feat(src/brainyData, src/utils): enhance embedding efficiency with batch processing and initialize safeguards

- Added batch embedding support with `defaultBatchEmbeddingFunction`, leveraging shared model instances for optimized performance.
- Integrated `isInitializing` flag to prevent recursive initialization and ensure smooth concurrent operation handling during `BrainyData` initialization.
- Pre-loaded Universal Sentence Encoder in `BrainyData` to prevent delays during embedding.
- Introduced fallback mechanisms in embedding initialization for better error resiliency and model reusability.
- Updated `addBatch` with support for batchSize and refactored text/vector processing logic for clearer separation and memory management.
- Improved GPU and CPU backend selection in Universal Sentence Encoder for compatibility across environments.
- Enhanced memory management by cleaning tensors after embedding operations.
- Updated README with instructions for batch embedding, threading updates, and GPU/CPU optimizations.
This commit is contained in:
David Snelling 2025-06-27 14:06:59 -07:00
parent e81979dc84
commit bba9a0c219
11 changed files with 729 additions and 633 deletions

View file

@ -13,7 +13,10 @@ import { isThreadingAvailable } from './environment.js'
* Lower values indicate higher similarity
* Optimized using array methods for Node.js 23.11+
*/
export const euclideanDistance: DistanceFunction = (a: Vector, b: Vector): number => {
export const euclideanDistance: DistanceFunction = (
a: Vector,
b: Vector
): number => {
if (a.length !== b.length) {
throw new Error('Vectors must have the same dimensions')
}
@ -21,7 +24,7 @@ export const euclideanDistance: DistanceFunction = (a: Vector, b: Vector): numbe
// Use array.reduce for better performance in Node.js 23.11+
const sum = a.reduce((acc, val, i) => {
const diff = val - b[i]
return acc + (diff * diff)
return acc + diff * diff
}, 0)
return Math.sqrt(sum)
@ -33,19 +36,25 @@ export const euclideanDistance: DistanceFunction = (a: Vector, b: Vector): numbe
* Range: 0 (identical) to 2 (opposite)
* Optimized using array methods for Node.js 23.11+
*/
export const cosineDistance: DistanceFunction = (a: Vector, b: Vector): number => {
export const cosineDistance: DistanceFunction = (
a: Vector,
b: Vector
): number => {
if (a.length !== b.length) {
throw new Error('Vectors must have the same dimensions')
}
// Use array.reduce to calculate all values in a single pass
const { dotProduct, normA, normB } = a.reduce((acc, val, i) => {
return {
dotProduct: acc.dotProduct + (val * b[i]),
normA: acc.normA + (val * val),
normB: acc.normB + (b[i] * b[i])
}
}, { dotProduct: 0, normA: 0, normB: 0 })
const { dotProduct, normA, normB } = a.reduce(
(acc, val, i) => {
return {
dotProduct: acc.dotProduct + val * b[i],
normA: acc.normA + val * val,
normB: acc.normB + b[i] * b[i]
}
},
{ dotProduct: 0, normA: 0, normB: 0 }
)
if (normA === 0 || normB === 0) {
return 2 // Maximum distance for zero vectors
@ -61,7 +70,10 @@ export const cosineDistance: DistanceFunction = (a: Vector, b: Vector): number =
* Lower values indicate higher similarity
* Optimized using array methods for Node.js 23.11+
*/
export const manhattanDistance: DistanceFunction = (a: Vector, b: Vector): number => {
export const manhattanDistance: DistanceFunction = (
a: Vector,
b: Vector
): number => {
if (a.length !== b.length) {
throw new Error('Vectors must have the same dimensions')
}
@ -76,209 +88,211 @@ export const manhattanDistance: DistanceFunction = (a: Vector, b: Vector): numbe
* Converted to a distance metric (lower is better)
* Optimized using array methods for Node.js 23.11+
*/
export const dotProductDistance: DistanceFunction = (a: Vector, b: Vector): number => {
export const dotProductDistance: DistanceFunction = (
a: Vector,
b: Vector
): number => {
if (a.length !== b.length) {
throw new Error('Vectors must have the same dimensions')
}
// Use array.reduce for better performance in Node.js 23.11+
const dotProduct = a.reduce((sum, val, i) => sum + (val * b[i]), 0)
const dotProduct = a.reduce((sum, val, i) => sum + val * b[i], 0)
// Convert to a distance metric (lower is better)
return -dotProduct
}
/**
* GPU-accelerated batch distance calculation
* Uses TensorFlow.js with WebGL backend when available for optimal performance
* Falls back to CPU processing when GPU is not available
* Batch distance calculation
* Uses TensorFlow.js with CPU backend for optimized performance
*
* @param queryVector The query vector to compare against all vectors
* @param vectors Array of vectors to compare against
* @param distanceFunction The distance function to use
* @returns Promise resolving to array of distances
*/
export async function calculateDistancesWithGPU(
export async function calculateDistancesBatch(
