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, it, expect, vi, beforeEach } from 'vitest';
// Stub the LLM driver so deriveAudioCues never hits a provider.
const runStagedLLMMock = vi.fn();
vi.mock('../stageRunner.js', () => ({
runStagedLLM: (...a) => runStagedLLMMock(...a),
}));
const {
deriveAudioCues,
preserveRenderedCues,
ERR_NO_SOURCE,
ERR_EMPTY_RESULT,
__testing,
} = await import('./audioCues.js');
describe('audioCues — pure internals', () => {
describe('episodeBeatText', () => {
it('prefers idea.output (beats) over idea.input (synopsis)', () => {
const issue = { stages: { idea: { output: 'BEATS', input: 'SYNOPSIS' } } };
expect(__testing.episodeBeatText(issue)).toBe('BEATS');
});
it('falls back to idea.input when output is empty', () => {
const issue = { stages: { idea: { output: ' ', input: 'SYNOPSIS' } } };
expect(__testing.episodeBeatText(issue)).toBe('SYNOPSIS');
});
it('returns empty string when neither is present', () => {
expect(__testing.episodeBeatText({ stages: { idea: {} } })).toBe('');
expect(__testing.episodeBeatText({})).toBe('');
});
});
describe('renderScenesForPrompt', () => {
it('returns the no-scenes placeholder for an empty list', () => {
expect(__testing.renderScenesForPrompt([])).toMatch(/no storyboard scenes/i);
expect(__testing.renderScenesForPrompt(null)).toMatch(/no storyboard scenes/i);
});
it('numbers scenes and prefers heading + summary', () => {
const out = __testing.renderScenesForPrompt([
{ heading: 'The kitchen', summary: 'Jean waits.' },
{ heading: 'The roof' },
]);
expect(out).toContain('1. The kitchen — Jean waits.');
expect(out).toContain('2. The roof');
});
});
describe('sanitizeDerivedCue', () => {
it('drops a cue with no prompt', () => {
expect(__testing.sanitizeDerivedCue({ label: 'x' }, 0, 'musicgen')).toBeNull();
});
it('builds an un-rendered, un-placed cue with sequential id + default engine', () => {
const cue = __testing.sanitizeDerivedCue({ label: 'Act I', prompt: 'pads' }, 0, 'audioldm2');
expect(cue).toMatchObject({
id: 'cue-001', label: 'Act I', prompt: 'pads', engine: 'audioldm2',
startSec: null, endSec: null, trackFilename: null, durationSec: null, gain: null,
});
});
it('honors a per-cue engine hint over the default', () => {
const cue = __testing.sanitizeDerivedCue({ prompt: 'p', engine: 'suno' }, 3, 'musicgen');
expect(cue.engine).toBe('suno');
expect(cue.id).toBe('cue-004');
});
it('defaults the label when absent', () => {
const cue = __testing.sanitizeDerivedCue({ prompt: 'p' }, 1, 'musicgen');
expect(cue.label).toBe('Cue 2');
});
});
});
describe('preserveRenderedCues', () => {
it('returns fresh cues unchanged when there is no prior rendered audio', () => {
const fresh = [{ id: 'cue-001', label: 'A', trackFilename: null }];
expect(preserveRenderedCues(fresh, [])).toEqual(fresh);
expect(preserveRenderedCues(fresh, [{ label: 'A', trackFilename: null }])).toEqual(fresh);
});
it('carries a rendered track forward onto the matching fresh label (case-insensitive)', () => {
const fresh = [
{ id: 'cue-001', label: 'Act I', prompt: 'new prompt', trackFilename: null, durationSec: null },
{ id: 'cue-002', label: 'Climax', trackFilename: null, durationSec: null },
];
const prior = [{ label: 'act i', trackFilename: 'music-gen-x.wav', durationSec: 42 }];
const merged = preserveRenderedCues(fresh, prior);
expect(merged[0].trackFilename).toBe('music-gen-x.wav');
expect(merged[0].durationSec).toBe(42);
// The fresh prompt is kept — only the render artifacts carry over.
expect(merged[0].prompt).toBe('new prompt');
// A label with no prior match stays un-rendered.
expect(merged[1].trackFilename).toBeNull();
});
});
describe('deriveAudioCues', () => {
beforeEach(() => runStagedLLMMock.mockReset());
const issueWithBeats = {
number: 3, title: 'Ep',
stages: { idea: { output: 'Act 1 setup. Act 2 climax.' }, storyboards: { scenes: [{ heading: 'A' }] } },
};
it('throws PIPELINE_AUDIO_CUES_NO_SOURCE when the episode has no beats', async () => {
await expect(deriveAudioCues({ stages: { idea: {} } })).rejects.toMatchObject({ code: ERR_NO_SOURCE, status: 400 });
expect(runStagedLLMMock).not.toHaveBeenCalled();
});
it('runs the cue-planner stage and returns sanitized cues with the default engine', async () => {
runStagedLLMMock.mockResolvedValue({
content: { cues: [
{ label: 'Act I', prompt: 'warm pads' },
{ label: 'Climax', prompt: 'tense strings' },
] },
runId: 'r1', providerId: 'p', model: 'm',
});
const { cues, runId } = await deriveAudioCues(issueWithBeats, { defaultEngine: 'audioldm2' });
expect(runStagedLLMMock).toHaveBeenCalledWith('pipeline-audio-cues', expect.any(Object), expect.objectContaining({ returnsJson: true }));
expect(cues).toHaveLength(2);
expect(cues[0]).toMatchObject({ id: 'cue-001', label: 'Act I', prompt: 'warm pads', engine: 'audioldm2', startSec: null, trackFilename: null });
expect(cues[1].id).toBe('cue-002');
expect(runId).toBe('r1');
});
it('drops prompt-less cues and re-indexes ids sequentially', async () => {
runStagedLLMMock.mockResolvedValue({
content: { cues: [
{ label: 'A', prompt: 'p1' },
{ label: 'B' }, // dropped — no prompt
{ label: 'C', prompt: 'p3' },
] },
});
const { cues } = await deriveAudioCues(issueWithBeats, { defaultEngine: 'musicgen' });
expect(cues.map((c) => c.id)).toEqual(['cue-001', 'cue-002']);
expect(cues.map((c) => c.label)).toEqual(['A', 'C']);
});
it('throws PIPELINE_AUDIO_CUES_EMPTY when no usable cues come back', async () => {
runStagedLLMMock.mockResolvedValue({ content: { cues: [{ label: 'no prompt' }] } });
await expect(deriveAudioCues(issueWithBeats)).rejects.toMatchObject({ code: ERR_EMPTY_RESULT });
});
});