Merge remote-tracking branch 'origin/main'

# Conflicts:
#	CHANGES.md
#	package-lock.json
#	package.json
#	src/storage/adapters/fileSystemStorage.ts
This commit is contained in:
David Snelling 2025-07-28 10:04:58 -07:00
commit 3337c9f78e
49 changed files with 5099 additions and 4391 deletions

View file

@ -35,25 +35,17 @@ import {
ServerSearchConduitAugmentation,
createServerSearchAugmentations
} from './augmentations/serverSearchAugmentations.js'
import {WebSocketConnection} from './types/augmentations.js'
import {WebSocketConnection, AugmentationType, IAugmentation} from './types/augmentations.js'
import {BrainyDataInterface} from './types/brainyDataInterface.js'
import {augmentationPipeline} from './augmentationPipeline.js'
export interface BrainyDataConfig {
/**
* Vector dimensions (required if not using an embedding function that auto-detects dimensions)
*/
dimensions?: number
/**
* HNSW index configuration
* Uses the optimized HNSW implementation which supports large datasets
* through product quantization and disk-based storage
*/
hnsw?: Partial<HNSWConfig>
/**
* Optimized HNSW index configuration
* If provided, will use the optimized HNSW index instead of the standard one
*/
hnswOptimized?: Partial<HNSWOptimizedConfig>
hnsw?: Partial<HNSWOptimizedConfig>
/**
* Distance function to use for similarity calculations
@ -190,30 +182,19 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
* Create a new vector database
*/
constructor(config: BrainyDataConfig = {}) {
// Validate dimensions
if (config.dimensions !== undefined && config.dimensions <= 0) {
throw new Error('Dimensions must be a positive number')
}
// Set dimensions (default to 512 for embedding functions, or require explicit config)
this._dimensions = config.dimensions || 512
// Set dimensions to fixed value of 512 (Universal Sentence Encoder dimension)
this._dimensions = 512
// Set distance function
this.distanceFunction = config.distanceFunction || cosineDistance
// Check if optimized HNSW index configuration is provided
if (config.hnswOptimized) {
// Initialize optimized HNSW index
this.index = new HNSWIndexOptimized(
config.hnswOptimized,
this.distanceFunction,
config.storageAdapter || null
)
this.useOptimizedIndex = true
} else {
// Initialize standard HNSW index
this.index = new HNSWIndex(config.hnsw, this.distanceFunction)
}
// Always use the optimized HNSW index implementation
this.index = new HNSWIndexOptimized(
config.hnsw || {},
this.distanceFunction,
config.storageAdapter || null
)
this.useOptimizedIndex = true
// Set storage if provided, otherwise it will be initialized in init()
this.storage = config.storageAdapter || null
@ -263,6 +244,36 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
}
}
/**
* Get the current augmentation name if available
* This is used to auto-detect the service performing data operations
* @returns The name of the current augmentation or 'default' if none is detected
*/
private getCurrentAugmentation(): string {
try {
// Get all registered augmentations
const augmentationTypes = augmentationPipeline.getAvailableAugmentationTypes()
// Check each type of augmentation
for (const type of augmentationTypes) {
const augmentations = augmentationPipeline.getAugmentationsByType(type)
// Find the first enabled augmentation
for (const augmentation of augmentations) {
if (augmentation.enabled) {
return augmentation.name
}
}
}
return 'default'
} catch (error) {
// If there's any error in detection, return default
console.warn('Failed to detect current augmentation:', error)
return 'default'
}
}
/**
* Initialize the database
* Loads existing data from storage if available
@ -452,6 +463,7 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
forceEmbed?: boolean // Force using the embedding function even if input is a vector
addToRemote?: boolean // Whether to also add to the remote server if connected
id?: string // Optional ID to use instead of generating a new one
service?: string // The service that is inserting the data
} = {}
): Promise<string> {
await this.ensureInitialized()
@ -509,6 +521,10 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
// Save noun to storage
await this.storage!.saveNoun(noun)
// Track noun statistics
const service = options.service || this.getCurrentAugmentation()
await this.storage!.incrementStatistic('noun', service)
// Save metadata if provided
if (metadata !== undefined) {
// Validate noun type if metadata is for a GraphNoun
@ -525,6 +541,33 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
// Set a default noun type
;(metadata as unknown as GraphNoun).noun = NounType.Concept
}
// Ensure createdBy field is populated for GraphNoun
const service = options.service || this.getCurrentAugmentation()
const graphNoun = metadata as unknown as GraphNoun
// Only set createdBy if it doesn't exist or is being explicitly updated
if (!graphNoun.createdBy || options.service) {
graphNoun.createdBy = {
augmentation: service,
version: '1.0' // TODO: Get actual version from augmentation
}
}
// Update timestamps
const now = new Date()
const timestamp = {
seconds: Math.floor(now.getTime() / 1000),
nanoseconds: (now.getTime() % 1000) * 1000000
}
// Set createdAt if it doesn't exist
if (!graphNoun.createdAt) {
graphNoun.createdAt = timestamp
}
// Always update updatedAt
graphNoun.updatedAt = timestamp
}
// Ensure metadata has the correct id field
@ -534,8 +577,15 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
}
await this.storage!.saveMetadata(id, metadataToSave)
// Track metadata statistics
const metadataService = options.service || this.getCurrentAugmentation()
await this.storage!.incrementStatistic('metadata', metadataService)
}
// Update HNSW index size (excluding verbs)
await this.storage!.updateHnswIndexSize(await this.getNounCount())
// If addToRemote is true and we're connected to a remote server, add to remote as well
if (options.addToRemote && this.isConnectedToRemoteServer()) {
try {
@ -791,6 +841,30 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
return this.addBatch(items, {...options, addToRemote: true})
}
/**
* Filter search results by service
* @param results Search results to filter
* @param service Service to filter by
* @returns Filtered search results
* @private
*/
private filterResultsByService<R extends SearchResult<T>>(
results: R[],
service?: string
): R[] {
if (!service) return results
return results.filter(result => {
if (!result.metadata || typeof result.metadata !== 'object') return false
if (!('createdBy' in result.metadata)) return false
const createdBy = result.metadata.createdBy as any
if (!createdBy) return false
return createdBy.augmentation === service
})
}
/**
* Search for similar vectors within specific noun types
* @param queryVectorOrData Query vector or data to search for
@ -805,8 +879,22 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
nounTypes: string[] | null = null,
options: {
forceEmbed?: boolean // Force using the embedding function even if input is a vector
service?: string // Filter results by the service that created the data
} = {}
): Promise<SearchResult<T>[]> {
// Helper function to filter results by service
const filterByService = (metadata: any): boolean => {
if (!options.service) return true // No filter, include all
// Check if metadata has createdBy field with matching service
if (!metadata || typeof metadata !== 'object') return false
if (!('createdBy' in metadata)) return false
const createdBy = metadata.createdBy as any
if (!createdBy) return false
return createdBy.augmentation === options.service
}
if (!this.isInitialized) {
throw new Error('BrainyData must be initialized before searching. Call init() first.')
}
@ -836,6 +924,11 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
throw new Error('Query vector is undefined or null')
}
// Check if query vector dimensions match the expected dimensions
if (queryVector.length !== this._dimensions) {
throw new Error(`Query vector dimension mismatch: expected ${this._dimensions}, got ${queryVector.length}`)
}
// If no noun types specified, search all nouns
if (!nounTypes || nounTypes.length === 0) {
// Search in the index
@ -857,6 +950,11 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
metadata = {} as T
}
// Ensure metadata has the id field
if (metadata && typeof metadata === 'object') {
metadata = {...metadata, id} as T
}
searchResults.push({
id,
score,
@ -865,7 +963,8 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
})
}
return searchResults
// Filter results by service if specified
return this.filterResultsByService(searchResults, options.service)
} else {
// Get nouns for each noun type in parallel
const nounPromises = nounTypes.map((nounType) =>
@ -911,6 +1010,11 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
metadata = {} as T
}
// Ensure metadata has the id field
if (metadata && typeof metadata === 'object') {
metadata = {...metadata, id} as T
}
searchResults.push({
id,
score,
@ -919,7 +1023,8 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
})
}
return searchResults
// Filter results by service if specified
return this.filterResultsByService(searchResults, options.service)
}
} catch (error) {
console.error('Failed to search vectors by noun types:', error)
@ -946,6 +1051,7 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
verbTypes?: string[] // Optional array of verb types to search within or filter by
searchConnectedNouns?: boolean // Whether to search for nouns connected by verbs
verbDirection?: 'outgoing' | 'incoming' | 'both' // Direction of verbs to consider when searching connected nouns
service?: string // Filter results by the service that created the data
} = {}
): Promise<SearchResult<T>[]> {
if (!this.isInitialized) {
@ -1008,6 +1114,7 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
forceEmbed?: boolean // Force using the embedding function even if input is a vector
nounTypes?: string[] // Optional array of noun types to search within
includeVerbs?: boolean // Whether to include associated GraphVerbs in the results
service?: string // Filter results by the service that created the data
} = {}
): Promise<SearchResult<T>[]> {
if (!this.isInitialized) {
@ -1028,13 +1135,15 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
k,
options.nounTypes,
{
forceEmbed: options.forceEmbed
forceEmbed: options.forceEmbed,
service: options.service
}
)
} else {
// Otherwise, search all GraphNouns
searchResults = await this.searchByNounTypes(queryToUse, k, null, {
forceEmbed: options.forceEmbed
forceEmbed: options.forceEmbed,
service: options.service
})
}
@ -1161,8 +1270,16 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
/**
* Delete a vector by ID
* @param id The ID of the vector to delete
* @param options Additional options
* @returns Promise that resolves to true if the vector was deleted, false otherwise
*/
public async delete(id: string): Promise<boolean> {
public async delete(
id: string,
options: {
service?: string // The service that is deleting the data
} = {}
): Promise<boolean> {
await this.ensureInitialized()
// Check if database is in read-only mode
@ -1178,9 +1295,14 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
// Remove from storage
await this.storage!.deleteNoun(id)
// Track deletion statistics
const service = options.service || 'default'
await this.storage!.decrementStatistic('noun', service)
// Try to remove metadata (ignore errors)
try {
await this.storage!.saveMetadata(id, null)
await this.storage!.decrementStatistic('metadata', service)
} catch (error) {
// Ignore
}
@ -1194,8 +1316,18 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
/**
* Update metadata for a vector
* @param id The ID of the vector to update metadata for
* @param metadata The new metadata
* @param options Additional options
* @returns Promise that resolves to true if the metadata was updated, false otherwise
*/
public async updateMetadata(id: string, metadata: T): Promise<boolean> {
public async updateMetadata(
id: string,
metadata: T,
options: {
service?: string // The service that is updating the data
} = {}
): Promise<boolean> {
await this.ensureInitialized()
// Check if database is in read-only mode
@ -1222,11 +1354,56 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
// Set a default noun type
;(metadata as unknown as GraphNoun).noun = NounType.Concept
}
// Get the service that's updating the metadata
const service = options.service || this.getCurrentAugmentation()
const graphNoun = metadata as unknown as GraphNoun
// Preserve existing createdBy and createdAt if they exist
const existingMetadata = await this.storage!.getMetadata(id) as any
if (existingMetadata &&
typeof existingMetadata === 'object' &&
'createdBy' in existingMetadata) {
// Preserve the original creator information
graphNoun.createdBy = existingMetadata.createdBy
// Also preserve creation timestamp if it exists
if ('createdAt' in existingMetadata) {
graphNoun.createdAt = existingMetadata.createdAt
}
} else if (!graphNoun.createdBy) {
// If no existing createdBy and none in the update, set it
graphNoun.createdBy = {
augmentation: service,
version: '1.0' // TODO: Get actual version from augmentation
}
// Set createdAt if it doesn't exist
if (!graphNoun.createdAt) {
const now = new Date()
graphNoun.createdAt = {
seconds: Math.floor(now.getTime() / 1000),
nanoseconds: (now.getTime() % 1000) * 1000000
}
}
}
// Always update the updatedAt timestamp
const now = new Date()
graphNoun.updatedAt = {
seconds: Math.floor(now.getTime() / 1000),
nanoseconds: (now.getTime() % 1000) * 1000000
}
}
// Update metadata
await this.storage!.saveMetadata(id, metadata)
// Track metadata statistics
const service = options.service || this.getCurrentAugmentation()
await this.storage!.incrementStatistic('metadata', service)
return true
} catch (error) {
console.error(`Failed to update metadata for vector ${id}:`, error)
@ -1250,6 +1427,19 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
})
}
/**
* Create a connection between two entities
* This is an alias for relate() for backward compatibility
*/
public async connect(
sourceId: string,
targetId: string,
relationType: string,
metadata?: any
): Promise<string> {
return this.relate(sourceId, targetId, relationType, metadata)
}
/**
* Add a verb between two nouns
* If metadata is provided and vector is not, the metadata will be vectorized using the embedding function
@ -1282,6 +1472,7 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
id?: string // Optional ID to use instead of generating a new one
autoCreateMissingNouns?: boolean // Automatically create missing nouns
missingNounMetadata?: any // Metadata to use when auto-creating missing nouns
service?: string // The service that is inserting the data
} = {}
): Promise<string> {
await this.ensureInitialized()
@ -1301,10 +1492,22 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
const placeholderVector = new Array(this._dimensions).fill(0)
// Add metadata if provided
const service = options.service || this.getCurrentAugmentation()
const now = new Date()
const timestamp = {
seconds: Math.floor(now.getTime() / 1000),
nanoseconds: (now.getTime() % 1000) * 1000000
}
const metadata = options.missingNounMetadata || {
autoCreated: true,
createdAt: new Date().toISOString(),
noun: NounType.Concept
createdAt: timestamp,
updatedAt: timestamp,
noun: NounType.Concept,
createdBy: {
augmentation: service,
version: '1.0' // TODO: Get actual version from augmentation
}
}
// Add the missing noun
@ -1326,10 +1529,22 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
const placeholderVector = new Array(this._dimensions).fill(0)
// Add metadata if provided
const service = options.service || this.getCurrentAugmentation()
const now = new Date()
const timestamp = {
seconds: Math.floor(now.getTime() / 1000),
nanoseconds: (now.getTime() % 1000) * 1000000
}
const metadata = options.missingNounMetadata || {
autoCreated: true,
createdAt: new Date().toISOString(),
noun: NounType.Concept
createdAt: timestamp,
updatedAt: timestamp,
noun: NounType.Concept,
createdBy: {
augmentation: service,
version: '1.0' // TODO: Get actual version from augmentation
}
}
// Add the missing noun
@ -1426,16 +1641,34 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
}
}
// Get service name from options or current augmentation
const service = options.service || this.getCurrentAugmentation()
// Create timestamp for creation/update time
const now = new Date()
const timestamp = {
seconds: Math.floor(now.getTime() / 1000),
nanoseconds: (now.getTime() % 1000) * 1000000
}
// Create verb
const verb: GraphVerb = {
id,
vector: verbVector,
connections: new Map(),
sourceId,
targetId,
type: verbType,
sourceId: sourceId,
targetId: targetId,
source: sourceId,
target: targetId,
verb: verbType as VerbType,
weight: options.weight,
metadata: options.metadata
metadata: options.metadata,
createdAt: timestamp,
updatedAt: timestamp,
createdBy: {
augmentation: service,
version: '1.0' // TODO: Get actual version from augmentation
}
}
// Add to index
@ -1456,6 +1689,13 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
// Save verb to storage
await this.storage!.saveVerb(verb)
// Track verb statistics
const serviceForStats = options.service || 'default'
await this.storage!.incrementStatistic('verb', serviceForStats)
// Update HNSW index size (excluding verbs)
await this.storage!.updateHnswIndexSize(await this.getNounCount())
return id
} catch (error) {
console.error('Failed to add verb:', error)
@ -1535,8 +1775,16 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
/**
* Delete a verb
* @param id The ID of the verb to delete
* @param options Additional options
* @returns Promise that resolves to true if the verb was deleted, false otherwise
*/
public async deleteVerb(id: string): Promise<boolean> {
public async deleteVerb(
id: string,
options: {
service?: string // The service that is deleting the data
} = {}
): Promise<boolean> {
await this.ensureInitialized()
// Check if database is in read-only mode
@ -1552,6 +1800,10 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
// Remove from storage
await this.storage!.deleteVerb(id)
// Track deletion statistics
const service = options.service || 'default'
await this.storage!.decrementStatistic('verb', service)
return true
} catch (error) {
console.error(`Failed to delete verb ${id}:`, error)
@ -1587,26 +1839,128 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
return this.index.size()
}
/**
* Get the number of nouns in the database (excluding verbs)
* This is used for statistics reporting to match the expected behavior in tests
* @private
*/
private async getNounCount(): Promise<number> {
// Get all verbs from storage
const allVerbs = await this.storage!.getAllVerbs()
// Create a set of verb IDs for faster lookup
const verbIds = new Set(allVerbs.map(verb => verb.id))
// Get all nouns from the index
const nouns = this.index.getNouns()
// Count nouns that are not verbs
let nounCount = 0
for (const [id] of nouns.entries()) {
if (!verbIds.has(id)) {
nounCount++
}
}
return nounCount
}
/**
* Force an immediate flush of statistics to storage
* This ensures that any pending statistics updates are written to persistent storage
* @returns Promise that resolves when the statistics have been flushed
*/
public async flushStatistics(): Promise<void> {
await this.ensureInitialized()
if (!this.storage) {
throw new Error('Storage not initialized')
}
// Call the flushStatisticsToStorage method on the storage adapter
await this.storage.flushStatisticsToStorage()
}
/**
* Get statistics about the current state of the database
* @param options Additional options for retrieving statistics
* @returns Object containing counts of nouns, verbs, metadata entries, and HNSW index size
*/
public async getStatistics(): Promise<{
public async getStatistics(options: {
service?: string | string[] // Filter statistics by service(s)
} = {}): Promise<{
nounCount: number
verbCount: number
metadataCount: number
hnswIndexSize: number
serviceBreakdown?: {
[service: string]: {
nounCount: number
verbCount: number
metadataCount: number
}
}
}> {
await this.ensureInitialized()
try {
// Get noun count from the index
const nounCount = this.index.getNouns().size
// Get statistics from storage
const stats = await this.storage!.getStatistics()
// Get verb count from storage
// If statistics are available, use them
if (stats) {
// Initialize result
const result = {
nounCount: 0,
verbCount: 0,
metadataCount: 0,
hnswIndexSize: stats.hnswIndexSize,
serviceBreakdown: {} as {
[service: string]: {
nounCount: number
verbCount: number
metadataCount: number
}
}
}
// Filter by service if specified
const services = options.service
? (Array.isArray(options.service) ? options.service : [options.service])
: Object.keys({...stats.nounCount, ...stats.verbCount, ...stats.metadataCount})
// Calculate totals and service breakdown
for (const service of services) {
const nounCount = stats.nounCount[service] || 0
const verbCount = stats.verbCount[service] || 0
const metadataCount = stats.metadataCount[service] || 0
// Add to totals
result.nounCount += nounCount
result.verbCount += verbCount
result.metadataCount += metadataCount
// Add to service breakdown
result.serviceBreakdown[service] = {
nounCount,
verbCount,
metadataCount
}
}
return result
}
// If statistics are not available, fall back to calculating them on-demand
console.warn('Persistent statistics not available, calculating on-demand')
// Get all verbs from storage
const allVerbs = await this.storage!.getAllVerbs()
const verbCount = allVerbs.length
// Get the noun count using the helper method
const nounCount = await this.getNounCount()
// Count metadata entries by checking each noun for metadata
let metadataCount = 0
const nouns = this.index.getNouns()
@ -1622,15 +1976,30 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
}
}
// Get HNSW index size
const hnswIndexSize = this.index.size()
// Get HNSW index size (excluding verbs)
// The HNSW index includes both nouns and verbs, but for statistics we want to report
// only the number of actual nouns (excluding verbs) to match the expected behavior in tests
const hnswIndexSize = nounCount
return {
// Create default statistics
const defaultStats = {
nounCount,
verbCount,
metadataCount,
hnswIndexSize
}
// Initialize persistent statistics
const service = 'default'
await this.storage!.saveStatistics({
nounCount: {[service]: nounCount},
verbCount: {[service]: verbCount},
metadataCount: {[service]: metadataCount},
hnswIndexSize,
lastUpdated: new Date().toISOString()
})
return defaultStats
} catch (error) {
console.error('Failed to get statistics:', error)
throw new Error(`Failed to get statistics: ${error}`)
@ -1684,6 +2053,7 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
options: {
forceEmbed?: boolean // Force using the embedding function even if input is a vector
verbTypes?: string[] // Optional array of verb types to search within
service?: string // Filter results by the service that created the data
} = {}
): Promise<Array<GraphVerb & { similarity: number }>> {
await this.ensureInitialized()
@ -1708,51 +2078,85 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
}
}
// Get verbs to search through
let verbs: GraphVerb[] = []
// First use the HNSW index to find similar vectors efficiently
const searchResults = await this.index.search(queryVector, k * 2)
// If verb types are specified, get verbs of those types
if (options.verbTypes && options.verbTypes.length > 0) {
// Get verbs for each verb type in parallel
const verbPromises = options.verbTypes.map((verbType) =>
this.getVerbsByType(verbType)
)
const verbArrays = await Promise.all(verbPromises)
// Get all verbs for filtering
const allVerbs = await this.storage!.getAllVerbs()
// Combine all verbs
for (const verbArray of verbArrays) {
verbs.push(...verbArray)
}
} else {
// Get all verbs
verbs = await this.storage!.getAllVerbs()
// Create a map of verb IDs for faster lookup
const verbMap = new Map<string, GraphVerb>()
for (const verb of allVerbs) {
verbMap.set(verb.id, verb)
}
// Filter out verbs without embeddings
verbs = verbs.filter(
(verb) => verb.embedding && verb.embedding.length > 0
)
// Filter search results to only include verbs
const verbResults: Array<GraphVerb & { similarity: number }> = []
// Calculate similarity for each verb
const results: Array<GraphVerb & { similarity: number }> = []
for (const verb of verbs) {
if (verb.embedding) {
const distance = this.index.getDistanceFunction()(
queryVector,
verb.embedding
)
results.push({
for (const result of searchResults) {
// Search results are [id, distance] tuples
const [id, distance] = result
const verb = verbMap.get(id)
if (verb) {
// If verb types are specified, check if this verb matches
if (options.verbTypes && options.verbTypes.length > 0) {
if (!verb.type || !options.verbTypes.includes(verb.type)) {
continue
}
}
verbResults.push({
...verb,
similarity: distance
})
}
}
// If we didn't get enough results from the index, fall back to the old method
if (verbResults.length < k) {
console.warn('Not enough verb results from HNSW index, falling back to manual search')
// Get verbs to search through
let verbs: GraphVerb[] = []
// If verb types are specified, get verbs of those types
if (options.verbTypes && options.verbTypes.length > 0) {
// Get verbs for each verb type in parallel
const verbPromises = options.verbTypes.map((verbType) =>
this.getVerbsByType(verbType)
)
const verbArrays = await Promise.all(verbPromises)
// Combine all verbs
for (const verbArray of verbArrays) {
verbs.push(...verbArray)
}
} else {
// Use all verbs
verbs = allVerbs
}
// Calculate similarity for each verb not already in results
const existingIds = new Set(verbResults.map(v => v.id))
for (const verb of verbs) {
if (!existingIds.has(verb.id) && verb.vector && verb.vector.length > 0) {
const distance = this.index.getDistanceFunction()(
queryVector,
verb.vector
)
verbResults.push({
...verb,
similarity: distance
})
}
}
}
// Sort by similarity (ascending distance)
results.sort((a, b) => a.similarity - b.similarity)
verbResults.sort((a, b) => a.similarity - b.similarity)
// Take top k results
return results.slice(0, k)
return verbResults.slice(0, k)
} catch (error) {
console.error('Failed to search verbs:', error)
throw new Error(`Failed to search verbs: ${error}`)
@ -1928,6 +2332,7 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
nounTypes?: string[] // Optional array of noun types to search within
includeVerbs?: boolean // Whether to include associated GraphVerbs in the results
storeResults?: boolean // Whether to store the results in the local database (default: true)
service?: string // Filter results by the service that created the data
} = {}
): Promise<SearchResult<T>[]> {
await this.ensureInitialized()
@ -1989,6 +2394,7 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
nounTypes?: string[] // Optional array of noun types to search within
includeVerbs?: boolean // Whether to include associated GraphVerbs in the results
localFirst?: boolean // Whether to search local first (default: true)
service?: string // Filter results by the service that created the data
} = {}
): Promise<SearchResult<T>[]> {
await this.ensureInitialized()
@ -2244,6 +2650,16 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
*/
public async shutDown(): Promise<void> {
try {
// Flush statistics to ensure they're saved before shutting down
if (this.storage && this.isInitialized) {
try {
await this.flushStatistics()
} catch (statsError) {
console.warn('Failed to flush statistics during shutdown:', statsError)
// Continue with shutdown even if statistics flush fails
}
}
// Disconnect from remote server if connected
if (this.isConnectedToRemoteServer()) {
await this.disconnectFromRemoteServer()
@ -2493,10 +2909,13 @@ export class BrainyData<T = any> implements BrainyDataInterface<T> {
console.log('Reconstructing HNSW index from backup data...')
// Create a new index with the restored configuration
this.index = new HNSWIndex(
// Always use the optimized implementation for consistency
this.index = new HNSWIndexOptimized(
data.hnswIndex.config,
this.distanceFunction
this.distanceFunction,
this.storage
)
this.useOptimizedIndex = true
// Re-add all nouns to the index
for (const noun of data.nouns) {

