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';
const { getBoard, updateBoard, updateBoardItem, backfill, promptFromMedia, composeBoardPrompt, emit } = vi.hoisted(() => ({
getBoard: vi.fn(),
updateBoard: vi.fn(async () => ({})),
updateBoardItem: vi.fn(async (_b, id, patch) => ({ id, ...patch })),
backfill: vi.fn(),
promptFromMedia: vi.fn(),
composeBoardPrompt: vi.fn(async () => ({ prompt: 'composed' })),
emit: vi.fn(),
}));
vi.mock('./index.js', () => ({ getBoard, updateBoard, updateBoardItem, backfillGalleryPrompts: backfill }));
vi.mock('../mediaPromptFromMedia.js', () => ({ promptFromMedia }));
vi.mock('../moodBoardCompositeStyle.js', () => ({ composeBoardPrompt }));
vi.mock('../socket.js', () => ({ getIo: () => ({ emit }) }));
import { startAnalyzeJob, getAnalyzeJob } from './analyzeJob.js';
const waitTerminal = async (id) => {
for (let i = 0; i < 100; i += 1) {
const job = getAnalyzeJob(id);
if (job && job.status !== 'running') return job;
await new Promise((r) => setTimeout(r, 5));
}
throw new Error('job never finished');
};
beforeEach(() => vi.clearAllMocks());
describe('board analyze job', () => {
it('analyzes only unprompted pins server-side, then composes the style', async () => {
const items = [
{ id: 'a', type: 'image', mediaKey: 'image:a.png' },
{ id: 'b', type: 'image', mediaKey: 'image:b.png', analysis: { prompt: 'done already' } },
{ id: 'c', type: 'image', imageUrl: 'https://example.com/x.png' },
];
backfill.mockResolvedValue({ id: 'mb-run', items });
getBoard.mockResolvedValue({ id: 'mb-run', items });
promptFromMedia.mockResolvedValue({ imagePrompt: 'ink' });
const started = startAnalyzeJob('mb-run', { providerId: 'p1', model: 'm' });
expect(started.status).toBe('running');
// A second start while running joins the live job instead of double-running.
expect(startAnalyzeJob('mb-run', { providerId: 'p1' }).startedAt).toBe(started.startedAt);
const job = await waitTerminal('mb-run');
expect(job).toMatchObject({ status: 'done', total: 1, done: 1, failures: 0 });
expect(promptFromMedia).toHaveBeenCalledTimes(1);
expect(promptFromMedia).toHaveBeenCalledWith(expect.objectContaining({ filename: 'a.png', providerId: 'p1' }));
expect(updateBoardItem).toHaveBeenCalledWith('mb-run', 'a', { analysis: expect.objectContaining({ prompt: 'ink' }) });
expect(updateBoard).toHaveBeenCalledWith('mb-run', { style: { prompt: 'composed' } });
expect(emit).toHaveBeenCalledWith('mood-board:analyze', expect.objectContaining({ boardId: 'mb-run' }));
});
it('fails without composing when no pin could be analyzed', async () => {
const items = [{ id: 'a', type: 'image', mediaKey: 'image:a.png' }];
backfill.mockResolvedValue({ id: 'mb-fail', items });
getBoard.mockResolvedValue({ id: 'mb-fail', items });
promptFromMedia.mockRejectedValue(new Error('vision down'));
startAnalyzeJob('mb-fail', { providerId: 'p1' });
const job = await waitTerminal('mb-fail');
expect(job.status).toBe('failed');
expect(job.failures).toBe(1);
expect(composeBoardPrompt).not.toHaveBeenCalled();
});
});