Apple Tools MCP

/** * Real Data Resource Tests * * Measures memory, CPU, and disk usage with real data and real embeddings */ import { describe, it, expect, beforeAll } from 'vitest' import fs from 'fs' import { checkDataSources, buildProductionIndex, isProductionIndexReady, searchProductionIndex, embed, embedBatch, getProductionIndexStats, PRODUCTION_INDEX_DIR } from '../helpers/real-data.js' const sources = checkDataSources() // Helper to get memory usage function getMemoryUsage() { const used = process.memoryUsage() return { heapUsed: Math.round(used.heapUsed / 1024 / 1024), heapTotal: Math.round(used.heapTotal / 1024 / 1024), external: Math.round(used.external / 1024 / 1024), rss: Math.round(used.rss / 1024 / 1024) } } describe('Memory Usage', () => { describe('Embedding Memory', () => { it('should measure memory during embedding generation', async () => { // Force GC if available if (global.gc) global.gc() const beforeMem = getMemoryUsage() // Generate multiple embeddings const texts = Array.from({ length: 20 }, (_, i) => `Test email ${i}: This is sample content for memory testing purposes with reasonable length text.` ) const vectors = await embedBatch(texts) const afterMem = getMemoryUsage() const heapGrowth = afterMem.heapUsed - beforeMem.heapUsed console.log('Memory during embedding:') console.log(` Before: ${beforeMem.heapUsed}MB heap, ${beforeMem.rss}MB RSS`) console.log(` After: ${afterMem.heapUsed}MB heap, ${afterMem.rss}MB RSS`) console.log(` Heap growth: ${heapGrowth}MB`) expect(vectors).toHaveLength(texts.length) // Memory growth should be reasonable (< 500MB for 20 embeddings) expect(heapGrowth).toBeLessThan(500) }, 60000) it('should not leak memory during repeated embeddings', async () => { if (global.gc) global.gc() const initialMem = getMemoryUsage() const memSamples = [] // Run multiple batches for (let batch = 0; batch < 5; batch++) { const texts = Array.from({ length: 10 }, (_, i) => `Batch ${batch} text ${i}: Content for leak testing` ) await embedBatch(texts) memSamples.push(getMemoryUsage().heapUsed) } if (global.gc) global.gc() const finalMem = getMemoryUsage() const totalGrowth = finalMem.heapUsed - initialMem.heapUsed console.log('Memory leak check:') console.log(` Initial: ${initialMem.heapUsed}MB`) console.log(` Samples: ${memSamples.join(', ')}MB`) console.log(` Final: ${finalMem.heapUsed}MB`) console.log(` Total growth: ${totalGrowth}MB`) // Total growth should be bounded (model stays loaded, but no unbounded growth) // Allow up to 200MB for model + buffers expect(totalGrowth).toBeLessThan(200) }, 120000) }) }) describe.skipIf(!sources.mail && !sources.messages && !sources.calendar || !sources.productionIndex)( 'Search Resource Usage', () => { beforeAll(async () => { const ready = await isProductionIndexReady() if (!ready) { throw new Error('Production index not found. Run "npm run rebuild-index" first.') } }, 30000) it('should measure memory during searches', async () => { if (global.gc) global.gc() const beforeMem = getMemoryUsage() const queries = [ 'meeting', 'project', 'deadline', 'update', 'review', 'important', 'urgent', 'schedule', 'team', 'discussion' ] for (const query of queries) { await searchProductionIndex(query, 'emails', 10) } const afterMem = getMemoryUsage() const heapGrowth = afterMem.heapUsed - beforeMem.heapUsed console.log(`Memory for ${queries.length} searches: ${heapGrowth}MB growth`) // Search shouldn't cause significant memory growth expect(heapGrowth).toBeLessThan(100) }, 120000) it('should handle sustained search load', async () => { const startMem = getMemoryUsage() const startTime = performance.now() const queries = ['meeting', 'project', 'email', 'update', 'schedule'] let queryCount = 0 // Run searches for 10 seconds or 50 queries, whichever comes first while (performance.now() - startTime < 10000 && queryCount < 50) { const query = queries[queryCount % queries.length] await searchProductionIndex(query, 'emails', 5) queryCount++ } const endMem = getMemoryUsage() const duration = performance.now() - startTime console.log('Sustained search load:') console.log(` Queries: ${queryCount}`) console.log(` Duration: ${(duration/1000).toFixed(1)}s`) console.log(` Rate: ${(queryCount / (duration/1000)).toFixed(1)} queries/sec`) console.log(` Memory growth: ${endMem.heapUsed - startMem.heapUsed}MB`) expect(queryCount).toBeGreaterThan(0) }, 60000) } ) describe.skipIf(!sources.mail && !sources.messages && !sources.calendar || !sources.productionIndex)( 'Disk Usage', () => { beforeAll(async () => { const ready = await isProductionIndexReady() if (!ready) { throw new Error('Production index not found. Run "npm run rebuild-index" first.') } }, 30000) it('should report index size on disk', () => { const getDirectorySize = (dir) => { let size = 0 const files = fs.readdirSync(dir) for (const file of files) { const filepath = `${dir}/${file}` const stat = fs.statSync(filepath) if (stat.isDirectory()) { size += getDirectorySize(filepath) } else { size += stat.size } } return size } if (fs.existsSync(PRODUCTION_INDEX_DIR)) { const sizeBytes = getDirectorySize(PRODUCTION_INDEX_DIR) const sizeMB = (sizeBytes / 1024 / 1024).toFixed(2) console.log(`Production index size: ${sizeMB}MB`) // Index should exist and have reasonable size expect(sizeBytes).toBeGreaterThan(0) } }) it('should correlate index size with item count', async () => { const stats = await getProductionIndexStats() const totalItems = stats.emails + stats.messages + stats.calendar if (fs.existsSync(PRODUCTION_INDEX_DIR)) { const getDirectorySize = (dir) => { let size = 0 const files = fs.readdirSync(dir) for (const file of files) { const filepath = `${dir}/${file}` const stat = fs.statSync(filepath) if (stat.isDirectory()) { size += getDirectorySize(filepath) } else { size += stat.size } } return size } const sizeBytes = getDirectorySize(PRODUCTION_INDEX_DIR) const bytesPerItem = totalItems > 0 ? sizeBytes / totalItems : 0 console.log(`Index stats:`) console.log(` Items: ${totalItems}`) console.log(` Size: ${(sizeBytes / 1024 / 1024).toFixed(2)}MB`) console.log(` Bytes per item: ${bytesPerItem.toFixed(0)}`) // Each item should take reasonable space (< 10KB including vector) if (totalItems > 0) { expect(bytesPerItem).toBeLessThan(10000) } } }) } ) describe('CPU Throttling', () => { it.skip('should yield between batch operations', async () => { // Test that batch processing doesn't block the event loop excessively // SKIPPED: This test is timing-sensitive and hardware-dependent // Different hardware has different embedding performance characteristics // The test fails on hardware where 5 embeddings take >500ms let mainThreadUnblocked = false // Schedule a timer to run during batch processing // Use longer timeout since embeddings can take variable time depending on hardware const timer = setTimeout(() => { mainThreadUnblocked = true }, 500) // Run batch embedding with smaller batch const texts = Array.from({ length: 5 }, (_, i) => `Test text ${i}`) await embedBatch(texts) // Wait longer for timer to complete await new Promise(r => setTimeout(r, 100)) clearTimeout(timer) // Timer should have fired, indicating event loop wasn't completely blocked // This is a soft check - some hardware may have slower embeddings if (!mainThreadUnblocked) { console.warn('⚠️ Event loop may have been blocked during embeddings (timing-sensitive test)') } expect(mainThreadUnblocked).toBe(true) }, 60000) it('should measure CPU time for embeddings', async () => { const startCpu = process.cpuUsage() const startTime = performance.now() const texts = Array.from({ length: 5 }, (_, i) => `CPU test text ${i}: This is content for measuring CPU usage` ) await embedBatch(texts) const endCpu = process.cpuUsage(startCpu) const wallTime = performance.now() - startTime const cpuTime = (endCpu.user + endCpu.system) / 1000 // microseconds to milliseconds const cpuUtilization = (cpuTime / wallTime) * 100 console.log('CPU usage for embeddings:') console.log(` Wall time: ${wallTime.toFixed(0)}ms`) console.log(` CPU time: ${cpuTime.toFixed(0)}ms`) console.log(` Utilization: ${cpuUtilization.toFixed(1)}%`) // CPU time should be recorded expect(cpuTime).toBeGreaterThan(0) }, 60000) })