainfrom
claude/wake-up-22piwp
Aug 21, 2026
Merged

Rectangular mesh split API update: RectangularAdapt* → RectangularBilinearAdapt*#74
Jammy2211 merged 2 commits into
mainfrom
claude/wake-up-22piwp

Conversation

@Jammy2211

Copy link
Copy Markdown
Collaborator

What

The HowToLens leg of the rectangular mesh split rollout (PyAutoArray#462 / PyAutoGalaxy#579 / PyAutoLens#707), which retired the RectangularAdaptDensity / RectangularAdaptImage names:

  • Chapter-3 scripts + notebooks (12 tutorials): every reference moves to RectangularBilinearAdaptDensity / RectangularBilinearAdaptImage — the workspace default, per the no-workspace-example-uses-RTU decision recorded on autolens_workspace's feature/rectangular-bilinear-rtu-mesh-split.
  • config/priors/mesh/: rectangular_adapt_*.yaml renamed to rectangular_bilinear_adapt_*.yaml (keys updated) with rectangular_rtu_adapt_*.yaml twins added (same priors), matching that branch's config layout.
  • The navigator catalogue does not embed the class names, so llms-full.txt / workspace_index.json are unchanged.

Why

main smoke has been red since 2026-08-20 22:30 UTC — first from the fnnls numba regression fixed in PyAutoArray#463, and since this morning's mesh-split merges every tutorial constructing RectangularAdaptDensity additionally fails with AttributeError (confirmed on the main re-run). This red is the HowToLens: Smoke Tests failure on main that blocked the nightly release.

This PR's smoke run picks up PyAutoArray's same-named claude/wake-up-22piwp fix branch via the workflow's branch matching, so the two fixes are validated together; PyAutoArray#463 should merge first.

Validation

  • tutorial_8_adaptive_pixelization + tutorial_11_brightness_adaption (the two originally-red scripts) pass locally against current library mains + PyAutoArray#463.
  • Full local run_smoke.py sweep across all 57 scripts running at time of writing; CI on this PR is the authoritative gate.

🤖 Generated with Claude Code

https://claude.ai/code/session_015GsUfbCwPd4XC8kpsiUJp7


Generated by Claude Code

claude added 2 commits August 21, 2026 20:05
…linearAdapt*

PyAutoArray#462 / PyAutoGalaxy#579 / PyAutoLens#707 split the rectangular
adaptive mesh family into Bilinear (rank-CDF, the workspace default) and RTU
(kernel-CDF, advanced/interferometer) variants, retiring the old
RectangularAdaptDensity / RectangularAdaptImage names. This is the HowToLens
leg of that rollout, mirroring autolens_workspace's
feature/rectangular-bilinear-rtu-mesh-split: every tutorial reference moves to
the Bilinear default (per the no-workspace-example-uses-RTU decision recorded
there), notebooks carry the same textual rename, and the prior configs are
renamed with RTU twins added (same priors), matching that branch's config
layout.

Chapter-3 scripts + notebooks (12 tutorials), config/priors/mesh. The
navigator catalogue does not embed the class names, so llms-full.txt and
workspace_index.json are unchanged.

Smoke-verified locally against current library mains (with the PyAutoArray
fnnls JAX-coercion fix): the two tutorials red since 2026-08-20
(tutorial_8_adaptive_pixelization, tutorial_11_brightness_adaption) pass.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015GsUfbCwPd4XC8kpsiUJp7
tutorial_5_bayesian_formalism imports overlay_grid_from by its module path,
which the mesh split also renamed: rectangular_adapt_density ->
rectangular_rtu_adapt_density (where the function now lives). Caught by the
full local run_smoke.py sweep (49/50 passed, this the one failure); the
script now passes.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_015GsUfbCwPd4XC8kpsiUJp7
@Jammy2211
Jammy2211 merged commit ecf8785 into main Aug 21, 2026
7 checks passed
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.

2 participants