[anyingiit]

Link to issue

Closes #2059

Description of change

With a continuous color, marginal traces are colored with args["color_continuous_scale"][0]. For a named scale like "Viridis" that is the first character ('V'), and for a list of [position, color] pairs it is the pair itself, so px.scatter(..., color=<numeric>, marginal_x="box", color_continuous_scale="Viridis") raises an invalid marker.color error. This PR normalizes the scale with ColorscaleValidator (as is already done for the coloraxis) before taking its first color.

Motivation / disclosure: I had some spare AI-assistant quota (Claude Code) and am using it to try to help projects with open issues. The change was prepared with Claude Code and verified as listed below. If it doesn't fit, or you'd simply rather not take it, please feel free to close it, no hard feelings at all 🙂

Demo

import plotly.express as px
df = px.data.iris()
fig = px.scatter(df, x="sepal_width", y="sepal_length", color="sepal_length",
                 marginal_x="box", color_continuous_scale="Viridis")
fig.data[1].marker.color  # before: ValueError (received 'V'); after: '#440154'

Testing strategy

Added test_marginal_color_with_continuous_color_scale in tests/test_optional/test_px/test_marginals.py (named, lowercase, reversed _r, list of colors and list of pairs × histogram/box/violin/rug). 16 of its 20 cases fail on main (the list-of-colors cases already worked) and all pass with this change.

• python -m pytest tests/test_optional/test_px --ignore=.../test_imshow.py --ignore=.../test_trendline.py: 1083 passed, 3 failed (test_render_mode, which needs statsmodels, not installed locally)
• ruff format --check . and ruff check on the changed files: clean

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).