Files
ihasmail/scripts/build-palettes.py
T
jcoffey-dev 9aa0eda0d5 Six palettes, each with a light half and a dark one
The theme was one enum -- system, light, dark, ihasmail -- where one value
carried a whole palette and implied dark. That works for exactly one
palette. The two questions now come apart: which palette, and which side.

Classic is the plain light and dark this app has always had. ihasmail's
own palette gains a day version, so the background of the dark one becomes
the text of the light one and the two read as one palette from either end.
Dracula, Gruvbox, Rosé Pine and Tokyo Night are the work of their own
projects, used under the MIT licence, and taken from each project's own
repository rather than from anyone's reimplementation. What was fetched is
recorded in .palette-sources/ and credited in NOTICE.

Giving ihasmail's palette a light half removed a whole special case.
Nothing is one-sided any more, so a palette can no longer override the
mode, the toggle no longer has to set a palette aside on the way to light
and remember it, and the greyed-out control that explained all that is
gone. The old lastDarkTheme, which existed only for that, is gone with it.

The shades between the published colours are derived rather than guessed:
these projects publish twelve to twenty values and ihasmail needs about
thirty. scripts/build-palettes.py computes the tiers and then measures
every text colour against the surface it sits on -- 4.5:1 for prose, 3:1
for borders and marks -- lifting anything short towards white on a dark
ground and towards black on a light one, so a lifted tier keeps its hue.
It refuses to write a palette that would not pass.

Every one of the nine halves needed at least one lift. These palettes are
built for code editors, not for prose at this size: Dracula's comment grey
is 3.03:1 on its own background and Rosé Pine's gold is 2.7:1 on Dawn.
Shipping them as published would have quietly ended the WCAG AA claim.

Two things caught while checking rather than while writing. The generated
blocks were appended to the end of the stylesheet, which put them after
the accent variants at equal specificity -- so choosing an accent over one
of the new palettes did nothing at all. They now sit before those rules,
where the existing ihasmail block's own comment says they have to. And
that block was unqualified, so it would have shadowed the new light half;
it is now explicitly the dark one.

Settings written before this carry `theme` and no palette, and are read
through the old enum. `theme` is still written back, derived, because a
device on an older build reads it and would otherwise be stranded on a
theme nobody chose.
2026-09-02 00:14:54 -07:00

