Fix NaN-unsafe vmax computation in log-scale LogNorm

[Copilot]

When an array contains NaNs (e.g., masked pixels), clipped.max() propagates nan into vmax, silently breaking LogNorm and imshow. The existing MatWrap path already uses np.nanmin/np.nanmax for this reason.

Change

• autoarray/plot/plots/array.py: replace clipped.max() with np.nanmax(clipped) when computing vmax for LogNorm in the log-scale branch.

# Before
norm = LogNorm(vmin=vmin or log10_min, vmax=vmax or clipped.max())

# After
norm = LogNorm(vmin=vmin or log10_min, vmax=vmax or np.nanmax(clipped))

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.