Euclidean & probability rhythms per part

[philoking]

Summary

Per-part Euclidean and trigger-probability rhythm options: distribute K hits across N steps as evenly as possible, and/or give each step a chance of firing. Applied as a rhythmic mask over whatever the part already generates.

Why it fits

"K of N" Euclidean patterns and per-step probability are the most modular-native rhythm tools there are, and they pair perfectly with per-part clock divide — a 3-of-8 arp at 1/8 against a 5-of-16 bass at 1/16 is instant generative interplay. It's also a tiny, well-bounded addition that adds a lot of life.

Proposed scope

Rhythm mask (js/generator.js)

• Per part: euclid: { k, n } (off when unset) producing a boolean trigger mask via the standard Bjorklund/Bresenham distribution, with a rotation offset.
• Per part: triggerProb (0–1) — each otherwise-active step fires with this probability, using the seeded PRNG so it's reproducible.
• The mask gates the part's existing note stream: a masked-out step is a rest, and the note pool advances using the same rest semantics as the step editor, so the two features agree.

UI (js/app.js)

• In each part's ⚙: K, N, rotation, and a probability knob. Show the resulting pattern as a small dotted readout.
• Persist in state.

Where

js/generator.js (mask generation + application), js/app.js (controls + state).

Done when

• Each part can take a K-of-N Euclidean mask with rotation.
• Per-step trigger probability works and is seed-deterministic.
• Masks compose cleanly with per-part clock divide.
• Settings persist in presets / JSON.

Open questions

• Should the Euclidean mask count in the part's own divided steps, or absolute 16ths?
• Default to off for all parts so existing songs are unchanged?