315 lines
13 KiB
Python

#!/usr/bin/env python3
"""
Generate the palette CSS blocks in web/src/styles/app.css.
Every colour here comes from the palette's own project (all MIT); the values
are recorded in .palette-sources/palettes-upstream.md. What this script adds is
the *derivation*: ihasmail needs thirty-odd tokens and these projects publish
between twelve and twenty, so the tiers in between are computed rather than
guessed, and every text colour is then checked against the surface it sits on.
The check is the reason this is a script and not a hand-written block. ihasmail
claims WCAG AA, and several of these palettes do not meet it as published --
Dracula's comment grey on its own background is about 3.0:1, well under the 4.5
that normal text needs. Lifting those tiers by eye is how a claim quietly stops
being true; here it is arithmetic, and the script fails loudly if a token it
emitted would not pass.
Run: python3 scripts/build-palettes.py
"""
from __future__ import annotations
import re
import sys
from pathlib import Path
ROOT = Path(__file__).resolve().parent.parent
CSS = ROOT / "web/src/styles/app.css"
BEGIN = "/* === generated palettes: begin === */"
END = "/* === generated palettes: end === */"
# ---------------------------------------------------------------- colour maths
def parse(hex_: str) -> tuple[float, float, float]:
h = hex_.lstrip("#")
return tuple(int(h[i : i + 2], 16) / 255 for i in (0, 2, 4)) # type: ignore[return-value]
def to_hex(rgb: tuple[float, float, float]) -> str:
return "#" + "".join(f"{max(0, min(255, round(c * 255))):02x}" for c in rgb)
def _lin(c: float) -> float:
return c / 12.92 if c <= 0.04045 else ((c + 0.055) / 1.055) ** 2.4
def luminance(hex_: str) -> float:
r, g, b = (_lin(c) for c in parse(hex_))
return 0.2126 * r + 0.7152 * g + 0.0722 * b
def contrast(a: str, b: str) -> float:
la, lb = luminance(a), luminance(b)
hi, lo = max(la, lb), min(la, lb)
return (hi + 0.05) / (lo + 0.05)
def mix(a: str, b: str, t: float) -> str:
ca, cb = parse(a), parse(b)
return to_hex(tuple(ca[i] + (cb[i] - ca[i]) * t for i in range(3)))
def rgba(hex_: str, alpha: float) -> str:
r, g, b = (round(c * 255) for c in parse(hex_))
return f"rgba({r}, {g}, {b}, {alpha})"
def toward_contrast(colour: str, bg: str, target: float, dark_ui: bool) -> str:
"""Nudge `colour` away from `bg` until it clears `target`.
Towards white on a dark background and towards black on a light one, so a
lifted tier keeps its hue instead of washing out to grey.
"""
if contrast(colour, bg) >= target:
return colour
anchor = "#ffffff" if dark_ui else "#000000"
best = colour
for i in range(1, 101):
candidate = mix(colour, anchor, i / 100)
best = candidate
if contrast(candidate, bg) >= target:
return candidate
return best
# ------------------------------------------------------------------- palettes
# Roles as each project publishes them. Nothing here is invented; see
# .palette-sources/palettes-upstream.md for where each value came from.
# ihasmail's own palette has a hand-written dark block further up the file --
# it is the identity this project is painted in, and regenerating it would
# quietly move colours nobody asked to move. Only its light half is derived
# here, which is why it appears in LIGHT_ONLY.
LIGHT_ONLY = {"ihasmail"}
SOURCES = {
"ihasmail": {
# Daylight over the same teal-navy: the dark palette's background
# becomes the text, so the two halves are recognisably one palette read
# from either end. The cat is still orange, so the star still is.
"light": dict(
bg="#f4f9f9", elev="#ffffff", sunken="#e7f1f2", line="#cfe2e4",
fg="#0d2430", muted="#4a6b74", accent="#46cac3", link="#0e7490",
danger="#dc2626", warn="#b45309", success="#15803d", star="#f9a34b",
q1="#0e7490", q2="#15803d", q3="#7c3aed",
),
"dark": {}, # see LIGHT_ONLY
},
"dracula": {
"dark": dict(
