[Jammy2211]

Part of PyAutoLabs/autogalaxy_workspace#268 (phase 3 of PyAutoLabs/PyAutoArray#669).

Merge only after the PyAutoArray release that carries PyAutoLabs/PyAutoArray#670. The tutorials call al.util.dataset.simulate_if_needed, which released PyAutoArray doesn't have yet.

Overlap: the Mind task howto-chapter2-format also claims this repo. Its branch isn't on GitHub, so overlap couldn't be checked. If it touches the same auto-simulate blocks, whichever merges second takes a small rebase.

Summary

Replaces the 8-line should_simulate + subprocess.run auto-simulate block with the one-line al.util.dataset.simulate_if_needed(dataset_path, simulator=...) in 24 tutorial scripts (29 blocks).

• The rewrite is mechanical and done with an AST matcher, so it only touches the exact standard block. Dataset and simulator paths are kept exactly. A redundant str(...) around the path is dropped because the helper calls str() itself. Behaviour is unchanged.
• One in-block comment, "Use the same Python interpreter as this tutorial.", is dropped together with the sys.executable call it described.
• The 24 matching notebooks were regenerated with PyAutoHands generate.py howtolens --assistant-root <autolens_assistant>. I reverted unrelated generator noise (3 Colab-context refreshes and Unicode escaping in 3 notebooks), so each notebook diff is only this rewrite.

Testing

• Every inserted call is identical to black's output, and all changed scripts compile.
• check_dataset_allowlist: 30 guarded call sites before and 30 after.
• Smoke run: chapter_1_introduction/tutorial_6_fitting.py passed against released autolens with merged PyAutoArray main, simulating its dataset through the helper.

Not changed

scripts/chapter_1_introduction/tutorial_5_data.py:696 simulates inline, so it isn't the standard block and was left as-is.

Merge mode: human /prm, after the release.

---

Generated by Claude Code