queryVector: Vector,
vectors: Vector[],
distanceFunction: DistanceFunction = euclideanDistance
): Promise<number[]> {
// For small batches, use the standard distance function
if (vectors.length < 10) {
return vectors.map(vector => distanceFunction(queryVector, vector))
return vectors.map((vector) => distanceFunction(queryVector, vector))
}
try {
// Function to be executed in a worker thread
const distanceCalculator = async (
args: {
queryVector: Vector,
vectors: Vector[],
distanceFnString: string
}
) => {
const distanceCalculator = async (args: {
queryVector: Vector
vectors: Vector[]
distanceFnString: string
}) => {
const { queryVector, vectors, distanceFnString } = args
// Try to use TensorFlow.js with GPU acceleration if available
// Use TensorFlow.js with CPU processing
const useTensorFlow = async () => {
try {
// TensorFlow.js will use its default EPSILON value
// TensorFlow.js will use its default EPSILON value
// Use the importTensorFlow function if available (in worker context)
// or directly import TensorFlow.js (in main thread)
let tf
if (
typeof self !== 'undefined' &&
typeof self.importTensorFlow === 'function'
) {
// In worker context, use the importTensorFlow function
tf = await self.importTensorFlow()
} else {
// Dynamically import TensorFlow.js core module and backends
const tf = await import('@tensorflow/tfjs-core')
tf = await import('@tensorflow/tfjs-core')
// Import CPU backend as fallback
// Import CPU backend
await import('@tensorflow/tfjs-backend-cpu')
let usingGPU = false
// Set CPU as the backend
await tf.setBackend('cpu')
}
try {
// Try to import and use WebGL backend (GPU)
await import('@tensorflow/tfjs-backend-webgl')
// Convert vectors to tensors
const queryTensor = tf.tensor2d([queryVector])
const vectorsTensor = tf.tensor2d(vectors)
// Check if WebGL is available and set it as the backend
if (await tf.findBackend('webgl') || await tf.ready().then(() => tf.findBackend('webgl'))) {
await tf.setBackend('webgl')
usingGPU = true
} else {
await tf.setBackend('cpu')
}
} catch (err) {
// If WebGL fails, use CPU
await tf.setBackend('cpu')
}
let distances: number[]
// Convert vectors to tensors
const queryTensor = tf.tensor2d([queryVector])
const vectorsTensor = tf.tensor2d(vectors)
// Calculate distances based on the distance function type
if (distanceFnString.includes('euclideanDistance')) {
// Euclidean distance using GPU-optimized operations
// Formula: sqrt(sum((a - b)^2))
const expanded = tf.sub(
(queryTensor as any).expandDims(1),
(vectorsTensor as any).expandDims(0)
)
const squaredDiff = tf.square(expanded)
const sumSquaredDiff = tf.sum(squaredDiff, -1)
const distancesTensor = tf.sqrt(sumSquaredDiff)
distances = (await (distancesTensor as any)
.squeeze()
.array()) as number[]
let distances: number[]
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
expanded.dispose()
squaredDiff.dispose()
sumSquaredDiff.dispose()
distancesTensor.dispose()
} else if (distanceFnString.includes('cosineDistance')) {
// Cosine distance using GPU-optimized operations
// Formula: 1 - (a·b / (||a|| * ||b||))
const dotProduct = tf.matMul(
queryTensor,
(vectorsTensor as any).transpose()
)
// Calculate distances based on the distance function type
if (distanceFnString.includes('euclideanDistance')) {
// Euclidean distance using GPU-optimized operations
// Formula: sqrt(sum((a - b)^2))
const expanded = tf.sub((queryTensor as any).expandDims(1), (vectorsTensor as any).expandDims(0))
const squaredDiff = tf.square(expanded)
const sumSquaredDiff = tf.sum(squaredDiff, -1)
const distancesTensor = tf.sqrt(sumSquaredDiff)
distances = await (distancesTensor as any).squeeze().array() as number[]
const queryNorm = tf.norm(queryTensor, 2, 1)
const vectorsNorm = tf.norm(vectorsTensor, 2, 1)
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
expanded.dispose()
squaredDiff.dispose()
sumSquaredDiff.dispose()
distancesTensor.dispose()
} else if (distanceFnString.includes('cosineDistance')) {
// Cosine distance using GPU-optimized operations
// Formula: 1 - (a·b / (||a|| * ||b||))
const dotProduct = tf.matMul(queryTensor, (vectorsTensor as any).transpose())
const normProduct = tf.outerProduct(
queryNorm as any,
vectorsNorm as any
)
const cosineSimilarity = tf.div(dotProduct, normProduct)
const distancesTensor = tf.sub(tf.scalar(1), cosineSimilarity)
const queryNorm = tf.norm(queryTensor, 2, 1)
const vectorsNorm = tf.norm(vectorsTensor, 2, 1)
distances = (await (distancesTensor as any)
.squeeze()
.array()) as number[]
const normProduct = tf.outerProduct(queryNorm as any, vectorsNorm as any)
const cosineSimilarity = tf.div(dotProduct, normProduct)
const distancesTensor = tf.sub(tf.scalar(1), cosineSimilarity)
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
dotProduct.dispose()
queryNorm.dispose()