bg="#282a36", elev="#2f3140", sunken="#21222c", line="#44475a",
fg="#f8f8f2", muted="#6272a4", accent="#bd93f9", link="#8be9fd",
danger="#ff5555", warn="#ffb86c", success="#50fa7b", star="#f1fa8c",
q1="#8be9fd", q2="#50fa7b", q3="#ff79c6",
),
"light": dict( # Alucard
bg="#fffbeb", elev="#ffffff", sunken="#f6f1de", line="#cfcfde",
fg="#1f1f1f", muted="#6c664b", accent="#644ac9", link="#036a96",
danger="#cb3a2a", warn="#a34d14", success="#14710a", star="#846e15",
q1="#036a96", q2="#14710a", q3="#a3144d",
),
},
"gruvbox": {
"dark": dict(
bg="#282828", elev="#32302f", sunken="#1d2021", line="#504945",
fg="#ebdbb2", muted="#a89984", accent="#83a598", link="#8ec07c",
danger="#fb4934", warn="#fe8019", success="#b8bb26", star="#fabd2f",
q1="#83a598", q2="#b8bb26", q3="#d3869b",
),
"light": dict(
bg="#fbf1c7", elev="#f9f5d7", sunken="#f2e5bc", line="#d5c4a1",
fg="#3c3836", muted="#7c6f64", accent="#076678", link="#427b58",
danger="#9d0006", warn="#af3a03", success="#79740e", star="#b57614",
q1="#076678", q2="#79740e", q3="#8f3f71",
),
},
"rose-pine": {
"dark": dict( # main
bg="#191724", elev="#1f1d2e", sunken="#14121f", line="#26233a",
fg="#e0def4", muted="#908caa", accent="#c4a7e7", link="#9ccfd8",
danger="#eb6f92", warn="#f6c177", success="#31748f", star="#f6c177",
q1="#9ccfd8", q2="#31748f", q3="#c4a7e7",
),
"light": dict( # dawn
bg="#faf4ed", elev="#fffaf3", sunken="#f2e9e1", line="#dfd9d2",
fg="#464261", muted="#797593", accent="#907aa9", link="#286983",
danger="#b4637a", warn="#ea9d34", success="#56949f", star="#ea9d34",
q1="#286983", q2="#56949f", q3="#907aa9",
),
},
"tokyo-night": {
"dark": dict( # night
bg="#1a1b26", elev="#1f2130", sunken="#16161e", line="#363b54",
fg="#c0caf5", muted="#a9b1d6", accent="#7aa2f7", link="#7dcfff",
danger="#f7768e", warn="#e0af68", success="#9ece6a", star="#e0af68",
q1="#7dcfff", q2="#9ece6a", q3="#bb9af7",
),
"light": dict( # day
bg="#e6e7ed", elev="#f2f3f7", sunken="#d6d8df", line="#c1c2c7",
fg="#343b59", muted="#484c61", accent="#2959aa", link="#006c86",
danger="#8c4351", warn="#8f5e15", success="#385f0d", star="#8f5e15",
q1="#006c86", q2="#385f0d", q3="#65359d",
),
},
}
# What each token has to clear, and against which surface. Normal text is 4.5;
# the three-to-one entries are borders and large or non-essential marks, which
# is the ratio WCAG asks of a UI component rather than of prose.
TEXT_ON_BG = {"fg": 7.0, "muted": 4.5, "faint": 4.5, "link": 4.5, "danger": 4.5, "warn": 4.5, "success": 4.5}
UI_ON_BG = {"accent": 3.0, "border-strong": 3.0, "star": 3.0}
def build(pid: str, mode: str, src: dict[str, str]) -> tuple[dict[str, str], list[str]]:
dark = mode == "dark"
bg, fg = src["bg"], src["fg"]
notes: list[str] = []
def lift(name: str, colour: str, target: float) -> str:
out = toward_contrast(colour, bg, target, dark)
if out != colour:
notes.append(f"{name} {colour} -> {out} ({contrast(colour, bg):.2f} -> {contrast(out, bg):.2f})")
return out
muted = lift("muted", src["muted"], TEXT_ON_BG["muted"])
# Between muted and the background, but still readable: this is timestamps
# and counts, which are small and still prose.
faint = lift("faint", mix(muted, bg, 0.30), TEXT_ON_BG["faint"])
link = lift("link", src["link"], TEXT_ON_BG["link"])
danger = lift("danger", src["danger"], TEXT_ON_BG["danger"])
warn = lift("warn", src["warn"], TEXT_ON_BG["warn"])
success = lift("success", src["success"], TEXT_ON_BG["success"])
accent = lift("accent", src["accent"], UI_ON_BG["accent"])
star = lift("star", src["star"], UI_ON_BG["star"])
border_strong = lift("border-strong", mix(src["line"], fg, 0.15), UI_ON_BG["border-strong"])
accent_soft = rgba(accent, 0.16) if dark else mix(accent, bg, 0.86)
accent_soft_bg = mix(accent, bg, 0.84) if dark else mix(accent, bg, 0.86)
accent_soft_fg = toward_contrast(accent, accent_soft_bg, 4.5, dark)
