Fix px.imshow drawing the next animation frame in empty facet_col_wrap cells

[breken-ai]

Link to issue

No existing issue that I could find (searched open and closed issues and PRs for imshow facet_col_wrap and imshow animation facet). I'm happy to open one first if you'd like it tracked separately.

Description of change

px.imshow with both animation_frame and facet_col_wrap draws the wrong image when the facets don't fill the wrapped grid. The loop that builds fig.data kept every trace with index < nrows * ncols. With 5 facets wrapped at 2 columns that is 6, so trace 5, which is facet 0 of the second animation frame, went into the empty sixth cell. Frames only update the first 5 traces, so that extra image stays there for the whole animation, with no facet title. The check now uses nslices_facets, the number of facets in one frame.

Demo

import numpy as np
import plotly.express as px

img = np.arange(2 * 5 * 3 * 4).reshape(2, 5, 3, 4).astype(float)  # 2 frames x 5 facets
fig = px.imshow(img, animation_frame=0, facet_col=1, facet_col_wrap=2)
print(len(fig.data), [(t.name, t.xaxis, t.z[0][0]) for t in fig.data][-1])

Before: 6 ('5', 'x2', 60.0). Frame 1's first facet (values starting at 60) is drawn in the bottom-right cell, which should be empty.
After: 5 ('4', 'x', 48.0). Only frame 0's five facets are drawn, and the sixth cell is empty.

Testing strategy

I added test_animation_and_facet_col_wrap_incomplete_grid to tests/test_optional/test_px/test_imshow.py, covering both binary_string=False (Heatmap) and True (Image). Both cases fail on main (6 traces instead of 5) and pass with this change. The full test_imshow.py passes (92 tests), and ruff format --check is clean.

Additional information (optional)

Without animation there is only one frame's worth of traces, so the old bound had no effect. With facet_row, or facet_col without wrapping, nrows * ncols already equals the facet count, so only the wrapped layout with an incomplete last row was affected. I found this bug and wrote the fix with the help of an AI coding assistant, and I checked the change and the test output myself.

Guidelines

• I have reviewed the pull request guidelines and the Code of Conduct and confirm that this PR follows them.
• I have added an entry to the changelog if needed (not required for documentation PRs).