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 { beforeEach, describe, expect, it, vi } from 'vitest';
vi.mock('./aiProvider.js', async (importActual) => {
const actual = await importActual();
return { ...actual, resolveAPIProvider: vi.fn() };
});
vi.mock('./promptRunner.js', async (importActual) => {
const actual = await importActual();
return { ...actual, runPromptThroughProvider: vi.fn() };
});
const aiProvider = await import('./aiProvider.js');
const promptRunner = await import('./promptRunner.js');
const {
analyzeUniverseStyleReference,
buildStyleReferenceDiff,
buildStyleReferencePrompt,
} = await import('./universeStyleReference.js');
const apiProvider = { id: 'ollama', type: 'api', defaultModel: 'qwen-vl' };
const analysisText = JSON.stringify({
title: 'Dust-lit ink wash',
prompt: 'A weathered city rendered in granular ink wash and muted ochre.',
styleNotes: 'Tactile ink-wash science fiction with restrained, dusty light.',
influences: {
embrace: ['granular ink wash', 'muted ochre', 'weathered silhouettes'],
avoid: ['glossy 3D', 'neon'],
},
rationale: 'The image replaces polish with tactile marks and restrained color.',
});
beforeEach(() => {
vi.clearAllMocks();
aiProvider.resolveAPIProvider.mockResolvedValue(apiProvider);
promptRunner.runPromptThroughProvider.mockResolvedValue({
text: analysisText,
model: 'qwen-vl',
});
});
describe('universeStyleReference', () => {
it('analyzes the image and returns a reference plus a reviewable diff', async () => {
const result = await analyzeUniverseStyleReference({
imagePath: '/mock/images/reference.png',
imageFilename: 'reference.png',
styleNotes: 'Clean vector art',
influences: { embrace: ['clean vectors'], avoid: ['grain'] },
model: 'qwen-vl',
});
expect(result.reference).toMatchObject({
title: 'Dust-lit ink wash',
prompt: expect.stringContaining('granular ink wash'),
imageRefs: ['reference.png'],
});
expect(result.reference.id).toMatch(/^style-ref-/);
expect(result.diff).toMatchObject({
hasChanges: true,
influences: {
embrace: { added: ['granular ink wash', 'muted ochre', 'weathered silhouettes'], removed: ['clean vectors'] },
avoid: { added: ['glossy 3D', 'neon'], removed: ['grain'] },
},
});
expect(promptRunner.runPromptThroughProvider).toHaveBeenCalledWith(
expect.objectContaining({
screenshots: ['/mock/images/reference.png'],
source: 'universe-style-reference',
}),
);
});
it('preserves user-supplied metadata and locked guidance', async () => {
const result = await analyzeUniverseStyleReference({
imagePath: '/mock/images/reference.png',
imageFilename: 'reference.png',
title: 'My title',
prompt: 'My exact recreation prompt',
styleNotes: 'Pinned prose',
influences: { embrace: ['pinned positive'], avoid: ['pinned negative'] },
locked: { styleNotes: true, influencesEmbrace: true, influencesAvoid: true },
});
expect(result.reference).toMatchObject({
title: 'My title',
prompt: 'My exact recreation prompt',
});
expect(result.proposed).toEqual({
styleNotes: 'Pinned prose',
influences: { embrace: ['pinned positive'], avoid: ['pinned negative'] },
});
expect(result.diff.hasChanges).toBe(false);
});
it('distinguishes an explicit empty recommendation from an omitted list', async () => {
promptRunner.runPromptThroughProvider.mockResolvedValue({
text: JSON.stringify({
title: 'Reference',
prompt: 'Prompt',
styleNotes: 'Notes',
influences: { embrace: [] },
}),
});
const result = await analyzeUniverseStyleReference({
imagePath: '/mock/reference.png',
imageFilename: 'reference.png',
influences: { embrace: ['remove me'], avoid: ['preserve me'] },
});
expect(result.proposed.influences).toEqual({ embrace: [], avoid: ['preserve me'] });
});
it('rejects invalid JSON and vision runs that dropped the image', async () => {
promptRunner.runPromptThroughProvider.mockResolvedValueOnce({ text: 'not json' });
await expect(analyzeUniverseStyleReference({
imagePath: '/mock/reference.png',
imageFilename: 'reference.png',
})).rejects.toMatchObject({ code: 'VISION_BAD_JSON', status: 502 });
promptRunner.runPromptThroughProvider.mockResolvedValueOnce({
text: analysisText,
provider: { id: 'cli-provider', type: 'cli' },
});
await expect(analyzeUniverseStyleReference({
imagePath: '/mock/reference.png',
imageFilename: 'reference.png',
})).rejects.toMatchObject({ code: 'VISION_FALLBACK_DROPPED_IMAGES', status: 502 });
});
it('builds deterministic list and prose diffs', () => {
expect(buildStyleReferenceDiff(
{ styleNotes: 'before', influences: { embrace: ['A'], avoid: [] } },
{ styleNotes: 'after', influences: { embrace: ['A', 'B'], avoid: [] } },
)).toMatchObject({
hasChanges: true,
styleNotes: { before: 'before', after: 'after', changed: true },
influences: { embrace: { added: ['B'], removed: [] } },
});
});
it('embeds supplied metadata and locks as JSON context', () => {
const prompt = buildStyleReferencePrompt({
title: 'Reference',
prompt: '',
styleNotes: 'Notes',
influences: { embrace: ['ink'], avoid: [] },
locked: { styleNotes: true },
});
expect(prompt).toContain('"suppliedTitle":"Reference"');
expect(prompt).toContain('"locked":{"styleNotes":true}');
});
});