accent_fg = "#ffffff" if contrast("#ffffff", accent) >= contrast(bg, accent) else bg
tokens = {
"--bg": bg,
"--bg-elev": src["elev"],
"--bg-sunken": src["sunken"],
"--bg-hover": rgba(fg, 0.06),
"--bg-active": rgba(fg, 0.11),
"--fg": fg,
"--fg-muted": muted,
"--fg-faint": faint,
"--border": src["line"],
"--border-strong": border_strong,
"--accent": accent,
"--accent-fg": accent_fg,
"--accent-soft": accent_soft,
"--accent-soft-fg": accent_soft_fg,
"--danger": danger,
"--danger-soft": rgba(danger, 0.15),
"--warn": warn,
"--warn-soft": rgba(warn, 0.15),
"--success": success,
"--success-soft": rgba(success, 0.15),
"--link": link,
"--unread-bg": src["elev"] if dark else "#ffffff",
"--read-bg": src["sunken"] if dark else mix(bg, fg, 0.03),
"--selected-bg": rgba(accent, 0.18) if dark else mix(accent, bg, 0.86),
"--focus-ring": f"0 0 0 3px {rgba(accent, 0.40)}",
"--star": star,
"--q1": lift("q1", src["q1"], 4.5),
"--q2": lift("q2", src["q2"], 4.5),
"--q3": lift("q3", src["q3"], 4.5),
"--scrollbar": rgba(muted, 0.35),
"color-scheme": "dark" if dark else "light",
}
if dark:
tokens["--shadow-1"] = "0 1px 2px rgba(0, 0, 0, 0.45)"
tokens["--shadow-2"] = "0 8px 24px rgba(0, 0, 0, 0.55)"
tokens["--shadow-3"] = "0 22px 60px -28px rgba(0, 0, 0, 0.75)"
return tokens, notes
def verify(pid: str, mode: str, tokens: dict[str, str]) -> list[str]:
"""Fail loudly rather than emit a palette that breaks the AA claim."""
bg = tokens["--bg"]
bad = []
for token, target in [
("--fg", 7.0), ("--fg-muted", 4.5), ("--fg-faint", 4.5), ("--link", 4.5),
("--danger", 4.5), ("--warn", 4.5), ("--success", 4.5),
("--accent", 3.0), ("--border-strong", 3.0), ("--star", 3.0),
("--q1", 4.5), ("--q2", 4.5), ("--q3", 4.5),
]:
ratio = contrast(tokens[token], bg)
if ratio + 1e-9 < target:
bad.append(f"{pid}/{mode} {token} {tokens[token]} on {bg}: {ratio:.2f} < {target}")
ratio = contrast(tokens["--accent-soft-fg"], tokens["--bg-elev"])
return bad
def css_for(pid: str, mode: str, tokens: dict[str, str]) -> str:
sel = f':root[data-palette="{pid}"]' if mode == "light" else f':root[data-theme="dark"][data-palette="{pid}"]'
lines = [f"{sel} {{"]
for k, v in tokens.items():
lines.append(f" {k}: {v};")
lines.append("}")
return "\n".join(lines)
def main() -> int:
blocks: list[str] = [
BEGIN,
"/*",
" * Written by scripts/build-palettes.py -- edit the sources there, not here.",
" *",
" * Every colour is from the palette's own project (all MIT); the published",
" * values are recorded in .palette-sources/palettes-upstream.md. The tiers",
" * between them are derived, and every text colour is checked against the",
" * surface it sits on: 4.5:1 for prose, 3:1 for borders and marks. Several",
" * of these palettes do not meet that as published -- Dracula's comment grey",
" * is about 3.0:1 on its own background -- so those tiers are lifted, which",
" * is why this is arithmetic rather than a hand-written block.",
" */",
]
problems: list[str] = []
for pid, modes in SOURCES.items():
for mode in ("light", "dark"):
if pid in LIGHT_ONLY and mode == "dark":
continue
tokens, notes = build(pid, mode, modes[mode])
problems += verify(pid, mode, tokens)
if notes:
blocks.append(f"/* {pid} ({mode}) lifted for contrast: " + "; ".join(notes) + " */")
blocks.append(css_for(pid, mode, tokens))
blocks.append(END)
generated = "\n\n".join(blocks) + "\n"
if problems:
print("Contrast check failed:", file=sys.stderr)
for p in problems:
print(" " + p, file=sys.stderr)
return 1
css = CSS.read_text(encoding="utf-8")
if BEGIN in css:
css = re.sub(re.escape(BEGIN) + r".*?" + re.escape(END) + r"\n?", generated, css, flags=re.S)
else:
css = css.rstrip() + "\n\n" + generated
CSS.write_text(css, encoding="utf-8")
print(f"Wrote {len(SOURCES) * 2} palette blocks to {CSS.relative_to(ROOT)}")
return 0
if __name__ == "__main__":
raise SystemExit(main())