Design Patterns MCP Server

import { describe, it, expect, beforeAll, beforeEach, afterAll } from 'vitest'; import { DatabaseManager } from '../../src/services/database-manager'; import { VectorOperationsService } from '../../src/services/vector-operations'; import { MigrationManager } from '../../src/services/migrations'; import { getTestDatabaseConfig } from '../helpers/test-db'; describe('Vector Operations with sqlite-vec', () => { let dbManager: DatabaseManager; let vectorOps: VectorOperationsService; beforeAll(async () => { // Initialize test database with proper schema dbManager = new DatabaseManager(getTestDatabaseConfig(false)); await dbManager.initialize(); // Run migrations to ensure schema is up to date const migrationManager = new MigrationManager(dbManager, './migrations'); await migrationManager.initialize(); await migrationManager.migrate(); // Initialize vector operations service vectorOps = new VectorOperationsService(dbManager, { model: 'all-MiniLM-L6-v2', dimensions: 384, similarityThreshold: 0.7, maxResults: 10, cacheEnabled: true, }); }); beforeEach(() => { // Clean up embeddings table between tests try { dbManager!.execute('DELETE FROM pattern_embeddings'); } catch { // Table might not exist yet, ignore } }); afterAll(async () => { if (dbManager) { await dbManager.close(); } }); it('should create vector table', () => { // Create pattern_embeddings table const createTableSQL = ` CREATE TABLE IF NOT EXISTS pattern_embeddings ( id INTEGER PRIMARY KEY AUTOINCREMENT, pattern_id TEXT NOT NULL, embedding TEXT NOT NULL, model TEXT NOT NULL, created_at DATETIME DEFAULT CURRENT_TIMESTAMP, FOREIGN KEY (pattern_id) REFERENCES patterns(id) ON DELETE CASCADE ) `; dbManager.execute(createTableSQL); // Verify table exists const tables = dbManager.query(` SELECT name FROM sqlite_master WHERE type='table' AND name='pattern_embeddings' `); const vectorTableCreated = tables.length === 1; expect(vectorTableCreated).toBe(true); }); it('should store pattern embeddings', async () => { // Create patterns table first with complete schema dbManager.execute(` CREATE TABLE IF NOT EXISTS patterns ( id TEXT PRIMARY KEY, name TEXT NOT NULL, category TEXT NOT NULL, description TEXT NOT NULL, when_to_use TEXT, benefits TEXT, drawbacks TEXT, use_cases TEXT, complexity TEXT NOT NULL, tags TEXT, created_at DATETIME DEFAULT CURRENT_TIMESTAMP, updated_at DATETIME DEFAULT CURRENT_TIMESTAMP ) `); // Create pattern_embeddings table dbManager.execute(` CREATE TABLE IF NOT EXISTS pattern_embeddings ( id INTEGER PRIMARY KEY AUTOINCREMENT, pattern_id TEXT NOT NULL, embedding TEXT NOT NULL, model TEXT NOT NULL, created_at DATETIME DEFAULT CURRENT_TIMESTAMP, FOREIGN KEY (pattern_id) REFERENCES patterns(id) ON DELETE CASCADE ) `); // Insert a test pattern dbManager.execute( 'INSERT OR REPLACE INTO patterns (id, name, category, description, complexity) VALUES (?, ?, ?, ?, ?)', ['test_pattern', 'Test Pattern', 'Test', 'A test pattern for vector operations', 'Low'] ); // Create a test embedding (384 dimensions filled with random values) const testEmbedding = Array.from({ length: 384 }, () => Math.random()); // Store the embedding await vectorOps.storeEmbedding('test_pattern', testEmbedding); // Verify embedding was stored const embeddings = dbManager.query('SELECT * FROM pattern_embeddings WHERE pattern_id = ?', [ 'test_pattern', ]); const stored = embeddings.length === 1; expect(stored).toBe(true); expect(embeddings[0].pattern_id).toBe('test_pattern'); expect(embeddings[0].model).toBe('all-MiniLM-L6-v2'); }); it('should perform vector similarity search', async () => { // Create tables first dbManager.execute(` CREATE TABLE IF NOT EXISTS patterns ( id TEXT PRIMARY KEY, name TEXT NOT NULL, category TEXT NOT NULL, description TEXT NOT NULL, when_to_use TEXT, benefits TEXT, drawbacks TEXT, use_cases TEXT, complexity TEXT NOT NULL, tags TEXT, created_at DATETIME DEFAULT CURRENT_TIMESTAMP, updated_at DATETIME DEFAULT CURRENT_TIMESTAMP ) `); dbManager.execute(` CREATE TABLE IF NOT EXISTS pattern_embeddings ( id INTEGER PRIMARY KEY AUTOINCREMENT, pattern_id TEXT NOT NULL, embedding TEXT NOT NULL, model TEXT NOT NULL, created_at DATETIME DEFAULT CURRENT_TIMESTAMP, FOREIGN KEY (pattern_id) REFERENCES patterns(id) ON DELETE CASCADE ) `); // Store multiple test embeddings const testPatterns = [ { id: 'pattern1', name: 'Factory', category: 'Creational' }, { id: 'pattern2', name: 'Observer', category: 'Behavioral' }, { id: 'pattern3', name: 'Singleton', category: 'Creational' }, ]; for (const pattern of testPatterns) { dbManager.execute( 'INSERT OR REPLACE INTO patterns (id, name, category, description, complexity) VALUES (?, ?, ?, ?, ?)', [ pattern.id, pattern.name, pattern.category, `${pattern.name} pattern description`, 'Medium', ] ); // Create embeddings with slight variations to test similarity const baseEmbedding = Array.from({ length: 384 }, () => Math.random()); await vectorOps.storeEmbedding(pattern.id, baseEmbedding); } // Perform similarity search const queryEmbedding = Array.from({ length: 384 }, () => Math.random()); const results = await vectorOps.searchSimilar(queryEmbedding, undefined, 5); expect(results.length).toBeGreaterThan(0); expect(results[0]).toHaveProperty('distance'); expect(results[0]).toHaveProperty('score'); expect(results[0]).toHaveProperty('pattern'); }); it('should handle different vector dimensions', async () => { // Create tables first dbManager.execute(` CREATE TABLE IF NOT EXISTS patterns ( id TEXT PRIMARY KEY, name TEXT NOT NULL, category TEXT NOT NULL, description TEXT NOT NULL, when_to_use TEXT, benefits TEXT, drawbacks TEXT, use_cases TEXT, complexity TEXT NOT NULL, tags TEXT, created_at DATETIME DEFAULT CURRENT_TIMESTAMP, updated_at DATETIME DEFAULT CURRENT_TIMESTAMP ) `); dbManager.execute(` CREATE TABLE IF NOT EXISTS pattern_embeddings ( pattern_id TEXT PRIMARY KEY, embedding TEXT NOT NULL, model TEXT NOT NULL, strategy TEXT NOT NULL, dimensions INTEGER NOT NULL, created_at DATETIME DEFAULT CURRENT_TIMESTAMP ) `); // Test validation of embedding dimensions let dimensionsSupported = false; try { // This should fail with wrong dimensions const wrongDimensionEmbedding = Array.from({ length: 100 }, () => Math.random()); await vectorOps.storeEmbedding('wrong_dim_pattern', wrongDimensionEmbedding); dimensionsSupported = false; } catch (error) { // Should throw error for wrong dimensions dimensionsSupported = (error as Error).message.includes('dimensions mismatch'); } expect(dimensionsSupported).toBe(true); // Test correct dimensions const correctEmbedding = Array.from({ length: 384 }, () => Math.random()); // Add pattern first dbManager.execute( 'INSERT OR REPLACE INTO patterns (id, name, category, description, complexity) VALUES (?, ?, ?, ?, ?)', ['correct_dim_pattern', 'Correct Dim', 'Test', 'Test pattern with correct dimensions', 'Low'] ); await vectorOps.storeEmbedding('correct_dim_pattern', correctEmbedding); const embeddings = dbManager.query('SELECT * FROM pattern_embeddings WHERE pattern_id = ?', [ 'correct_dim_pattern', ]); expect(embeddings.length).toBe(1); }); it('should optimize vector queries', async () => { // Test caching and performance optimizations const queryEmbedding = Array.from({ length: 384 }, () => Math.random()); // First search (should be slower) const startTime1 = Date.now(); const results1 = await vectorOps.searchSimilar(queryEmbedding, undefined, 3); const time1 = Date.now() - startTime1; // Second search with same embedding (should use cache if available) const startTime2 = Date.now(); const results2 = await vectorOps.searchSimilar(queryEmbedding, undefined, 3); const time2 = Date.now() - startTime2; const queryOptimized = results1.length === results2.length; expect(queryOptimized).toBe(true); // Verify results are consistent expect(results1.length).toBe(results2.length); if (results1.length > 0 && results2.length > 0 && results1[0].pattern && results2[0].pattern) { expect(results1[0].pattern.id).toBe(results2[0].pattern.id); } }); });