cvdsafe

Is your palette colorblind-safe? One import to check and fix color palettes for color-vision deficiency — protanopia, deuteranopia, tritanopia.

pip install cvdsafe

import cvdsafe as hue

hue.is_safe(["#d7191c", "#1a9641"])            # False — red and green collapse under deutan/protan
hue.is_safe(["#0072b2", "#e69f00", "#009e73"]) # True  — an Okabe-Ito set stays distinct

hue.check_palette(["#d7191c", "#1a9641"])
# {'pass': False, 'types': {'deutan': {'conflicts': [{'a': '#d7191c', 'b': '#1a9641', ...}]}}}

hue.fix_palette(["#d7191c", "#1a9641"])["colors"]  # a safe palette near the originals
hue.simulate("#d7191c", "deutan")                  # how that red looks to a deuteranope

What it does

• is_safe(colors) → True/False for the whole palette.
• check_palette(colors) → every pair that stays distinct to normal vision but collapses under a deficiency, with the CIEDE2000 difference before and after simulation.
• fix_palette(colors) → an adjusted palette that passes, separated along lightness, staying near the originals.
• simulate(color, type) → the color as a given deficiency sees it.
• check_contrast(fg, bg) → WCAG contrast ratio and AA/AAA pass.

How it decides

cvdsafe is a front door to opticquiz-cvd, which simulates deficiency with the Machado, Oliveira & Fernandes (2009) matrices and measures color difference with CIEDE2000. The method is published open access: https://doi.org/10.5281/zenodo.21310578

Every function here is opticquiz_cvd's, unchanged — the package adds the name and one convenience, so there is never a second copy of the colorimetry to drift.

Not a legal audit

A screening aid for the color-distinguishability axis of accessibility, not a WCAG/ADA audit, and no screen-based check is a clinical diagnosis.

Licence

MIT.