peterfei/ai-agent-team / .claude/skills/thread-manager
thread-manager
Flaggedtests/unit/vector-search.test.ts
View on GitHubimport { DatabaseManager, ThreadsDAO, MessagesDAO } from '../../src/database';
// No direct import of XenovaEmbeddingService, as it's fully mocked below.
import { MockEmbeddingService } from '../unit/__mocks__/embedding-service'; // Import the mock service directly
import { cosineSimilarity } from '../../src/core/vector-utils';
import path from 'path';
import fs from 'fs-extra';
import { v4 as uuidv4 } from 'uuid';
jest.mock('../../src/core/embedding-service', () => ({
// Use a factory function to ensure MockEmbeddingService is defined when accessed
XenovaEmbeddingService: jest.fn(() => new MockEmbeddingService()),
}));
describe('Vector Search Integration', () => {
const testDir = path.join(__dirname, '../../tmp-test-vector');
let dbPath: string;
let dbManager: DatabaseManager;
let threadsDAO: ThreadsDAO;
let messagesDAO: MessagesDAO;
let embeddingService: MockEmbeddingService;
let threadId: string;
beforeAll(() => {
fs.ensureDirSync(testDir);
});
afterAll(() => {
fs.removeSync(testDir);
});
beforeEach(() => {
dbPath = path.join(testDir, `test-vector-${uuidv4()}.db`);
dbManager = new DatabaseManager(dbPath);
dbManager.init();
embeddingService = new MockEmbeddingService();
threadsDAO = new ThreadsDAO(dbManager);
messagesDAO = new MessagesDAO(dbManager, embeddingService);
const thread = threadsDAO.create({ title: 'Vector Test Thread' });
threadId = thread.id;
});
afterEach(() => {
dbManager.close();
if (fs.existsSync(dbPath)) {
fs.unlinkSync(dbPath);
}
});
it('should generate embedding when creating message', async () => {
const message = await messagesDAO.create({
threadId,
role: 'user',
content: 'I like apple'
});
expect(message.embedding).toBeDefined();
expect(message.embedding).toHaveLength(384);
// Verify stored in DB
const db = dbManager.getDatabase();
const row = db.prepare('SELECT embedding_blob FROM messages WHERE id = ?').get(message.id) as any;
expect(row.embedding_blob).toBeDefined();
expect(row.embedding_blob.length).toBe(384 * 4); // Float32 (4 bytes)
});
it('should perform semantic search', async () => {
// 1. Add messages
await messagesDAO.create({ threadId, role: 'user', content: 'I like eating an apple' }); // Related
await messagesDAO.create({ threadId, role: 'assistant', content: 'Apple is a fruit' }); // Related
await messagesDAO.create({ threadId, role: 'user', content: 'My car is fast' }); // Unrelated
// 2. Search
const results = await messagesDAO.searchSimilar('fruit', {
threadId,
topK: 5,
minScore: 0.1
});
// 3. Verify results
expect(results.length).toBeGreaterThanOrEqual(2);
// Check that we found relevant messages (case-insensitive)
const contents = results.map(r => r.content.toLowerCase());
expect(contents.some(c => c.includes('apple'))).toBe(true);
expect(contents.some(c => c.includes('fruit'))).toBe(true);
// 'car' should be much lower score or filtered if we set higher threshold
const carMsg = results.find(r => r.content.includes('car'));
if (carMsg) {
// The car message vector has 0.9 at index 1, fruit has 0.9 at index 0.
// Dot product should be 0.
expect(carMsg.score).toBeLessThan(0.1);
}
});
it('should calculate cosine similarity correctly', () => {
const vecA = [1, 0, 0];
const vecB = [1, 0, 0];
const vecC = [0, 1, 0];
const vecD = [0.7071, 0.7071, 0];
expect(cosineSimilarity(vecA, vecB)).toBeCloseTo(1);
expect(cosineSimilarity(vecA, vecC)).toBeCloseTo(0);
expect(cosineSimilarity(vecA, vecD)).toBeCloseTo(0.7071);
});
});
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |