Self-hosted, Self-Improving, Federated Everything App — dev management, agent orchestration, persistent mind, digital twin, second brain, and creative pipeline from a single dashboard over Tailscale to port the user into the OS
JavaScript ai-agentsdashboarddeveloper-toolsdigital-twinquantified-selfsecond-brainself-hostedtailscale
import { describe, expect, it } from 'vitest';
import { PORTOS_SCHEMA_VERSIONS } from '../../lib/schemaVersions.js';
import { normalizeComposition } from './composition.js';
import { buildCodeTimeline, buildSongDocument, quantizeSongDuration } from './codeTimeline.js';
const fixtureSectionSource = (color) => `function render(ctx, env) {\n ctx.fillStyle = ${JSON.stringify(color)};\n ctx.fillRect(env.safe.x, env.safe.y, 12 + (env.frame % 3), 12);\n}`;
const project = {
name: 'Click',
audioAnalysis: {
durationSec: 4,
beats: [0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5],
downbeats: [0, 2],
sections: [
{ label: 'Verse', startSec: 0, endSec: 2 },
{ label: 'Chorus', startSec: 2, endSec: 4 },
],
},
lyricCues: [
{ id: 'l1', text: 'hello world', startSec: 0.55, endSec: 1.4 },
{ id: 'l2', text: 'second line', startSec: 2.2, endSec: 3.2, words: [
{ text: 'second', startSec: 2.25, endSec: 2.6 },
{ text: 'line', startSec: 2.7, endSec: 3.1 },
] },
],
scenes: [{ sceneId: 's1', label: 'Shot', startSec: 0, endSec: 4, takes: [{ id: 'take-1' }] }],
composition: { mode: 'concat', textCues: [{ id: 'c1', text: 'kept', startSec: 1, endSec: 2 }], style: { color: '#ffffff', font: 'sans' } },
};
describe('buildCodeTimeline (#9076)', () => {
it('snaps later section boundaries to the beat at or before the first word', () => {
const timeline = buildCodeTimeline(project);
expect(timeline.sections.map((section) => [section.label, section.startSec, section.endSec])).toEqual([
['Verse', 0, 2],
['Chorus', 2, 4],
]);
for (const section of timeline.sections.slice(1)) {
expect(timeline.beats).toContain(section.startSec);
}
const song = buildSongDocument(project, timeline);
const chorus = song.lyrics.find((line) => line.id === 'l2');
expect(chorus.words.map((word) => word.text)).toEqual(['second', 'line']);
// No word timings: the line itself is the only cue, not invented word times.
expect(song.lyrics.find((line) => line.id === 'l1').words).toEqual([
{ text: 'hello world', startSec: 0.55, endSec: 1.4 },
]);
});
it('snaps to the aligned word, which is stored as w', () => {
const aligned = {
...project,
audioAnalysis: {
...project.audioAnalysis,
sections: [
{ label: 'Verse', startSec: 0, endSec: 1.2 },
{ label: 'Chorus', startSec: 1.2, endSec: 4 },
],
},
lyricCues: [{
id: 'l2', text: 'second line', startSec: 1.25, endSec: 3.2,
words: [
{ w: 'second', startSec: 1.8, endSec: 2.1, conf: 'matched' },
{ w: 'line', startSec: 2.2, endSec: 2.8, conf: 'matched' },
],
}],
};
const timeline = buildCodeTimeline(aligned);
// The line starts at 1.25 (beat 1). The first sung word is at 1.8, so the
// cut is the beat at or before that word.
expect(timeline.sections[1].startSec).toBe(1.5);
const song = buildSongDocument(aligned, timeline);
expect(song.lyrics[0].words.map((word) => word.text)).toEqual(['second', 'line']);
});
it('uses the nearest downbeat when a section has no lyric', () => {
const instrumental = {
...project,
lyricCues: [],
audioAnalysis: { ...project.audioAnalysis, sections: [
{ label: 'Intro', startSec: 0, endSec: 2.2 },
{ label: 'Drop', startSec: 2.2, endSec: 4 },
] },
};
const timeline = buildCodeTimeline(instrumental);
expect(timeline.sections.map((section) => section.startSec)).toEqual([0, 2]);
expect(timeline.sections[1].snapKind).toBe('downbeat');
});
it('keeps a section when its last lyric snaps onto the following boundary', () => {
const lateLyric = {
...project,
audioAnalysis: { durationSec: 30, beats: [0, 10, 20], downbeats: [0, 10, 20], sections: [
{ id: 'intro', startSec: 0, endSec: 10 },
{ id: 'verse', startSec: 10, endSec: 20 },
{ id: 'hook', startSec: 20, endSec: 30 },
] },
lyricCues: [{ id: 'last-line', text: 'on the cut', startSec: 19.8, endSec: 20.1,
words: [{ w: 'on', startSec: 20 }] }],
};
expect(buildCodeTimeline(lateLyric).sections.map((section) => [section.id, section.startSec])).toEqual([
['intro', 0], ['verse', 10], ['hook', 20],
]);
});
it('prefers treatment windows over analyzed sections', () => {
const directed = {
...project,
treatment: { arc: { beats: [
{ id: 'b1', label: 'Open', startSec: 0, endSec: 1.2 },
{ id: 'b2', label: 'Payoff', startSec: 1.2, endSec: 4 },
] } },
};
const timeline = buildCodeTimeline(directed);
expect(timeline.sections.map((section) => section.id)).toEqual(['b1', 'b2']);
expect(timeline.beats).toContain(timeline.sections[1].startSec);
});
it('covers the song within one frame', () => {
for (const duration of [1 / 24, 2, 2.01, 3.333, 10]) {
const quantized = quantizeSongDuration(duration, 24);
expect(quantized.durationSec).toBeGreaterThanOrEqual(duration - 1e-9);
expect(quantized.durationSec - duration).toBeLessThan(1 / 24 + 1e-9);
expect(Math.round(quantized.durationSec * 24)).toBe(quantized.frames);
}
});
});
describe('code mode keeps the rest of the project (#9076)', () => {
it('defaults unknown modes to concat and round-trips code without dropping cues', () => {
expect(normalizeComposition({ mode: 'not-a-mode', textCues: [{ text: 'Hi', startSec: 1, endSec: 2 }] }).mode).toBe('concat');
const source = fixtureSectionSource('#112233');
const code = normalizeComposition({
mode: 'code',
textCues: project.composition.textCues,
codeVideo: { providerId: 'stub', model: 'fixture', sections: [{ id: 'Verse', source }, { id: 'bad', source: 'Math.random()' }] },
});
expect(code.mode).toBe('code');
expect(code.textCues.map((cue) => cue.text)).toEqual(['kept']);
expect(code.codeVideo.sections).toEqual([{ id: 'Verse', source }]);
expect(normalizeComposition({ ...code, mode: 'concat' }).textCues).toHaveLength(1);
expect(normalizeComposition({ ...code, mode: 'composed' }).codeVideo.sections).toHaveLength(1);
expect(PORTOS_SCHEMA_VERSIONS.musicVideoProjects).toBeGreaterThanOrEqual(7);
});
});