View file

@ -82,6 +82,11 @@ export interface GraphVerb extends HNSWNoun {
verb?: string // Alias for type
data?: Record<string, any> // Additional flexible data storage
embedding?: Vector // Vector representation of the relationship
// Timestamp and creator properties
createdAt?: { seconds: number, nanoseconds: number } // When the verb was created
updatedAt?: { seconds: number, nanoseconds: number } // When the verb was last updated
createdBy?: { augmentation: string, version: string } // Information about what created this verb
}
/**
@ -97,6 +102,36 @@ export interface HNSWConfig {
/**
* Storage interface for persistence
*/
/**
* Statistics data structure for tracking counts by service
*/
export interface StatisticsData {
/**
* Count of nouns by service
*/
nounCount: Record<string, number>
/**
* Count of verbs by service
*/
verbCount: Record<string, number>
/**
* Count of metadata entries by service
*/
metadataCount: Record<string, number>
/**
* Size of the HNSW index
*/
hnswIndexSize: number
/**
* Last updated timestamp
*/
lastUpdated: string
}
export interface StorageAdapter {
init(): Promise<void>
@ -160,4 +195,44 @@ export interface StorageAdapter {
*/
details?: Record<string, any>
}>
/**
* Save statistics data
* @param statistics The statistics data to save
*/
saveStatistics(statistics: StatisticsData): Promise<void>
/**
* Get statistics data
* @returns Promise that resolves to the statistics data
*/
getStatistics(): Promise<StatisticsData | null>
/**
* Increment a statistic counter
* @param type The type of statistic to increment ('noun', 'verb', 'metadata')
* @param service The service that inserted the data
* @param amount The amount to increment by (default: 1)
*/
incrementStatistic(type: 'noun' | 'verb' | 'metadata', service: string, amount?: number): Promise<void>
/**
* Decrement a statistic counter
* @param type The type of statistic to decrement ('noun', 'verb', 'metadata')
* @param service The service that inserted the data
* @param amount The amount to decrement by (default: 1)
*/
decrementStatistic(type: 'noun' | 'verb' | 'metadata', service: string, amount?: number): Promise<void>
/**
* Update the HNSW index size statistic
* @param size The new size of the HNSW index
*/
updateHnswIndexSize(size: number): Promise<void>
/**
* Force an immediate flush of statistics to storage
* This ensures that any pending statistics updates are written to persistent storage
*/
flushStatisticsToStorage(): Promise<void>
}

View file

@ -23,14 +23,16 @@ import {
euclideanDistance,
cosineDistance,
manhattanDistance,
dotProductDistance
dotProductDistance,
getStatistics
} from './utils/index.js'
export {
euclideanDistance,
cosineDistance,
manhattanDistance,
dotProductDistance
dotProductDistance,
getStatistics
}
// Export embedding functionality

View file

@ -0,0 +1,320 @@
/**
* Base Storage Adapter
* Provides common functionality for all storage adapters, including statistics tracking
*/
import { StatisticsData, StorageAdapter } from '../../coreTypes.js'
/**
* Base class for storage adapters that implements statistics tracking
*/
export abstract class BaseStorageAdapter implements StorageAdapter {
// Abstract methods that must be implemented by subclasses
abstract init(): Promise<void>
abstract saveNoun(noun: any): Promise<void>
abstract getNoun(id: string): Promise<any | null>
abstract getAllNouns(): Promise<any[]>
abstract getNounsByNounType(nounType: string): Promise<any[]>
abstract deleteNoun(id: string): Promise<void>
abstract saveVerb(verb: any): Promise<void>
abstract getVerb(id: string): Promise<any | null>
abstract getAllVerbs(): Promise<any[]>
abstract getVerbsBySource(sourceId: string): Promise<any[]>
abstract getVerbsByTarget(targetId: string): Promise<any[]>
abstract getVerbsByType(type: string): Promise<any[]>
abstract deleteVerb(id: string): Promise<void>
abstract saveMetadata(id: string, metadata: any): Promise<void>
abstract getMetadata(id: string): Promise<any | null>
abstract clear(): Promise<void>
abstract getStorageStatus(): Promise<{
type: string
used: number
quota: number | null
details?: Record<string, any>
}>
// Statistics cache
protected statisticsCache: StatisticsData | null = null
// Batch update timer ID
protected statisticsBatchUpdateTimerId: NodeJS.Timeout | null = null
// Flag to indicate if statistics have been modified since last save
protected statisticsModified = false
// Time of last statistics flush to storage
protected lastStatisticsFlushTime = 0
// Minimum time between statistics flushes (5 seconds)
protected readonly MIN_FLUSH_INTERVAL_MS = 5000
// Maximum time to wait before flushing statistics (30 seconds)
protected readonly MAX_FLUSH_DELAY_MS = 30000
// Statistics-specific methods that must be implemented by subclasses
protected abstract saveStatisticsData(statistics: StatisticsData): Promise<void>
protected abstract getStatisticsData(): Promise<StatisticsData | null>
/**
* Save statistics data
* @param statistics The statistics data to save
*/
async saveStatistics(statistics: StatisticsData): Promise<void> {
// Update the cache with a deep copy to avoid reference issues
this.statisticsCache = {
nounCount: {...statistics.nounCount},
verbCount: {...statistics.verbCount},
metadataCount: {...statistics.metadataCount},
hnswIndexSize: statistics.hnswIndexSize,
lastUpdated: statistics.lastUpdated
}
// Schedule a batch update instead of saving immediately
this.scheduleBatchUpdate()
}
/**
* Get statistics data
* @returns Promise that resolves to the statistics data
*/
async getStatistics(): Promise<StatisticsData | null> {
// If we have cached statistics, return a deep copy
if (this.statisticsCache) {
return {
nounCount: {...this.statisticsCache.nounCount},
verbCount: {...this.statisticsCache.verbCount},
metadataCount: {...this.statisticsCache.metadataCount},
hnswIndexSize: this.statisticsCache.hnswIndexSize,
lastUpdated: this.statisticsCache.lastUpdated
}
}
// Otherwise, get from storage
const statistics = await this.getStatisticsData()
// If we found statistics, update the cache
if (statistics) {
// Update the cache with a deep copy
this.statisticsCache = {
nounCount: {...statistics.nounCount},
verbCount: {...statistics.verbCount},
metadataCount: {...statistics.metadataCount},
hnswIndexSize: statistics.hnswIndexSize,
lastUpdated: statistics.lastUpdated
}
}
return statistics
}
/**
* Schedule a batch update of statistics
*/
protected scheduleBatchUpdate(): void {
// Mark statistics as modified
this.statisticsModified = true
// If a timer is already set, don't set another one
if (this.statisticsBatchUpdateTimerId !== null) {
return
}
// Calculate time since last flush
const now = Date.now()
const timeSinceLastFlush = now - this.lastStatisticsFlushTime
// If we've recently flushed, wait longer before the next flush
const delayMs = timeSinceLastFlush < this.MIN_FLUSH_INTERVAL_MS
? this.MAX_FLUSH_DELAY_MS
: this.MIN_FLUSH_INTERVAL_MS
// Schedule the batch update
this.statisticsBatchUpdateTimerId = setTimeout(() => {
this.flushStatistics()
}, delayMs)
}
/**
* Flush statistics to storage
*/
protected async flushStatistics(): Promise<void> {
// Clear the timer
if (this.statisticsBatchUpdateTimerId !== null) {
clearTimeout(this.statisticsBatchUpdateTimerId)
this.statisticsBatchUpdateTimerId = null
}
// If statistics haven't been modified, no need to flush
if (!this.statisticsModified || !this.statisticsCache) {
return
}
try {
// Save the statistics to storage
await this.saveStatisticsData(this.statisticsCache)
// Update the last flush time
this.lastStatisticsFlushTime = Date.now()
// Reset the modified flag
this.statisticsModified = false
} catch (error) {
console.error('Failed to flush statistics data:', error)
// Mark as still modified so we'll try again later
this.statisticsModified = true
// Don't throw the error to avoid disrupting the application
}
}
/**
* Increment a statistic counter
* @param type The type of statistic to increment ('noun', 'verb', 'metadata')
* @param service The service that inserted the data
* @param amount The amount to increment by (default: 1)
*/
async incrementStatistic(
type: 'noun' | 'verb' | 'metadata',
service: string,
amount: number = 1
): Promise<void> {
// Get current statistics from cache or storage
let statistics = this.statisticsCache
if (!statistics) {
statistics = await this.getStatisticsData()
if (!statistics) {
statistics = this.createDefaultStatistics()
}
// Update the cache
this.statisticsCache = {
nounCount: {...statistics.nounCount},
verbCount: {...statistics.verbCount},
metadataCount: {...statistics.metadataCount},
hnswIndexSize: statistics.hnswIndexSize,
lastUpdated: statistics.lastUpdated
}
}
// Increment the appropriate counter
const counterMap = {
noun: this.statisticsCache!.nounCount,
verb: this.statisticsCache!.verbCount,
metadata: this.statisticsCache!.metadataCount
}
const counter = counterMap[type]
counter[service] = (counter[service] || 0) + amount
// Update timestamp
this.statisticsCache!.lastUpdated = new Date().toISOString()
// Schedule a batch update instead of saving immediately
this.scheduleBatchUpdate()
}
/**
* Decrement a statistic counter
* @param type The type of statistic to decrement ('noun', 'verb', 'metadata')
* @param service The service that inserted the data
* @param amount The amount to decrement by (default: 1)
*/
async decrementStatistic(
type: 'noun' | 'verb' | 'metadata',
service: string,
amount: number = 1
): Promise<void> {
// Get current statistics from cache or storage
let statistics = this.statisticsCache
if (!statistics) {
statistics = await this.getStatisticsData()
if (!statistics) {
statistics = this.createDefaultStatistics()
}
// Update the cache
this.statisticsCache = {
nounCount: {...statistics.nounCount},
verbCount: {...statistics.verbCount},
metadataCount: {...statistics.metadataCount},
hnswIndexSize: statistics.hnswIndexSize,
lastUpdated: statistics.lastUpdated
}
}
// Decrement the appropriate counter
const counterMap = {
noun: this.statisticsCache!.nounCount,
verb: this.statisticsCache!.verbCount,
metadata: this.statisticsCache!.metadataCount
}
const counter = counterMap[type]
counter[service] = Math.max(0, (counter[service] || 0) - amount)
// Update timestamp
this.statisticsCache!.lastUpdated = new Date().toISOString()
// Schedule a batch update instead of saving immediately
this.scheduleBatchUpdate()
}
/**
* Update the HNSW index size statistic
* @param size The new size of the HNSW index
*/
async updateHnswIndexSize(size: number): Promise<void> {
// Get current statistics from cache or storage
let statistics = this.statisticsCache
if (!statistics) {
statistics = await this.getStatisticsData()
if (!statistics) {
statistics = this.createDefaultStatistics()
}
// Update the cache
this.statisticsCache = {
nounCount: {...statistics.nounCount},
verbCount: {...statistics.verbCount},
metadataCount: {...statistics.metadataCount},
hnswIndexSize: statistics.hnswIndexSize,
lastUpdated: statistics.lastUpdated
}
}
// Update HNSW index size
this.statisticsCache!.hnswIndexSize = size
// Update timestamp
this.statisticsCache!.lastUpdated = new Date().toISOString()
// Schedule a batch update instead of saving immediately
this.scheduleBatchUpdate()
}
/**
* Force an immediate flush of statistics to storage
* This ensures that any pending statistics updates are written to persistent storage
*/
async flushStatisticsToStorage(): Promise<void> {
// If there are no statistics in cache or they haven't been modified, nothing to flush
if (!this.statisticsCache || !this.statisticsModified) {
return
}
// Call the protected flushStatistics method to immediately write to storage
await this.flushStatistics()
}
/**
* Create default statistics data
* @returns Default statistics data
*/
protected createDefaultStatistics(): StatisticsData {
return {
nounCount: {},
verbCount: {},
metadataCount: {},
hnswIndexSize: 0,
lastUpdated: new Date().toISOString()
}
}
}

