docs: surface datacube modeling and document 14 undocumented example folders

[Jammy2211]

Summary

Datacube modeling shipped in May 2026 under scripts/interferometer/features/datacube/, but was effectively invisible: the parent features/README.md "Folders" list never gained a datacube entry, and nothing in the top-level dataset-type listings hinted that anything beyond single-channel continuum data was supported. The original requester's thread (PyAutoMind#18) sat open for three months after the feature it asked for had already landed.

Auditing that one omission surfaced a broader instance of the same defect, fixed in the second commit.

Commit 1 — surface datacube modeling

File	Change
scripts/interferometer/features/README.md	Add the missing datacube bullet
scripts/interferometer/README.md	State that both continuum data and spectral-line cubes are supported; point at features/datacube
scripts/README.md, root README.md	Note cubes on the interferometer entry
llms.txt	Add a datacube route to the "I want to…" routing list

Commit 2 — the same defect, everywhere else

An audit comparing every scripts/**/README.md against the folders actually beside it found 14 subfolders that no parent README mentioned at all:

• group/features: scaling_relation; advanced adds double_source_plane_lens, mass_stellar_dark
• guides/results: aggregator, database
• imaging/data_preparation: gui, manual
• imaging/features/advanced: los_halos
• interferometer/data_preparation: examples
• interferometer/features: linear_light_profiles, multi_gaussian_expansion, scaling_relation
• point_source/features: scaling_relation

Also fixes a stale name: imaging/features/advanced/subhalo/README.md listed detection, but the folder is detect.

Every description is taken from that folder's own README or a script docstring — none inferred from the folder name.

Verification

• check_navigator passes (path check + banner lint) — every new folder reference resolves.
• Re-ran the audit after the fix: 0 undocumented folders remain.
• No catalogue regeneration needed: navigator.py builds llms-full.txt / workspace_index.json from script docstrings only, never from .md. llms.txt is hand-maintained and edited directly.
• Notebook READMEs are the generator's copies of the scripts/ originals (generate.py copies scripts/**/*.md) and are updated in step.

Scope

Documentation only — no .py, no notebooks, no datasets touched.

Pairs with PyAutoLabs/PyAutoLens#694, which documents the same feature in the RTD overview.

---

Generated by Claude Code