brainy/tests/model-loading-priority.test.ts
David Snelling f898f0ce7b feat\!: migrate from TensorFlow.js to Transformers.js with ONNX Runtime
BREAKING CHANGE: Complete migration from TensorFlow.js to Transformers.js for embedding generation

This is a major architectural change that replaces TensorFlow.js (USE model) with Transformers.js (all-MiniLM-L6-v2) for significantly improved performance and reduced complexity.

Key Changes:
- Replace TensorFlow.js Universal Sentence Encoder with Transformers.js all-MiniLM-L6-v2
- Reduce model size from 525MB to 87MB (83% reduction)
- Reduce embedding dimensions from 512 to 384 (faster distance calculations)
- Remove TensorFlow.js Float32Array patching (caused ONNX conflicts)
- Implement smart bundled model detection for offline operation
- Add explicit model download script for Docker deployments
- Remove complex environment variables in favor of simple configuration
- Update all distance functions to use optimized pure JavaScript
- Remove TensorFlow-specific utilities and type definitions

Performance Improvements:
- Model loading: 5x faster (87MB vs 525MB)
- Memory usage: 75% reduction (~200-400MB vs ~1.5GB)
- Distance calculations: Faster pure JS vs GPU overhead for small vectors
- Cold start performance: Significantly improved

Files Changed:
- Updated package.json: New dependencies, simplified scripts
- Rewrote src/utils/embedding.ts: Complete Transformers.js implementation
- Updated src/utils/distance.ts: Optimized JavaScript distance functions
- Simplified src/setup.ts: Removed TensorFlow-specific patching
- Simplified src/utils/textEncoding.ts: Only Node.js TextEncoder/Decoder patches
- Deleted src/utils/robustModelLoader.ts: TensorFlow-specific loader
- Deleted src/types/tensorflowTypes.ts: TensorFlow type definitions
- Added scripts/download-models.cjs: Docker-compatible model downloader
- Added comprehensive documentation: README.md, OFFLINE_MODELS.md, analysis docs

Testing:
- All 19 tests passing
- Removed test mocking in favor of real implementation testing
- Updated test environment for Transformers.js compatibility
- Performance tests validate improved efficiency

This migration resolves production issues with Docker egress limitations and provides a more robust, performant foundation for vector operations.
2025-08-05 19:29:59 -07:00

165 lines
No EOL
5.6 KiB
TypeScript

/**
* Model Loading Priority Test
*
* This test verifies that the model loading system correctly prioritizes
* local models from @soulcraft/brainy-models over remote URL loading.
*/
import { describe, it, expect, beforeAll, vi } from 'vitest'
import { RobustModelLoader } from '../src/utils/robustModelLoader.js'
describe('Model Loading Priority', () => {
let originalConsoleLog: any
let originalConsoleWarn: any
let logMessages: string[] = []
let warnMessages: string[] = []
beforeAll(() => {
// Capture console output
originalConsoleLog = console.log
originalConsoleWarn = console.warn
console.log = (...args: any[]) => {
logMessages.push(args.join(' '))
originalConsoleLog(...args)
}
console.warn = (...args: any[]) => {
warnMessages.push(args.join(' '))
originalConsoleWarn(...args)
}
})
afterAll(() => {
// Restore console
console.log = originalConsoleLog
console.warn = originalConsoleWarn
})
it('should try to load @soulcraft/brainy-models first', async () => {
logMessages = []
warnMessages = []
const loader = new RobustModelLoader({ verbose: false })
try {
// This will try to load the model
await loader.loadModelWithFallbacks()
} catch (error) {
// It's okay if it fails in test environment
console.log('Model loading failed (expected in test environment):', error)
}
// Check if it attempted to load local models (either @tensorflow-models or @soulcraft/brainy-models)
const hasCheckedForLocalModel = logMessages.some(msg =>
msg.includes('@soulcraft/brainy-models') ||
msg.includes('Checking for @soulcraft/brainy-models') ||
msg.includes('@tensorflow-models/universal-sentence-encoder') ||
msg.includes('Checking for @tensorflow-models/universal-sentence-encoder')
)
expect(hasCheckedForLocalModel).toBe(true)
})
it('should log warnings when falling back to URL loading', async () => {
logMessages = []
warnMessages = []
const loader = new RobustModelLoader({ verbose: false })
try {
await loader.loadModelWithFallbacks()
} catch (error) {
// Expected in test environment without actual model
}
// If @soulcraft/brainy-models is not installed, should see warning
const hasFallbackWarning = warnMessages.some(msg =>
msg.includes('Local model (@soulcraft/brainy-models) not found') ||
msg.includes('Falling back to remote model loading')
)
// We should see one of these: either local model found or fallback warning
const hasLocalModelSuccess = logMessages.some(msg =>
msg.includes('Found @soulcraft/brainy-models package installed') ||
msg.includes('Found @tensorflow-models/universal-sentence-encoder package')
)
// Either we found the local model OR we got a fallback warning
expect(hasLocalModelSuccess || hasFallbackWarning).toBe(true)
// If we're using fallback, should see installation suggestion
if (hasFallbackWarning) {
const hasInstallSuggestion = warnMessages.some(msg =>
msg.includes('npm install @soulcraft/brainy-models')
)
expect(hasInstallSuggestion).toBe(true)
}
})
it('should verify model correctness when loading from URL', async () => {
// This test is more of a documentation of the expected behavior
// The actual model loading would fail in test environment
const loader = new RobustModelLoader({ verbose: true })
// The loadModelWithFallbacks method now includes model verification
// It checks that embeddings have the correct dimensions (512)
// This ensures we're loading the Universal Sentence Encoder
expect(loader).toBeDefined()
})
it('should prioritize local model over URL when available', async () => {
// Mock the import to simulate @soulcraft/brainy-models being available
const mockBrainyModels = {
BundledUniversalSentenceEncoder: class {
constructor(options: any) {}
async load() { return true }
async embedToArrays(input: string[]) {
// Return mock embeddings with correct dimensions
return input.map(() => new Array(384).fill(0.1))
}
dispose() {}
}
}
// Create a custom loader that mocks the import
const loader = new RobustModelLoader({ verbose: true })
// Override the tryLoadLocalBundledModel to simulate local model
const originalTryLoad = (loader as any).tryLoadLocalBundledModel
;(loader as any).tryLoadLocalBundledModel = async function() {
console.log('✅ Found @soulcraft/brainy-models package installed')
console.log(' Using local bundled model for maximum performance and reliability')
// Return a mock model
return {
init: async () => {},
embed: async (sentences: string | string[]) => {
const input = Array.isArray(sentences) ? sentences : [sentences]
return new Array(384).fill(0.1)
},
dispose: async () => {}
}
}
logMessages = []
warnMessages = []
const model = await loader.loadModelWithFallbacks()
expect(model).toBeDefined()
// Should see success message for local model
const hasLocalSuccess = logMessages.some(msg =>
msg.includes('Using local bundled model')
)
expect(hasLocalSuccess).toBe(true)
// Should NOT see fallback warnings
const hasFallbackWarning = warnMessages.some(msg =>
msg.includes('Falling back to remote model loading')
)
expect(hasFallbackWarning).toBe(false)
})
})