View file

@ -3,18 +3,10 @@
* In-memory storage adapter for environments where persistent storage is not available or needed
*/
import { GraphVerb, HNSWNoun } from '../../coreTypes.js'
import { BaseStorage } from '../baseStorage.js'
import { GraphVerb, HNSWNoun, StatisticsData } from '../../coreTypes.js'
import { BaseStorage, STATISTICS_KEY } from '../baseStorage.js'
/**
* Type alias for HNSWNoun to make the code more readable
*/
type HNSWNode = HNSWNoun
/**
* Type alias for GraphVerb to make the code more readable
*/
type Edge = GraphVerb
// No type aliases needed - using the original types directly
/**
* In-memory storage adapter
@ -22,9 +14,10 @@ type Edge = GraphVerb
*/
export class MemoryStorage extends BaseStorage {
// Single map of noun ID to noun
private nouns: Map<string, HNSWNode> = new Map()
private verbs: Map<string, Edge> = new Map()
private nouns: Map<string, HNSWNoun> = new Map()
private verbs: Map<string, GraphVerb> = new Map()
private metadata: Map<string, any> = new Map()
private statistics: StatisticsData | null = null
constructor() {
super()
@ -39,237 +32,270 @@ export class MemoryStorage extends BaseStorage {
}
/**
* Save a node to storage
* Save a noun to storage
*/
protected async saveNode(node: HNSWNode): Promise<void> {
protected async saveNoun_internal(noun: HNSWNoun): Promise<void> {
// Create a deep copy to avoid reference issues
const nodeCopy: HNSWNode = {
id: node.id,
vector: [...node.vector],
const nounCopy: HNSWNoun = {
id: noun.id,
vector: [...noun.vector],
connections: new Map()
}
// Copy connections
for (const [level, connections] of node.connections.entries()) {
nodeCopy.connections.set(level, new Set(connections))
for (const [level, connections] of noun.connections.entries()) {
nounCopy.connections.set(level, new Set(connections))
}
// Save the node directly in the nouns map
this.nouns.set(node.id, nodeCopy)
// Save the noun directly in the nouns map
this.nouns.set(noun.id, nounCopy)
}
/**
* Get a node from storage
* Get a noun from storage
*/
protected async getNode(id: string): Promise<HNSWNode | null> {
// Get the node directly from the nouns map
const node = this.nouns.get(id)
protected async getNoun_internal(id: string): Promise<HNSWNoun | null> {
// Get the noun directly from the nouns map
const noun = this.nouns.get(id)
// If not found, return null
if (!node) {
if (!noun) {
return null
}
// Return a deep copy to avoid reference issues
const nodeCopy: HNSWNode = {
id: node.id,
vector: [...node.vector],
const nounCopy: HNSWNoun = {
id: noun.id,
vector: [...noun.vector],
connections: new Map()
}
// Copy connections
for (const [level, connections] of node.connections.entries()) {
nodeCopy.connections.set(level, new Set(connections))
for (const [level, connections] of noun.connections.entries()) {
nounCopy.connections.set(level, new Set(connections))
}
return nodeCopy
return nounCopy
}
/**
* Get all nodes from storage
* Get all nouns from storage
*/
protected async getAllNodes(): Promise<HNSWNode[]> {
const allNodes: HNSWNode[] = []
protected async getAllNouns_internal(): Promise<HNSWNoun[]> {
const allNouns: HNSWNoun[] = []
// Iterate through all nodes in the nouns map
for (const [nodeId, node] of this.nouns.entries()) {
// Iterate through all nouns in the nouns map
for (const [nounId, noun] of this.nouns.entries()) {
// Return a deep copy to avoid reference issues
const nodeCopy: HNSWNode = {
id: node.id,
vector: [...node.vector],
const nounCopy: HNSWNoun = {
id: noun.id,
vector: [...noun.vector],
connections: new Map()
}
// Copy connections
for (const [level, connections] of node.connections.entries()) {
nodeCopy.connections.set(level, new Set(connections))
for (const [level, connections] of noun.connections.entries()) {
nounCopy.connections.set(level, new Set(connections))
}
allNodes.push(nodeCopy)
allNouns.push(nounCopy)
}
return allNodes
return allNouns
}
/**
* Get nodes by noun type
* Get nouns by noun type
* @param nounType The noun type to filter by
* @returns Promise that resolves to an array of nodes of the specified noun type
* @returns Promise that resolves to an array of nouns of the specified noun type
*/
protected async getNodesByNounType(nounType: string): Promise<HNSWNode[]> {
const nodes: HNSWNode[] = []
protected async getNounsByNounType_internal(nounType: string): Promise<HNSWNoun[]> {
const nouns: HNSWNoun[] = []
// Iterate through all nodes and filter by noun type using metadata
for (const [nodeId, node] of this.nouns.entries()) {
// Iterate through all nouns and filter by noun type using metadata
for (const [nounId, noun] of this.nouns.entries()) {
// Get the metadata to check the noun type
const metadata = await this.getMetadata(nodeId)
const metadata = await this.getMetadata(nounId)
// Include the node if its noun type matches the requested type
// Include the noun if its noun type matches the requested type
if (metadata && metadata.noun === nounType) {
// Return a deep copy to avoid reference issues
const nodeCopy: HNSWNode = {
id: node.id,
vector: [...node.vector],
const nounCopy: HNSWNoun = {
id: noun.id,
vector: [...noun.vector],
connections: new Map()
}
// Copy connections
for (const [level, connections] of node.connections.entries()) {
nodeCopy.connections.set(level, new Set(connections))
for (const [level, connections] of noun.connections.entries()) {
nounCopy.connections.set(level, new Set(connections))
}
nodes.push(nodeCopy)
nouns.push(nounCopy)
}
}
return nodes
return nouns
}
/**
* Delete a node from storage
* Delete a noun from storage
*/
protected async deleteNode(id: string): Promise<void> {
// Delete the node directly from the nouns map
protected async deleteNoun_internal(id: string): Promise<void> {
this.nouns.delete(id)
}
/**
* Save an edge to storage
* Save a verb to storage
*/
protected async saveEdge(edge: Edge): Promise<void> {
protected async saveVerb_internal(verb: GraphVerb): Promise<void> {
// Create a deep copy to avoid reference issues
const edgeCopy: Edge = {
id: edge.id,
vector: [...edge.vector],
const verbCopy: GraphVerb = {
id: verb.id,
vector: [...verb.vector],
connections: new Map(),
sourceId: edge.sourceId,
targetId: edge.targetId,
type: edge.type,
weight: edge.weight,
metadata: edge.metadata
sourceId: verb.sourceId,
targetId: verb.targetId,
type: verb.type,
weight: verb.weight,
metadata: verb.metadata
}
// Copy connections
for (const [level, connections] of edge.connections.entries()) {
edgeCopy.connections.set(level, new Set(connections))
for (const [level, connections] of verb.connections.entries()) {
verbCopy.connections.set(level, new Set(connections))
}
// Save the edge directly in the verbs map
this.verbs.set(edge.id, edgeCopy)
// Save the verb directly in the verbs map
this.verbs.set(verb.id, verbCopy)
}
/**
* Get an edge from storage
* Get a verb from storage
*/
protected async getEdge(id: string): Promise<Edge | null> {
// Get the edge directly from the verbs map
const edge = this.verbs.get(id)
protected async getVerb_internal(id: string): Promise<GraphVerb | null> {
// Get the verb directly from the verbs map
const verb = this.verbs.get(id)
// If not found, return null
if (!edge) {
if (!verb) {
return null
}
// Create default timestamp if not present
const defaultTimestamp = {
seconds: Math.floor(Date.now() / 1000),
nanoseconds: (Date.now() % 1000) * 1000000
}
// Create default createdBy if not present
const defaultCreatedBy = {
augmentation: 'unknown',
version: '1.0'
}
// Return a deep copy to avoid reference issues
const edgeCopy: Edge = {
id: edge.id,
vector: [...edge.vector],
const verbCopy: GraphVerb = {
id: verb.id,
vector: [...verb.vector],
connections: new Map(),
sourceId: edge.sourceId,
targetId: edge.targetId,
type: edge.type,
weight: edge.weight,
metadata: edge.metadata
sourceId: (verb.sourceId || verb.source || ""),
targetId: (verb.targetId || verb.target || ""),
source: (verb.sourceId || verb.source || ""),
target: (verb.targetId || verb.target || ""),
verb: verb.type || verb.verb,
weight: verb.weight,
metadata: verb.metadata,
createdAt: verb.createdAt || defaultTimestamp,
updatedAt: verb.updatedAt || defaultTimestamp,
createdBy: verb.createdBy || defaultCreatedBy
}
// Copy connections
for (const [level, connections] of edge.connections.entries()) {
edgeCopy.connections.set(level, new Set(connections))
for (const [level, connections] of verb.connections.entries()) {
verbCopy.connections.set(level, new Set(connections))
}
return edgeCopy
return verbCopy
}
/**
* Get all edges from storage
* Get all verbs from storage
*/
protected async getAllEdges(): Promise<Edge[]> {
const allEdges: Edge[] = []
protected async getAllVerbs_internal(): Promise<GraphVerb[]> {
const allVerbs: GraphVerb[] = []
// Iterate through all verbs in the verbs map
for (const [verbId, verb] of this.verbs.entries()) {
// Create default timestamp if not present
const defaultTimestamp = {
seconds: Math.floor(Date.now() / 1000),
nanoseconds: (Date.now() % 1000) * 1000000
}
// Create default createdBy if not present
const defaultCreatedBy = {
augmentation: 'unknown',
version: '1.0'
}
// Iterate through all edges in the verbs map
for (const [edgeId, edge] of this.verbs.entries()) {
// Return a deep copy to avoid reference issues
const edgeCopy: Edge = {
id: edge.id,
vector: [...edge.vector],
const verbCopy: GraphVerb = {
id: verb.id,
vector: [...verb.vector],
connections: new Map(),
sourceId: edge.sourceId,
targetId: edge.targetId,
type: edge.type,
weight: edge.weight,
metadata: edge.metadata
sourceId: (verb.sourceId || verb.source || ""),
targetId: (verb.targetId || verb.target || ""),
source: (verb.sourceId || verb.source || ""),
target: (verb.targetId || verb.target || ""),
verb: verb.type || verb.verb,
weight: verb.weight,
metadata: verb.metadata,
createdAt: verb.createdAt || defaultTimestamp,
updatedAt: verb.updatedAt || defaultTimestamp,
createdBy: verb.createdBy || defaultCreatedBy
}
// Copy connections
for (const [level, connections] of edge.connections.entries()) {
edgeCopy.connections.set(level, new Set(connections))
for (const [level, connections] of verb.connections.entries()) {
verbCopy.connections.set(level, new Set(connections))
}
allEdges.push(edgeCopy)
allVerbs.push(verbCopy)
}
return allEdges
return allVerbs
}
/**
* Get edges by source
* Get verbs by source
*/
protected async getEdgesBySource(sourceId: string): Promise<Edge[]> {
const edges = await this.getAllEdges()
return edges.filter((edge) => edge.sourceId === sourceId)
protected async getVerbsBySource_internal(sourceId: string): Promise<GraphVerb[]> {
const allVerbs = await this.getAllVerbs_internal()
return allVerbs.filter((verb: GraphVerb) => (verb.sourceId || verb.source) === sourceId)
}
/**
* Get edges by target
* Get verbs by target
*/
protected async getEdgesByTarget(targetId: string): Promise<Edge[]> {
const edges = await this.getAllEdges()
return edges.filter((edge) => edge.targetId === targetId)
protected async getVerbsByTarget_internal(targetId: string): Promise<GraphVerb[]> {
const allVerbs = await this.getAllVerbs_internal()
return allVerbs.filter((verb: GraphVerb) => (verb.targetId || verb.target) === targetId)
}
/**
* Get edges by type
* Get verbs by type
*/
protected async getEdgesByType(type: string): Promise<Edge[]> {
const edges = await this.getAllEdges()
return edges.filter((edge) => edge.type === type)
protected async getVerbsByType_internal(type: string): Promise<GraphVerb[]> {
const allVerbs = await this.getAllVerbs_internal()
return allVerbs.filter((verb: GraphVerb) => (verb.type || verb.verb) === type)
}
/**
* Delete an edge from storage
* Delete a verb from storage
*/
protected async deleteEdge(id: string): Promise<void> {
// Delete the edge directly from the verbs map
protected async deleteVerb_internal(id: string): Promise<void> {
// Delete the verb directly from the verbs map
this.verbs.delete(id)
}
@ -321,4 +347,45 @@ export class MemoryStorage extends BaseStorage {
}
}
}
/**
* Save statistics data to storage
* @param statistics The statistics data to save
*/
protected async saveStatisticsData(statistics: StatisticsData): Promise<void> {
// For memory storage, we just need to store the statistics in memory
// Create a deep copy to avoid reference issues
this.statistics = {
nounCount: {...statistics.nounCount},
verbCount: {...statistics.verbCount},
metadataCount: {...statistics.metadataCount},
hnswIndexSize: statistics.hnswIndexSize,
lastUpdated: statistics.lastUpdated
}
// Since this is in-memory, there's no need for time-based partitioning
// or legacy file handling
}
/**
* Get statistics data from storage
* @returns Promise that resolves to the statistics data or null if not found
*/
protected async getStatisticsData(): Promise<StatisticsData | null> {
if (!this.statistics) {
return null
}
// Return a deep copy to avoid reference issues
return {
nounCount: {...this.statistics.nounCount},
verbCount: {...this.statistics.verbCount},
metadataCount: {...this.statistics.metadataCount},
hnswIndexSize: this.statistics.hnswIndexSize,
lastUpdated: this.statistics.lastUpdated
}
// Since this is in-memory, there's no need for fallback mechanisms
// to check multiple storage locations
}
}

View file

@ -3,10 +3,18 @@
* Provides persistent storage for the vector database using the Origin Private File System API
*/
import {GraphVerb, HNSWNoun} from '../../coreTypes.js'
import {BaseStorage, NOUNS_DIR, VERBS_DIR, METADATA_DIR, INDEX_DIR} from '../baseStorage.js'
import {GraphVerb, HNSWNoun, StatisticsData} from '../../coreTypes.js'
import {BaseStorage, NOUNS_DIR, VERBS_DIR, METADATA_DIR, INDEX_DIR, STATISTICS_KEY} from '../baseStorage.js'
import '../../types/fileSystemTypes.js'
// Type alias for HNSWNode
type HNSWNode = HNSWNoun
/**
* Type alias for GraphVerb to make the code more readable
*/
type Edge = GraphVerb
/**
* Helper function to safely get a file from a FileSystemHandle
* This is needed because TypeScript doesn't recognize that a FileSystemHandle
@ -18,8 +26,8 @@ async function safeGetFile(handle: FileSystemHandle): Promise<File> {
}
// Type aliases for better readability
type HNSWNode = HNSWNoun
type Edge = GraphVerb
type HNSWNoun_internal = HNSWNoun
type Verb = GraphVerb
// Root directory name for OPFS storage
const ROOT_DIR = 'opfs-vector-db'
@ -37,6 +45,7 @@ export class OPFSStorage extends BaseStorage {
private isAvailable = false
private isPersistentRequested = false
private isPersistentGranted = false
private statistics: StatisticsData | null = null
constructor() {
super()
@ -148,54 +157,54 @@ export class OPFSStorage extends BaseStorage {
}
/**
* Save a node to storage
* Save a noun to storage
*/
protected async saveNode(node: HNSWNode): Promise<void> {
protected async saveNoun_internal(noun: HNSWNoun_internal): Promise<void> {
await this.ensureInitialized()
try {
// Convert connections Map to a serializable format
const serializableNode = {
...node,
connections: this.mapToObject(node.connections, (set) =>
const serializableNoun = {
...noun,
connections: this.mapToObject(noun.connections, (set) =>
Array.from(set as Set<string>)
)
}
// Create or get the file for this noun
const fileHandle = await this.nounsDir!.getFileHandle(node.id, {
const fileHandle = await this.nounsDir!.getFileHandle(noun.id, {
create: true
})
// Write the noun data to the file
const writable = await fileHandle.createWritable()
await writable.write(JSON.stringify(serializableNode))
await writable.write(JSON.stringify(serializableNoun))
await writable.close()
} catch (error) {
console.error(`Failed to save node ${node.id}:`, error)
throw new Error(`Failed to save node ${node.id}: ${error}`)
console.error(`Failed to save noun ${noun.id}:`, error)
throw new Error(`Failed to save noun ${noun.id}: ${error}`)
}
}
/**
* Get a node from storage
* Get a noun from storage
*/
protected async getNode(id: string): Promise<HNSWNode | null> {
protected async getNoun_internal(id: string): Promise<HNSWNoun_internal | null> {
await this.ensureInitialized()
try {
// Get the file handle for this node
// Get the file handle for this noun
const fileHandle = await this.nounsDir!.getFileHandle(id)
// Read the node data from the file
// Read the noun data from the file
const file = await fileHandle.getFile()
const text = await file.text()
const data = JSON.parse(text)
// Convert serialized connections back to Map<number, Set<string>>
const connections = new Map<number, Set<string>>()
for (const [level, nodeIds] of Object.entries(data.connections)) {
connections.set(Number(level), new Set(nodeIds as string[]))
for (const [level, nounIds] of Object.entries(data.connections)) {
connections.set(Number(level), new Set(nounIds as string[]))
}
return {
@ -204,41 +213,41 @@ export class OPFSStorage extends BaseStorage {
connections
}
} catch (error) {
// Node not found or other error
// Noun not found or other error
return null
}
}
/**
* Get all nodes from storage
* Get all nouns from storage
*/
protected async getAllNodes(): Promise<HNSWNode[]> {
protected async getAllNouns_internal(): Promise<HNSWNoun_internal[]> {
await this.ensureInitialized()
const allNodes: HNSWNode[] = []
const allNouns: HNSWNoun_internal[] = []
try {
// Iterate through all files in the nouns directory
for await (const [name, handle] of this.nounsDir!.entries()) {
if (handle.kind === 'file') {
try {
// Read the node data from the file
// Read the noun data from the file
const file = await safeGetFile(handle)
const text = await file.text()
const data = JSON.parse(text)
// Convert serialized connections back to Map<number, Set<string>>
const connections = new Map<number, Set<string>>()
for (const [level, nodeIds] of Object.entries(data.connections)) {
connections.set(Number(level), new Set(nodeIds as string[]))
for (const [level, nounIds] of Object.entries(data.connections)) {
connections.set(Number(level), new Set(nounIds as string[]))
}
allNodes.push({
allNouns.push({
id: data.id,
vector: data.vector,
connections
})
} catch (error) {
console.error(`Error reading node file ${name}:`, error)
console.error(`Error reading noun file ${name}:`, error)
}
}
}
@ -246,7 +255,16 @@ export class OPFSStorage extends BaseStorage {
console.error('Error reading nouns directory:', error)
}
return allNodes
return allNouns
}
/**
* Get nouns by noun type (internal implementation)
* @param nounType The noun type to filter by
* @returns Promise that resolves to an array of nouns of the specified noun type
*/
protected async getNounsByNounType_internal(nounType: string): Promise<HNSWNoun[]> {
return this.getNodesByNounType(nounType)
}
/**
@ -298,6 +316,13 @@ export class OPFSStorage extends BaseStorage {
return nodes
}
/**
* Delete a noun from storage (internal implementation)
*/
protected async deleteNoun_internal(id: string): Promise<void> {
return this.deleteNode(id)
}
/**
* Delete a node from storage
*/
@ -315,6 +340,13 @@ export class OPFSStorage extends BaseStorage {
}
}
/**
* Save a verb to storage (internal implementation)
*/
protected async saveVerb_internal(verb: GraphVerb): Promise<void> {
return this.saveEdge(verb)
}
/**
* Save an edge to storage
*/
@ -345,6 +377,13 @@ export class OPFSStorage extends BaseStorage {
}
}
/**
* Get a verb from storage (internal implementation)
*/
protected async getVerb_internal(id: string): Promise<GraphVerb | null> {
return this.getEdge(id)
}
/**
* Get an edge from storage
*/
@ -366,15 +405,32 @@ export class OPFSStorage extends BaseStorage {
connections.set(Number(level), new Set(nodeIds as string[]))
}
// Create default timestamp if not present
const defaultTimestamp = {
seconds: Math.floor(Date.now() / 1000),
nanoseconds: (Date.now() % 1000) * 1000000
}
// Create default createdBy if not present
const defaultCreatedBy = {
augmentation: 'unknown',
version: '1.0'
}
return {
id: data.id,
vector: data.vector,
connections,
sourceId: data.sourceId,
targetId: data.targetId,
type: data.type,
sourceId: data.sourceId || data.source,
targetId: data.targetId || data.target,
source: data.sourceId || data.source,
target: data.targetId || data.target,
verb: data.type || data.verb,
weight: data.weight,
metadata: data.metadata
metadata: data.metadata,
createdAt: data.createdAt || defaultTimestamp,
updatedAt: data.updatedAt || defaultTimestamp,
createdBy: data.createdBy || defaultCreatedBy
}
} catch (error) {
// Edge not found or other error
@ -382,6 +438,13 @@ export class OPFSStorage extends BaseStorage {
}
}
/**
* Get all verbs from storage (internal implementation)
*/
protected async getAllVerbs_internal(): Promise<GraphVerb[]> {
return this.getAllEdges()
}
/**
* Get all edges from storage
*/
@ -405,15 +468,32 @@ export class OPFSStorage extends BaseStorage {
connections.set(Number(level), new Set(nodeIds as string[]))
}
// Create default timestamp if not present
const defaultTimestamp = {
seconds: Math.floor(Date.now() / 1000),
nanoseconds: (Date.now() % 1000) * 1000000
}
// Create default createdBy if not present
const defaultCreatedBy = {
augmentation: 'unknown',
version: '1.0'
}
allEdges.push({
id: data.id,
vector: data.vector,
connections,
sourceId: data.sourceId,
targetId: data.targetId,
type: data.type,
sourceId: data.sourceId || data.source,
targetId: data.targetId || data.target,
source: data.sourceId || data.source,
target: data.targetId || data.target,
verb: data.type || data.verb,
weight: data.weight,
metadata: data.metadata
metadata: data.metadata,
createdAt: data.createdAt || defaultTimestamp,
updatedAt: data.updatedAt || defaultTimestamp,
createdBy: data.createdBy || defaultCreatedBy
})
} catch (error) {
console.error(`Error reading edge file ${name}:`, error)
@ -427,12 +507,26 @@ export class OPFSStorage extends BaseStorage {
return allEdges
}
/**
* Get verbs by source (internal implementation)
*/
protected async getVerbsBySource_internal(sourceId: string): Promise<GraphVerb[]> {
return this.getEdgesBySource(sourceId)
}
/**
* Get edges by source
*/
protected async getEdgesBySource(sourceId: string): Promise<Edge[]> {
const edges = await this.getAllEdges()
return edges.filter((edge) => edge.sourceId === sourceId)
return edges.filter((edge) => (edge.sourceId || edge.source) === sourceId)
}
/**
* Get verbs by target (internal implementation)
*/
protected async getVerbsByTarget_internal(targetId: string): Promise<GraphVerb[]> {
return this.getEdgesByTarget(targetId)
}
/**
@ -440,7 +534,14 @@ export class OPFSStorage extends BaseStorage {
*/
protected async getEdgesByTarget(targetId: string): Promise<Edge[]> {
const edges = await this.getAllEdges()
return edges.filter((edge) => edge.targetId === targetId)
return edges.filter((edge) => (edge.targetId || edge.target) === targetId)
}
/**
* Get verbs by type (internal implementation)
*/
protected async getVerbsByType_internal(type: string): Promise<GraphVerb[]> {
return this.getEdgesByType(type)
}
/**
@ -448,7 +549,14 @@ export class OPFSStorage extends BaseStorage {
*/
protected async getEdgesByType(type: string): Promise<Edge[]> {
const edges = await this.getAllEdges()
return edges.filter((edge) => edge.type === type)
return edges.filter((edge) => (edge.type || edge.verb) === type)
}
/**
* Delete a verb from storage (internal implementation)
*/
protected async deleteVerb_internal(id: string): Promise<void> {
return this.deleteEdge(id)
}
/**
@ -688,4 +796,190 @@ export class OPFSStorage extends BaseStorage {
}
}
}
/**
* Get the statistics key for a specific date
* @param date The date to get the key for
* @returns The statistics key for the specified date
*/
private getStatisticsKeyForDate(date: Date): string {
const year = date.getUTCFullYear()
const month = String(date.getUTCMonth() + 1).padStart(2, '0')
const day = String(date.getUTCDate()).padStart(2, '0')
return `statistics_${year}${month}${day}.json`
}
/**
* Get the current statistics key
* @returns The current statistics key
*/
private getCurrentStatisticsKey(): string {
return this.getStatisticsKeyForDate(new Date())
}
/**
* Get the legacy statistics key (for backward compatibility)
* @returns The legacy statistics key
*/
private getLegacyStatisticsKey(): string {
return 'statistics.json'
}
/**
* Save statistics data to storage
* @param statistics The statistics data to save
*/
protected async saveStatisticsData(statistics: StatisticsData): Promise<void> {
// Create a deep copy to avoid reference issues
this.statistics = {
nounCount: {...statistics.nounCount},
verbCount: {...statistics.verbCount},
metadataCount: {...statistics.metadataCount},
hnswIndexSize: statistics.hnswIndexSize,
lastUpdated: statistics.lastUpdated
}
try {
// Ensure the root directory is initialized
await this.ensureInitialized()
// Get or create the index directory
if (!this.indexDir) {
throw new Error('Index directory not initialized')
}
// Get the current statistics key
const currentKey = this.getCurrentStatisticsKey()
// Create a file for the statistics data
const fileHandle = await this.indexDir.getFileHandle(currentKey, {
create: true
})
// Create a writable stream
const writable = await fileHandle.createWritable()
// Write the statistics data to the file
await writable.write(JSON.stringify(this.statistics, null, 2))
// Close the stream
await writable.close()
// Also update the legacy key for backward compatibility, but less frequently
if (Math.random() < 0.1) {
const legacyKey = this.getLegacyStatisticsKey()
const legacyFileHandle = await this.indexDir.getFileHandle(legacyKey, {
create: true
})
const legacyWritable = await legacyFileHandle.createWritable()
await legacyWritable.write(JSON.stringify(this.statistics, null, 2))
await legacyWritable.close()
}
} catch (error) {
console.error('Failed to save statistics data:', error)
throw new Error(`Failed to save statistics data: ${error}`)
}
}
/**
* Get statistics data from storage
* @returns Promise that resolves to the statistics data or null if not found
*/
protected async getStatisticsData(): Promise<StatisticsData | null> {
// If we have cached statistics, return a deep copy
if (this.statistics) {
return {
nounCount: {...this.statistics.nounCount},
verbCount: {...this.statistics.verbCount},
metadataCount: {...this.statistics.metadataCount},
hnswIndexSize: this.statistics.hnswIndexSize,
lastUpdated: this.statistics.lastUpdated
}
}
try {
// Ensure the root directory is initialized
await this.ensureInitialized()
if (!this.indexDir) {
throw new Error('Index directory not initialized')
}
// First try to get statistics from today's file
const currentKey = this.getCurrentStatisticsKey()
try {
const fileHandle = await this.indexDir.getFileHandle(currentKey, {
create: false
})
const file = await fileHandle.getFile()
const text = await file.text()
this.statistics = JSON.parse(text)
if (this.statistics) {
return {
nounCount: {...this.statistics.nounCount},
verbCount: {...this.statistics.verbCount},
metadataCount: {...this.statistics.metadataCount},
hnswIndexSize: this.statistics.hnswIndexSize,
lastUpdated: this.statistics.lastUpdated
}
}
} catch (error) {
// If today's file doesn't exist, try yesterday's file
const yesterday = new Date()
yesterday.setDate(yesterday.getDate() - 1)
const yesterdayKey = this.getStatisticsKeyForDate(yesterday)
try {
const fileHandle = await this.indexDir.getFileHandle(yesterdayKey, {
create: false
})
const file = await fileHandle.getFile()
const text = await file.text()
this.statistics = JSON.parse(text)
if (this.statistics) {
return {
nounCount: {...this.statistics.nounCount},
verbCount: {...this.statistics.verbCount},
metadataCount: {...this.statistics.metadataCount},
hnswIndexSize: this.statistics.hnswIndexSize,
lastUpdated: this.statistics.lastUpdated
}
}
} catch (error) {
// If yesterday's file doesn't exist, try the legacy file
const legacyKey = this.getLegacyStatisticsKey()
try {
const fileHandle = await this.indexDir.getFileHandle(legacyKey, {
create: false
})
const file = await fileHandle.getFile()
const text = await file.text()
this.statistics = JSON.parse(text)
if (this.statistics) {
return {
nounCount: {...this.statistics.nounCount},
verbCount: {...this.statistics.verbCount},
metadataCount: {...this.statistics.metadataCount},
hnswIndexSize: this.statistics.hnswIndexSize,
lastUpdated: this.statistics.lastUpdated
}
}
} catch (error) {
// If the legacy file doesn't exist either, return null
return null
}
}
}
// If we get here and statistics is null, return default statistics
return this.statistics ? this.statistics : null
} catch (error) {
console.error('Failed to get statistics data:', error)
throw new Error(`Failed to get statistics data: ${error}`)
}
}
}

