fix: unpark ellipse example scripts after JAX refactor

[Jammy2211]

Summary

Reactivate the five scripts/ellipse/* example scripts that were parked on 2026-04-24 in config/build/no_run.yaml pending the ellipse JAX refactor in PyAutoGalaxy. That refactor has now shipped (PyAutoGalaxy #408 ellipse-xp, #410 fit-ellipse-jax, #412 analysis-ellipse-jax, all merged 2026-05-14), and all five scripts (simulator, fit, modeling, multipoles, database) now pass cleanly under PYAUTO_TEST_MODE=2.

Also patches a leftover path.exists(...) / os.remove(...) pair in scripts/ellipse/database.py — the prior os.path → pathlib refactor removed the path (i.e. os.path) import but missed two call sites, leaving the script broken on import. Replaced with the pathlib idiom (Path(...) / name).exists() and .unlink().

Surfaced during this work, and shipped as a paired upstream fix: PyAutoFit #1270 (Drawer.__init__ from_dict round-trip crash on duplicate number_of_cores), which the aggregator scrape in database.py was hitting on every saved Drawer fit.

Scripts Changed

• config/build/no_run.yaml — remove the five ellipse/{simulator,fit,modeling,multipoles,database} NEEDS_FIX 2026-04-24 entries
• scripts/ellipse/database.py — replace 2× path.exists(...) / os.remove(...) (leftover from the os.path → pathlib refactor; path is no longer imported) with Path(...).exists() / .unlink()

Upstream PR

• fix: dedupe number_of_cores in Drawer for from_dict round-trip PyAutoFit#1270 — fix: dedupe number_of_cores in Drawer for from_dict round-trip (paired Drawer fix that database.py exposed during verification)

Test Plan

• All five ellipse scripts run cleanly under PYAUTO_TEST_MODE=2 end-to-end in the worktree
• database.py aggregator scrape of fit_all Drawer outputs now succeeds with PyAutoFit #1270 applied
• Smoke tests pass for autogalaxy_workspace

🤖 Generated with Claude Code