Summary

This PR fixes a persistent CI failure in tests/examples/test_examples_viz.py, specifically affecting the test_pd_grid_model test on the chromium runner.

The failure was caused by a Playwright Strict Mode Violation, where the test runner found multiple images matching the selector when only one was expected.

The Problem

The CI logs reported the following error:

Error: strict mode violation: locator("img") resolved to 2 elements:
    1) <img ... /> aka get_by_role("img").first
    2) <img ... /> aka get_by_role("img").nth(1)

This occurs because the test environment renders multiple visualization components (the Grid view and the Chart view) on the same page. When the test attempts to take a screenshot using page_session.locator("img"), Playwright finds both images and throws an error because the selector is ambiguous.

The Fix

I have updated the run_model_test function to strictly define which image to capture:

• Grid Capture: Changed .locator("img") to .locator("img").first.

  • Reasoning: The Grid visualization is rendered first at the top of the page.

• Graph/Chart Capture: Changed .locator("img") to .locator("img").last.

  • Reasoning: The Chart visualization is appended to the page later (when measure_config is present), so it corresponds to the last image element.

This change ensures the selectors are unique, resolving the ambiguity and preventing the test from crashing due to rendering timing or multiple elements