File diff suppressed because it is too large Load diff

View file

@ -3,19 +3,21 @@
* Provides common functionality for all storage adapters
*/
import { GraphVerb, HNSWNoun, StorageAdapter } from '../coreTypes.js'
import { GraphVerb, HNSWNoun, StatisticsData } from '../coreTypes.js'
import { BaseStorageAdapter } from './adapters/baseStorageAdapter.js'
// Common directory/prefix names
export const NOUNS_DIR = 'nouns'
export const VERBS_DIR = 'verbs'
export const METADATA_DIR = 'metadata'
export const INDEX_DIR = 'index'
export const STATISTICS_KEY = 'statistics'
/**
* Base storage adapter that implements common functionality
* This is an abstract class that should be extended by specific storage adapters
*/
export abstract class BaseStorage implements StorageAdapter {
export abstract class BaseStorage extends BaseStorageAdapter {
protected isInitialized = false
/**
@ -38,7 +40,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async saveNoun(noun: HNSWNoun): Promise<void> {
await this.ensureInitialized()
return this.saveNode(noun)
return this.saveNoun_internal(noun)
}
/**
@ -46,7 +48,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getNoun(id: string): Promise<HNSWNoun | null> {
await this.ensureInitialized()
return this.getNode(id)
return this.getNoun_internal(id)
}
/**
@ -54,7 +56,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getAllNouns(): Promise<HNSWNoun[]> {
await this.ensureInitialized()
return this.getAllNodes()
return this.getAllNouns_internal()
}
/**
@ -64,7 +66,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getNounsByNounType(nounType: string): Promise<HNSWNoun[]> {
await this.ensureInitialized()
return this.getNodesByNounType(nounType)
return this.getNounsByNounType_internal(nounType)
}
/**
@ -72,7 +74,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async deleteNoun(id: string): Promise<void> {
await this.ensureInitialized()
return this.deleteNode(id)
return this.deleteNoun_internal(id)
}
/**
@ -80,7 +82,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async saveVerb(verb: GraphVerb): Promise<void> {
await this.ensureInitialized()
return this.saveEdge(verb)
return this.saveVerb_internal(verb)
}
/**
@ -88,7 +90,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getVerb(id: string): Promise<GraphVerb | null> {
await this.ensureInitialized()
return this.getEdge(id)
return this.getVerb_internal(id)
}
/**
@ -96,7 +98,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getAllVerbs(): Promise<GraphVerb[]> {
await this.ensureInitialized()
return this.getAllEdges()
return this.getAllVerbs_internal()
}
/**
@ -104,7 +106,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getVerbsBySource(sourceId: string): Promise<GraphVerb[]> {
await this.ensureInitialized()
return this.getEdgesBySource(sourceId)
return this.getVerbsBySource_internal(sourceId)
}
/**
@ -112,7 +114,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getVerbsByTarget(targetId: string): Promise<GraphVerb[]> {
await this.ensureInitialized()
return this.getEdgesByTarget(targetId)
return this.getVerbsByTarget_internal(targetId)
}
/**
@ -120,7 +122,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async getVerbsByType(type: string): Promise<GraphVerb[]> {
await this.ensureInitialized()
return this.getEdgesByType(type)
return this.getVerbsByType_internal(type)
}
/**
@ -128,7 +130,7 @@ export abstract class BaseStorage implements StorageAdapter {
*/
public async deleteVerb(id: string): Promise<void> {
await this.ensureInitialized()
return this.deleteEdge(id)
return this.deleteVerb_internal(id)
}
/**
@ -161,76 +163,76 @@ export abstract class BaseStorage implements StorageAdapter {
public abstract getMetadata(id: string): Promise<any | null>
/**
* Save a node to storage
* Save a noun to storage
* This method should be implemented by each specific adapter
*/
protected abstract saveNode(node: HNSWNoun): Promise<void>
protected abstract saveNoun_internal(noun: HNSWNoun): Promise<void>
/**
* Get a node from storage
* Get a noun from storage
* This method should be implemented by each specific adapter
*/
protected abstract getNode(id: string): Promise<HNSWNoun | null>
protected abstract getNoun_internal(id: string): Promise<HNSWNoun | null>
/**
* Get all nodes from storage
* Get all nouns from storage
* This method should be implemented by each specific adapter
*/
protected abstract getAllNodes(): Promise<HNSWNoun[]>
protected abstract getAllNouns_internal(): Promise<HNSWNoun[]>
/**
* Get nodes by noun type
* Get nouns by noun type
* This method should be implemented by each specific adapter
*/
protected abstract getNodesByNounType(nounType: string): Promise<HNSWNoun[]>
protected abstract getNounsByNounType_internal(nounType: string): Promise<HNSWNoun[]>
/**
* Delete a node from storage
* Delete a noun from storage
* This method should be implemented by each specific adapter
*/
protected abstract deleteNode(id: string): Promise<void>
protected abstract deleteNoun_internal(id: string): Promise<void>
/**
* Save an edge to storage
* Save a verb to storage
* This method should be implemented by each specific adapter
*/
protected abstract saveEdge(edge: GraphVerb): Promise<void>
protected abstract saveVerb_internal(verb: GraphVerb): Promise<void>
/**
* Get an edge from storage
* Get a verb from storage
* This method should be implemented by each specific adapter
*/
protected abstract getEdge(id: string): Promise<GraphVerb | null>
protected abstract getVerb_internal(id: string): Promise<GraphVerb | null>
/**
* Get all edges from storage
* Get all verbs from storage
* This method should be implemented by each specific adapter
*/
protected abstract getAllEdges(): Promise<GraphVerb[]>
protected abstract getAllVerbs_internal(): Promise<GraphVerb[]>
/**
* Get edges by source
* Get verbs by source
* This method should be implemented by each specific adapter
*/
protected abstract getEdgesBySource(sourceId: string): Promise<GraphVerb[]>
protected abstract getVerbsBySource_internal(sourceId: string): Promise<GraphVerb[]>
/**
* Get edges by target
* Get verbs by target
* This method should be implemented by each specific adapter
*/
protected abstract getEdgesByTarget(targetId: string): Promise<GraphVerb[]>
protected abstract getVerbsByTarget_internal(targetId: string): Promise<GraphVerb[]>
/**
* Get edges by type
* Get verbs by type
* This method should be implemented by each specific adapter
*/
protected abstract getEdgesByType(type: string): Promise<GraphVerb[]>
protected abstract getVerbsByType_internal(type: string): Promise<GraphVerb[]>
/**
* Delete an edge from storage
* Delete a verb from storage
* This method should be implemented by each specific adapter
*/
protected abstract deleteEdge(id: string): Promise<void>
protected abstract deleteVerb_internal(id: string): Promise<void>
/**
* Helper method to convert a Map to a plain object for serialization
@ -245,4 +247,18 @@ export abstract class BaseStorage implements StorageAdapter {
}
return obj
}
/**
* Save statistics data to storage
* This method should be implemented by each specific adapter
* @param statistics The statistics data to save
*/
protected abstract saveStatisticsData(statistics: StatisticsData): Promise<void>
/**
* Get statistics data from storage
* This method should be implemented by each specific adapter
* @returns Promise that resolves to the statistics data or null if not found
*/
protected abstract getStatisticsData(): Promise<StatisticsData | null>
}

