…utorial subprocess paths

11 tutorials across chapters 1-4 subprocess to simulator paths that
don't exist in HowToGalaxy:

- 10 reference `scripts/imaging/simulator.py` (writes dataset `simple`)
- 1 references `scripts/guides/plot/simulator.py` (writes dataset `sersic_x2`)

Both are autogalaxy_workspace-style paths inherited when these
tutorials were originally adapted from that workspace. HowToGalaxy
has no equivalent local simulators, so every affected tutorial
exits 2 (file not found) during the subprocess auto-simulate step.

Retargeting to the existing `scripts/simulators/sersic.py` was
considered and rejected — its `simple__sersic` dataset is bulge-only
while several tutorials (e.g. `tutorial_3_realism_and_complexity.py`)
fit a bulge+disk model. Mirroring the matching workspace simulators
preserves model↔data integrity.

Two new files (verbatim copies):
- `scripts/simulators/simple.py` ← `autogalaxy_workspace/scripts/imaging/simulator.py`
- `scripts/simulators/sersic_x2.py` ← `autogalaxy_workspace/scripts/guides/plot/simulator.py`

11 mechanical one-line subprocess path edits in the affected tutorials.
No tutorial logic changes; the dataset names already used by the
tutorials (`simple`, `sersic_x2`) now match what the new local
simulators write.

Verified locally:
- `python scripts/simulators/simple.py` writes
  `dataset/imaging/simple/{data.fits, noise_map.fits, psf.fits, ...}`.
- `python scripts/simulators/sersic_x2.py` writes
  `dataset/imaging/sersic_x2/...`.
- `python scripts/chapter_1_introduction/tutorial_3_fitting.py` runs
  to completion under `PYAUTO_TEST_MODE=2`.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>