vectorsNorm.dispose()
normProduct.dispose()
cosineSimilarity.dispose()
distancesTensor.dispose()
} else if (distanceFnString.includes('manhattanDistance')) {
// Manhattan distance using GPU-optimized operations
// Formula: sum(|a - b|)
const diff = tf.sub(
(queryTensor as any).expandDims(1),
(vectorsTensor as any).expandDims(0)
)
const absDiff = tf.abs(diff)
const distancesTensor = tf.sum(absDiff, -1)
distances = await (distancesTensor as any).squeeze().array() as number[]
distances = (await (distancesTensor as any)
.squeeze()
.array()) as number[]
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
dotProduct.dispose()
queryNorm.dispose()
vectorsNorm.dispose()
normProduct.dispose()
cosineSimilarity.dispose()
distancesTensor.dispose()
} else if (distanceFnString.includes('manhattanDistance')) {
// Manhattan distance using GPU-optimized operations
// Formula: sum(|a - b|)
const diff = tf.sub((queryTensor as any).expandDims(1), (vectorsTensor as any).expandDims(0))
const absDiff = tf.abs(diff)
const distancesTensor = tf.sum(absDiff, -1)
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
diff.dispose()
absDiff.dispose()
distancesTensor.dispose()
} else if (distanceFnString.includes('dotProductDistance')) {
// Dot product distance using GPU-optimized operations
// Formula: -sum(a * b)
const dotProduct = tf.matMul(
queryTensor,
(vectorsTensor as any).transpose()
)
const distancesTensor = tf.neg(dotProduct)
distances = await (distancesTensor as any).squeeze().array() as number[]
distances = (await (distancesTensor as any)
.squeeze()
.array()) as number[]
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
diff.dispose()
absDiff.dispose()
distancesTensor.dispose()
} else if (distanceFnString.includes('dotProductDistance')) {
// Dot product distance using GPU-optimized operations
// Formula: -sum(a * b)
const dotProduct = tf.matMul(queryTensor, (vectorsTensor as any).transpose())
const distancesTensor = tf.neg(dotProduct)
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
dotProduct.dispose()
distancesTensor.dispose()
} else {
// For unknown distance functions, fall back to direct CPU implementation
throw new Error(
'Unsupported distance function for TensorFlow optimization'
)
}
distances = await (distancesTensor as any).squeeze().array() as number[]
// Clean up tensors
queryTensor.dispose()
vectorsTensor.dispose()
dotProduct.dispose()
distancesTensor.dispose()
} else {
// For unknown distance functions, fall back to CPU implementation
throw new Error('Unsupported distance function for GPU acceleration')
}
return {
distances,
usingGPU
}
} catch (error) {
// If TensorFlow.js fails, fall back to CPU implementation
throw error
return {
distances
}
}
// Try to use TensorFlow.js with GPU acceleration
// Try to use TensorFlow.js with CPU optimization
try {
return await useTensorFlow()
} catch (error) {
// Fall back to CPU implementation if TensorFlow.js fails
// Fall back to direct CPU implementation if TensorFlow.js fails
// Recreate the distance function from its string representation
const distanceFunction = new Function('return ' + distanceFnString)() as DistanceFunction
const distanceFunction = new Function(
'return ' + distanceFnString
)() as DistanceFunction
// Calculate distances for all vectors
const distances = vectors.map(vector => distanceFunction(queryVector, vector))
const distances = vectors.map((vector) =>
distanceFunction(queryVector, vector)
)
return {
distances,
usingGPU: false
distances
}
}
}
// Execute the distance calculation in a separate thread if threading is available
if (isThreadingAvailable()) {
try {
// Convert the distance function to a string for serialization
const distanceFnString = distanceFunction.toString()
// Execute in a separate thread
const result = await executeInThread<{ distances: number[], usingGPU: boolean }>(
distanceCalculator.toString(),
{ queryVector, vectors, distanceFnString }
)
return result.distances
} catch (error) {
// Fall back to main thread if threading fails
console.warn('Threaded distance calculation failed, falling back to main thread:', error)
}
}
// Threading is not available, so we'll always use the main thread implementation
// This comment is kept for clarity about the removed code
// If threading is not available or failed, calculate distances in the main thread
return vectors.map(vector => distanceFunction(queryVector, vector))
return vectors.map((vector) => distanceFunction(queryVector, vector))
} catch (error) {
// If anything fails, fall back to the standard distance function
console.error('GPU-accelerated distance calculation failed:', error)
return vectors.map(vector => distanceFunction(queryVector, vector))
console.error('Batch distance calculation failed:', error)
return vectors.map((vector) => distanceFunction(queryVector, vector))
}
}