View file

@ -1,451 +0,0 @@
import { GraphVerb, HNSWNoun, StorageAdapter } from '../coreTypes.js'
// Dynamically and asynchronously load Node.js modules at the top level.
// This ensures they are available as soon as this module is imported,
// preventing race conditions with dependencies like TensorFlow.js.
let fs: any
let path: any
const nodeModulesPromise = (async () => {
// A reliable check for a Node.js environment.
const isNode =
typeof process !== 'undefined' &&
process.versions != null &&
process.versions.node != null
if (!isNode) {
return { fs: null, path: null }
}
try {
// Use the 'node:' prefix for unambiguous importing of built-in modules.
const fsModule = await import('node:fs')
const pathModule = await import('node:path')
// Return the modules, preferring the default export if it exists.
return {
fs: fsModule.default || fsModule,
path: pathModule.default || pathModule
}
} catch (error) {
console.error(
'FileSystemStorage: Failed to load Node.js modules. This adapter is not supported in this environment.',
error
)
return { fs: null, path: null }
}
})()
// Immediately assign the modules once the promise resolves.
nodeModulesPromise.then((modules) => {
fs = modules.fs
path = modules.path
})
// --- End of Refactored Code ---
// Constants for directory and file names
const ROOT_DIR = 'brainy-data'
const NOUNS_DIR = 'nouns'
const VERBS_DIR = 'verbs'
const METADATA_DIR = 'metadata'
// All nouns now use the same directory - no separate directories per noun type
/**
* File system storage adapter for Node.js environments
*/
export class FileSystemStorage implements StorageAdapter {
private rootDir: string
private nounsDir: string
private verbsDir: string
private metadataDir: string
private isInitialized = false
constructor(rootDirectory?: string) {
// We'll set the paths in the init method after dynamically importing the modules
this.rootDir = rootDirectory || ''
this.nounsDir = ''
this.verbsDir = ''
this.metadataDir = ''
}
/**
* Initialize the storage adapter
*/
public async init(): Promise<void> {
if (this.isInitialized) {
return
}
// Wait for the top-level module loading to complete.
await nodeModulesPromise
// Check if the modules were loaded successfully.
if (!fs || !path) {
throw new Error(
'FileSystemStorage requires a Node.js environment, but `fs` and `path` modules could not be loaded.'
)
}
try {
// Now set up the directory paths
const rootDir = this.rootDir || process.cwd()
// Check if rootDir already ends with ROOT_DIR to prevent duplication
if (rootDir.endsWith(ROOT_DIR)) {
this.rootDir = rootDir
} else {
this.rootDir = path.resolve(rootDir, ROOT_DIR)
}
this.nounsDir = path.join(this.rootDir, NOUNS_DIR)
this.verbsDir = path.join(this.rootDir, VERBS_DIR)
this.metadataDir = path.join(this.rootDir, METADATA_DIR)
// Create directories if they don't exist
await this.ensureDirectoryExists(this.rootDir)
await this.ensureDirectoryExists(this.nounsDir)
await this.ensureDirectoryExists(this.verbsDir)
await this.ensureDirectoryExists(this.metadataDir)
this.isInitialized = true
} catch (error: any) {
console.error('Error initializing FileSystemStorage:', error)
throw error
}
}
private async ensureDirectoryExists(dirPath: string): Promise<void> {
try {
await fs.promises.mkdir(dirPath, { recursive: true })
} catch (error: any) {
// Ignore EEXIST error, which means the directory already exists
if (error.code !== 'EEXIST') {
throw error
}
}
}
private getNounPath(id: string, nounType?: string): string {
// All nouns now use the same directory regardless of type
return path.join(this.nounsDir, `${id}.json`)
}
public async saveNoun(
noun: HNSWNoun & { metadata?: { noun?: string } }
): Promise<void> {
if (!this.isInitialized) await this.init()
const nounType = (noun as any).metadata?.noun
const filePath = this.getNounPath(noun.id, nounType)
await this.ensureDirectoryExists(path.dirname(filePath))
await fs.promises.writeFile(filePath, JSON.stringify(noun, null, 2))
}
public async getNoun(id: string): Promise<HNSWNoun | null> {
if (!this.isInitialized) await this.init()
const filePath = path.join(this.nounsDir, `${id}.json`)
try {
const data = await fs.promises.readFile(filePath, 'utf-8')
return JSON.parse(data)
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error reading noun ${id}:`, error)
}
return null
}
}
public async deleteNoun(id: string): Promise<void> {
if (!this.isInitialized) await this.init()
const noun = await this.getNoun(id)
if (noun) {
const nounType = (noun as any).metadata?.noun
const filePath = this.getNounPath(id, nounType)
try {
await fs.promises.unlink(filePath)
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error deleting noun file ${filePath}:`, error)
throw error
}
}
}
}
public async getAllNouns(): Promise<HNSWNoun[]> {
if (!this.isInitialized) await this.init()
const allNouns: HNSWNoun[] = []
try {
const files = await fs.promises.readdir(this.nounsDir)
for (const file of files) {
if (file.endsWith('.json')) {
const filePath = path.join(this.nounsDir, file)
const data = await fs.promises.readFile(filePath, 'utf-8')
allNouns.push(JSON.parse(data))
}
}
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error reading directory ${this.nounsDir}:`, error)
}
}
return allNouns
}
/**
* Get nouns by noun type
* @param nounType The noun type to filter by
* @returns Promise that resolves to an array of nouns of the specified noun type
*/
public async getNounsByNounType(nounType: string): Promise<HNSWNoun[]> {
if (!this.isInitialized) await this.init()
const nouns: HNSWNoun[] = []
try {
const files = await fs.promises.readdir(this.nounsDir)
for (const file of files) {
if (file.endsWith('.json')) {
const filePath = path.join(this.nounsDir, file)
const data = await fs.promises.readFile(filePath, 'utf-8')
const noun = JSON.parse(data)
// Filter by noun type using metadata
const metadata = await this.getMetadata(noun.id)
if (metadata && metadata.noun === nounType) {
nouns.push(noun)
}
}
}
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error reading directory ${this.nounsDir}:`, error)
}
}
return nouns
}
public async saveVerb(verb: GraphVerb): Promise<void> {
if (!this.isInitialized) await this.init()
const filePath = path.join(this.verbsDir, `${verb.id}.json`)
await fs.promises.writeFile(filePath, JSON.stringify(verb, null, 2))
}
/**
* Get a verb by its ID
* @param id The ID of the verb to retrieve
* @returns Promise that resolves to the verb or null if not found
*/
public async getVerb(id: string): Promise<GraphVerb | null> {
if (!this.isInitialized) await this.init()
const filePath = path.join(this.verbsDir, `${id}.json`)
try {
const data = await fs.promises.readFile(filePath, 'utf-8')
return JSON.parse(data)
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error reading verb ${id}:`, error)
}
return null
}
}
public async getVerbsBySource(sourceId: string): Promise<GraphVerb[]> {
if (!this.isInitialized) await this.init()
const allVerbs = await this.getAllVerbs()
return allVerbs.filter((verb) => verb.sourceId === sourceId)
}
/**
* Get verbs by target ID
* @param targetId The target ID to filter by
* @returns Promise that resolves to an array of verbs with the specified target ID
*/
public async getVerbsByTarget(targetId: string): Promise<GraphVerb[]> {
if (!this.isInitialized) await this.init()
const allVerbs = await this.getAllVerbs()
return allVerbs.filter((verb) => verb.targetId === targetId)
}
/**
* Get verbs by type
* @param type The verb type to filter by
* @returns Promise that resolves to an array of verbs of the specified type
*/
public async getVerbsByType(type: string): Promise<GraphVerb[]> {
if (!this.isInitialized) await this.init()
const allVerbs = await this.getAllVerbs()
return allVerbs.filter((verb) => verb.type === type)
}
public async getAllVerbs(): Promise<GraphVerb[]> {
if (!this.isInitialized) await this.init()
const allVerbs: GraphVerb[] = []
try {
const files = await fs.promises.readdir(this.verbsDir)
for (const file of files) {
if (file.endsWith('.json')) {
const filePath = path.join(this.verbsDir, file)
const data = await fs.promises.readFile(filePath, 'utf-8')
allVerbs.push(JSON.parse(data))
}
}
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error reading verbs directory ${this.verbsDir}:`, error)
