Plots: subplot_fit_interferometer (zoomed panel)

[aparnavenkat27]

The right-hand side of the image attached below shows a fit image generated while PyAutolens was running. The zoom-to-brightest-pixel subplot appears to move the source to a different position relative to the caustic even though the code does not do it (the left hand side is the source code from my version of PyAutolens, v2026.9.19.1. The tracer image generated matches the no-zoom subplot. Is the zoom plot moving the image but not the position of the caustic lines?

[Jammy2211]

Hi @aparnavenkat27, thanks for the report and the screenshot, and sorry for the slow reply. I have checked the zoom path on current main: for a parametric source the zoomed panel re-evaluates the source on a grid centred on its brightest region and carries that grid's origin into the plotted extent, and the caustic is drawn in absolute arcsec on the same axes, so on main the two should agree. I could not reproduce an offset from the figure alone, so a few questions would help pin it down:

1. Is the source a light profile or a pixelization? The zoomed panel takes a different path for each.
2. What is imshow_origin set to in your config/visualize/general.yaml? Under lower there is a known remaining quirk in the zoom window (fix: make plot_array overlays respect imshow_origin PyAutoArray#565).
3. Could you paste the mass centre of the lens and the centre of the source from model.results (or the tracer you plotted)? If the lens is offset from (0, 0) the two panels can look inconsistent while being right.
4. Does it reproduce on 2026.10.18.1? The interferometer subplot changed once since 2026.9.19.1.

If you can share a minimal script (or the tracer.json plus real_space_mask) I will run it directly.