Repository navigation
docs: restore and complete reviewed Chapter 3 tutorials #125
Description
Overview
Implement the human-reviewed HowToLens Chapter 3 material (66 approved blocks) and audit early lessons for missing content. Restore Fit Problems and consolidate adaptive material into one tutorial.
Approved plan
- Implement the approved Goldilocks evidence explanation and explicit coefficient selection.
- Preserve accepted early content and audit Tutorial 1 / Tutorial 2 opening against pre-consolidation and repair history.
- Restore Tutorial 5 Fit Problems; consolidate old 8/11/12 into Tutorial 6 Adaptive Reconstruction with Delaunay/Hilbert/Split and an actual chained fit.
- Embed Maresca Figure 1, repair contents, numbering, cross-references and chapter navigation.
- Regenerate notebooks/catalogues and run structural, API, numerical and smoke checks.
Tier: judge — merge mode: human /prm.
Implementation
Repo: HowToLens, clean base origin/main a7bd32c (Chapter 2 PR #124 merged).
Branch: feature/howtolens-ch3-approved.
Worktree: /home/jammy/Code/PyAutoLabs/.worktrees/howtolens-ch3-approved/HowToLens.
Canonical sources: scripts/chapter_3_pixelizations/; notebooks generated through PyAutoHands.
Approved text recovered from chat 01a1a5a6-2da3-79e0-9b7b-ae8072a3ea02; latest revisions to blocks 1, 22, 25 and 27 take precedence. No library changes or merge authorized.
Original implementation request
Yes, go ahead and implement all the approved changes, audit tutorial 1 and the start of tutorial 2 for missing content, regenerate the notebooks and run the checks.
Activity
Jammy2211 commentedon Nov 4, 2026
Shipped — 2026-11-04
Merged #126 as 96bf477.
Restored the approved Chapter 3 narrative, Fit Problems and missing introductory content; consolidated adaptive reconstruction into Tutorial 6; regenerated notebooks and navigation. All six full-size numerical runs passed, and all eight GitHub jobs passed on 61eec78, including Python 3.12 and 3.13 smoke tests.
The strict CI configuration check required full-dataset declarations inside Tutorials 5 and 6 rather than profile overrides. The correction preserves numerical code and generated notebook content. Searches were checked using sampler bypass with actual likelihood evaluations, not posterior convergence tests.