Avoid NaN saturation when converting black from HSL to HSB

[Chessing234]

Bug

ColorConversion._hslaToHSBA returns a NaN saturation when the input
is pure black. Any p5.Color constructed from an HSL black
(e.g. color('hsl(0, 50%, 0%)')) ends up with sat = NaN once it's
converted to the internal HSBA representation, and downstream
getters/serializers propagate that NaN.

Root cause

let val;
if (li < 0.5) {
  val = (1 + sat) * li;
} else {
  val = li + sat - li * sat;
}

// Convert saturation.
sat = 2 * (val - li) / val;

For black, li === 0, and the first branch sets val = (1 + sat) * 0 = 0.
The subsequent 2 * (val - li) / val becomes 2 * 0 / 0, which is
NaN.

Why the fix is correct

• Pure black carries no hue and no meaningful saturation, so the
  conventional value is 0 - the same choice _rgbaToHSBA already
  makes for grayscale (chroma === 0) inputs elsewhere in the file.
• Gating the division on val === 0 with a ternary keeps the
  existing formula unchanged for every non-black input (floating
  point val is only ever exactly 0 when li === 0, so the
  existing val > 0 code path behaviour is identical).
• The return shape is the same; only the NaN output becomes 0.

Change

src/color/color_conversion.js: replace

sat = 2 * (val - li) / val;

with a ternary guard returning 0 when val === 0. One-line fix
(plus a brief comment).

[ksen0]

This does not link to an open bug report, so I will close it. Please review contribution guidelines before making future PRs. Thank you!