[kavyalavti]

Problem

ImageOverlay and folium.utilities.image_to_url document image as accepting an array-like object, but only real numpy.ndarray instances work. A nested list or tuple crashes with AttributeError: 'list' object has no attribute 'replace'.

Reproduce

import folium

m = folium.Map()
folium.raster_layers.ImageOverlay([[0.0, 1.0], [1.0, 0.0]], bounds=[[0, 0], [1, 1]]).add_to(m)

Fix

image_to_url chose the PNG branch with "ndarray" in image.__class__.__name__, so every other non-string input fell through to the JSON round-trip branch, which returns the list unchanged before .replace is called on it. The PNG branch now handles every non-string input. branca.utilities.write_png already accepts list-like data. Strings still take the same paths as before: file paths are embedded, and URLs are passed through.

Tests

• Added test_image_to_url_array_like, which checks that lists and tuples (mono and RGB) produce the same data URL as the equivalent np.array. It fails on main and passes with the fix.
• Added test_image_to_url_url to cover URL pass-through.
• pytest tests --ignore=tests/selenium --ignore=tests/snapshots: 400 passed. The 3 failures in tests/test_repr.py also fail on main locally because they need a Chrome webdriver.
• pre-commit hooks (ruff, black, codespell, etc.) and mypy pass.

AI disclosure: this change was prepared with the help of Claude Code (Anthropic). I reviewed the diff and test results.

Fixes #2343

🤖 Generated with Claude Code