feat: export subplot_fit_interferometer_dirty_images (autolens's own)

[Jammy2211]

Summary

Follow-up 4 from #667 — a live API defect.

aplt.subplot_fit_dirty_images resolves to the autogalaxy implementation inside autolens.plot:

>>> aplt.subplot_fit_dirty_images.__module__
'autogalaxy.interferometer.plot.fit_interferometer_plots'

autolens has its own version in autolens/interferometer/plot/fit_interferometer_plots.py, and it was reachable under no exported name at all.

That version is not merely a duplicate. When image_plane_lines is None it auto-derives the tracer's critical curves from the fit and overlays them on the dirty model image — lensing information the autogalaxy version cannot produce. Lens users calling aplt.subplot_fit_dirty_images were silently getting the galaxy version and losing that overlay.

API Changes

One symbol added: subplot_fit_interferometer_dirty_images, bound to autolens's own function. Purely additive — nothing removed, renamed, or resignatured.

subplot_fit_dirty_images is deliberately left bound to the autogalaxy version. The two signatures diverge (residuals_symmetric_cmap on the AG side, image_plane_lines / image_plane_line_colors on the AL side), so rebinding it would be a behaviour change, not an additive export.

This matches the convention already in the file: subplot_fit_real_space (autogalaxy) coexists with subplot_fit_interferometer_real_space (autolens's own). subplot_fit_dirty_images had simply never been given its subplot_fit_interferometer_* counterpart.

See full details below.

Test Plan

• Full PyAutoLens suite: 488 passed, 0 failed (55.7s)
• Additivity asserted programmatically against the branch build — new name resolves to autolens.…, old name still resolves to autogalaxy.…:
  • subplot_fit_interferometer_dirty_images → autolens.interferometer.plot.fit_interferometer_plots
  • subplot_fit_dirty_images → autogalaxy.interferometer.plot.fit_interferometer_plots (unchanged)
• Functionally exercised on a real FitInterferometer, not just signature-checked: the new export renders and auto-derives 1 critical curve via _compute_critical_curve_lines; the old export still renders unchanged.
• git diff --stat = 2 files, 2 insertions, 0 deletions

Full API Changes (for automation & release notes)

Added

• autolens.plot.subplot_fit_interferometer_dirty_images(fit, output_path=, output_format=, colormap=, use_log10=False, image_plane_lines=, image_plane_line_colors=, title_prefix=) — 2×3 subplot of dirty-image diagnostics for an interferometer fit, overlaying the tracer's critical curves on the dirty model image. Auto-derives the curves from the fit when image_plane_lines is None. Re-export of autolens.interferometer.plot.fit_interferometer_plots.subplot_fit_dirty_images.

Removed / Renamed / Changed Signature / Changed Behaviour

None. subplot_fit_dirty_images continues to resolve to the autogalaxy implementation exactly as before.

Migration

None required. Lens users wanting the critical-curve overlay switch call sites:

• Before: aplt.subplot_fit_dirty_images(fit=fit) — galaxy version, no overlay
• After: aplt.subplot_fit_interferometer_dirty_images(fit=fit) — overlay auto-derived

Deferred: the 11 existing workspace call sites

autolens_workspace calls aplt.subplot_fit_dirty_images at 11 sites (scripts/interferometer/{fit,modeling,plot}.py, four interferometer/features/*/fit.py, and multi/features/imaging_and_interferometer/modeling.py). All pass fit=fit only — none pass residuals_symmetric_cmap — so all 11 could be switched to the new name and would gain critical-curve overlays for free.

Not done here: that is a visual change across 11 lens examples and deserves its own review, separate from making the function reachable. Filed against #667.

Generated by the PyAutoLabs agent workflow.