docs: combine guides/results start_here.py files (simple + aggregator)

[Jammy2211]

Overview

The guides/results/ package currently has two entry-point tutorials: start_here.py (simple JSON/FITS loading, ~140 lines) and aggregator/start_here.py (full aggregator workflow, ~440 lines). Users hitting the results section have to read both to understand the choice between them. This task combines them into a single guides/results/start_here.py for both autogalaxy_workspace and autolens_workspace, with the aggregator introduced as the second half — making it explicit that downstream sections (samples, galaxies, fits, etc.) can be reached through either API.

Plan

• Combine the two files so a reader sees one entry point that explains:
  1. The simple, instant JSON/FITS loading workflow (current content of start_here.py).
  2. A clear pivot: "for many fits, use the aggregator instead — same objects, generator-based."
  3. The aggregator workflow itself (current content of aggregator/start_here.py), with each section noting that the result can also be reached via the simple-loading API where applicable.
• Drop the duplicate model-fit illustration so the merged file uses the JSON/FITS-loaded objects as the running example, falling back to the small in-script fit only where the aggregator illustration genuinely needs it.
• Delete the now-redundant guides/results/aggregator/start_here.py.
• Update internal cross-references in sibling files (samples.py, models.py, data_fitting.py, galaxies_fit(s).py, interferometer.py, queries.py, samples_via_aggregator.py) so any "see aggregator/start_here.py" pointer is rewritten to "see results/start_here.py — Aggregator section".
• Update __Contents__ block at the top of the merged file to reflect both halves.
• Mirror identical structure across both workspaces (autogalaxy uses Galaxies, autolens uses Tracer — content shape stays parallel).

Detailed implementation plan

Affected Repositories

• autolens_workspace (primary)
• autogalaxy_workspace

Work Classification

Workspace

Branch Survey

Repository	Current Branch	Dirty?
./autolens_workspace	main	clean
./autogalaxy_workspace	main	clean

Suggested branch: feature/merge-results-start-here
Worktree root: ~/Code/PyAutoLabs-wt/merge-results-start-here/ (created later by /start_workspace)

Implementation Steps

1. Rewrite autolens_workspace/scripts/guides/results/start_here.py as a single tutorial in two halves:

   • Simple loading (existing 145 lines, mostly preserved): docstring → result_path → tracer (from JSON) → model → samples → FITS files.
   • Aggregator pivot (new bridging paragraph): "Everything above loaded one fit at a time. To analyse many fits, use the aggregator." Note that you can keep using samples, tracer, model etc., just sourced from a generator.
   • Aggregator (adapted from current aggregator/start_here.py): keep the in-script Aggregator.from_directory(...) setup, generators explanation, database-file note, workflow examples, then the deeper sections (Result, Samples, Linear Light Profiles, Galaxies, Fits, Units & Cosmology, Linear Light Profiles / Basis, Pixelization). Where each section is parallel to the simple-loading API, add a one-line "(equivalent to from_json(...).tracer above)" note.

2. Rewrite autogalaxy_workspace/scripts/guides/results/start_here.py with the same merge, using Galaxies instead of Tracer, ag.Imaging instead of al.Imaging, etc. Keep section structure parallel to autolens version.

3. Delete obsolete files:

   • autogalaxy_workspace/scripts/guides/results/aggregator/start_here.py
   • autolens_workspace/scripts/guides/results/aggregator/start_here.py

4. Update sibling references — grep both aggregator/ subfolders for "start_here.py" pointers and rewrite them to point at guides/results/start_here.py with a section anchor (e.g. "see results/start_here.py § Aggregator").

5. Verify:

   • Run bash run_all_scripts.sh (or /run_all_script_fix_failures) on both workspaces in test mode.
   • Run scripts/check_sizes.sh after edits — start_here.py will grow ~3×, aggregator/start_here.py will be deleted (refresh snapshot to register the deletion via --update).
   • Re-generate notebooks via /generate_and_merge once code is approved.

Key Files

• autolens_workspace/scripts/guides/results/start_here.py — rewrite (combined ~520 lines)
• autolens_workspace/scripts/guides/results/aggregator/start_here.py — delete
• autogalaxy_workspace/scripts/guides/results/start_here.py — rewrite (combined ~500 lines)
• autogalaxy_workspace/scripts/guides/results/aggregator/start_here.py — delete
• Sibling files in both aggregator/ subfolders — update internal start_here.py cross-references

Original Prompt

Click to expand starting prompt

Can you combine guides/results/start_here.py and guides/results/aggregator/start_here.py into one,
at guides/results/start_here.py, for both autogalaxy_workspace and autolens_workspace.

I think its fine for the aggregator to be eplxained in the first results start_here.py file, after we introduce
the simple json loading, and make it clear steps like using the samples can be done using either API.

[Jammy2211]

Smoke Test Results — 2026-04-28

All 13 smoke tests passed across 2 workspaces.

Workspace	Passed	Failed	Total
autolens_workspace	7	0	7
autogalaxy_workspace	6	0	6