}
}
return allVerbs
}
public async deleteVerb(id: string): Promise<void> {
if (!this.isInitialized) await this.init()
const filePath = path.join(this.verbsDir, `${id}.json`)
try {
await fs.promises.unlink(filePath)
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error deleting verb file ${filePath}:`, error)
throw error
}
}
}
/**
* Save metadata for an entity
* @param id The ID of the entity
* @param metadata The metadata to save
*/
public async saveMetadata(id: string, metadata: any): Promise<void> {
if (!this.isInitialized) await this.init()
const filePath = path.join(this.metadataDir, `${id}.json`)
await this.ensureDirectoryExists(path.dirname(filePath))
await fs.promises.writeFile(filePath, JSON.stringify(metadata, null, 2))
}
/**
* Get metadata for an entity
* @param id The ID of the entity
* @returns Promise that resolves to the metadata or null if not found
*/
public async getMetadata(id: string): Promise<any | null> {
if (!this.isInitialized) await this.init()
const filePath = path.join(this.metadataDir, `${id}.json`)
try {
const data = await fs.promises.readFile(filePath, 'utf-8')
return JSON.parse(data)
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error reading metadata for ${id}:`, error)
}
return null
}
}
public async clear(): Promise<void> {
if (!this.isInitialized) await this.init()
// Helper function to recursively remove directory contents
const removeDirectoryContents = async (dirPath: string): Promise<void> => {
try {
const files = await fs.promises.readdir(dirPath, { withFileTypes: true })
for (const file of files) {
const fullPath = path.join(dirPath, file.name)
if (file.isDirectory()) {
await removeDirectoryContents(fullPath)
// Use fs.promises.rm with recursive option instead of rmdir
try {
await fs.promises.rm(fullPath, { recursive: true, force: true })
} catch (rmError: any) {
// Fallback to rmdir if rm fails
await fs.promises.rmdir(fullPath)
}
} else {
await fs.promises.unlink(fullPath)
}
}
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Error removing directory contents ${dirPath}:`, error)
throw error
}
}
}
try {
// First try the modern approach
await fs.promises.rm(this.rootDir, { recursive: true, force: true })
} catch (error: any) {
console.warn('Modern rm failed, falling back to manual cleanup:', error)
// Fallback: manually remove contents then directory
try {
await removeDirectoryContents(this.rootDir)
// Use fs.promises.rm with recursive option instead of rmdir
try {
await fs.promises.rm(this.rootDir, { recursive: true, force: true })
} catch (rmError: any) {
// Final fallback to rmdir if rm fails
await fs.promises.rmdir(this.rootDir)
}
} catch (fallbackError: any) {
if (fallbackError.code !== 'ENOENT') {
console.error('Manual cleanup also failed:', fallbackError)
throw fallbackError
}
}
}
this.isInitialized = false // Reset state
await this.init() // Re-create directories
}
public async getStorageStatus(): Promise<{
type: string
used: number
quota: number | null
details?: Record<string, any>
}> {
if (!this.isInitialized) await this.init()
const calculateSize = async (dirPath: string): Promise<number> => {
let size = 0
try {
const files = await fs.promises.readdir(dirPath, {
withFileTypes: true
})
for (const file of files) {
const fullPath = path.join(dirPath, file.name)
if (file.isDirectory()) {
size += await calculateSize(fullPath)
} else {
const stats = await fs.promises.stat(fullPath)
size += stats.size
}
}
} catch (error: any) {
if (error.code !== 'ENOENT') {
console.error(`Could not calculate size for ${dirPath}:`, error)
}
}
return size
}
const totalSize = await calculateSize(this.rootDir)
const nouns = await this.getAllNouns()
const verbs = await this.getAllVerbs()
return {
type: 'FileSystem',
used: totalSize,
quota: null, // File system quota is not easily available from Node.js
details: {
rootDir: this.rootDir,
nounCount: nouns.length,
verbCount: verbs.length
}
}
}
}

File diff suppressed because it is too large Load diff

View file

@ -1,952 +0,0 @@
import { GraphVerb, HNSWNoun, StorageAdapter } from '../coreTypes.js'
// Type aliases for compatibility
type HNSWNode = HNSWNoun
type Edge = GraphVerb
// Export R2Storage as an alias for S3CompatibleStorage
export { S3CompatibleStorage as R2Storage }
// Constants for S3 bucket prefixes
const NOUNS_PREFIX = 'nouns/'
const VERBS_PREFIX = 'verbs/'
const EDGES_PREFIX = 'verbs/' // Alias for VERBS_PREFIX for edge operations
const METADATA_PREFIX = 'metadata/'
// All nouns now use the same prefix - no separate directories per noun type
const NOUN_PREFIX = 'nouns/' // Single directory for all noun types
/**
* S3-compatible storage adapter for server environments
* Uses the AWS S3 client to interact with S3-compatible storage services
* including Amazon S3, Cloudflare R2, and Google Cloud Storage
*
* To use this adapter with Cloudflare R2, you need to provide:
* - bucketName: Your bucket name
* - accountId: Your Cloudflare account ID
* - accessKeyId: R2 access key ID
* - secretAccessKey: R2 secret access key
* - serviceType: 'r2'
*
* To use this adapter with Amazon S3, you need to provide:
* - bucketName: Your S3 bucket name
* - accessKeyId: AWS access key ID
* - secretAccessKey: AWS secret access key
* - region: AWS region (e.g., 'us-east-1')
* - serviceType: 's3'
*
* To use this adapter with Google Cloud Storage, you need to provide:
* - bucketName: Your GCS bucket name
* - accessKeyId: HMAC access key
* - secretAccessKey: HMAC secret
* - endpoint: GCS endpoint (e.g., 'https://storage.googleapis.com')
* - serviceType: 'gcs'
*
* For other S3-compatible services, provide:
* - bucketName: Your bucket name
* - accessKeyId: Access key ID
* - secretAccessKey: Secret access key
* - endpoint: Service endpoint URL
* - region: Region (if required)
* - serviceType: 'custom'
*/
export class S3CompatibleStorage implements StorageAdapter {
private bucketName: string
private accessKeyId: string
private secretAccessKey: string
private endpoint?: string
private region?: string
private accountId?: string
private serviceType: 'r2' | 's3' | 'gcs' | 'custom'
private s3Client: any // Will be initialized in init()
private isInitialized = false
// Alias methods to match StorageAdapter interface
public async saveNoun(noun: HNSWNoun): Promise<void> {
return this.saveNode(noun)
}
public async getNoun(id: string): Promise<HNSWNoun | null> {
return this.getNode(id)
}
public async getAllNouns(): Promise<HNSWNoun[]> {
return this.getAllNodes()
}
public async getNounsByNounType(nounType: string): Promise<HNSWNoun[]> {
return this.getNodesByNounType(nounType)
}
public async deleteNoun(id: string): Promise<void> {
return this.deleteNode(id)
}
public async saveVerb(verb: GraphVerb): Promise<void> {
return this.saveEdge(verb)
}
public async getVerb(id: string): Promise<GraphVerb | null> {
return this.getEdge(id)
}
public async getAllVerbs(): Promise<GraphVerb[]> {
return this.getAllEdges()
}
public async getVerbsBySource(sourceId: string): Promise<GraphVerb[]> {
return this.getEdgesBySource(sourceId)
}
public async getVerbsByTarget(targetId: string): Promise<GraphVerb[]> {
return this.getEdgesByTarget(targetId)
}
public async getVerbsByType(type: string): Promise<GraphVerb[]> {
return this.getEdgesByType(type)
}
public async deleteVerb(id: string): Promise<void> {
return this.deleteEdge(id)
}
constructor(options: {
bucketName: string
accessKeyId: string
secretAccessKey: string
serviceType: 'r2' | 's3' | 'gcs' | 'custom'
accountId?: string
region?: string
endpoint?: string
}) {
this.bucketName = options.bucketName
this.accessKeyId = options.accessKeyId
this.secretAccessKey = options.secretAccessKey
this.serviceType = options.serviceType
this.accountId = options.accountId
this.region = options.region
this.endpoint = options.endpoint
}
/**
* Initialize the storage adapter
*/
public async init(): Promise<void> {
if (this.isInitialized) {
return
}
try {
// Dynamically import the AWS SDK to avoid bundling it in browser environments
try {
const { S3Client } = await import('@aws-sdk/client-s3')
// Configure the S3 client based on the service type
const clientConfig: any = {
credentials: {
accessKeyId: this.accessKeyId,
secretAccessKey: this.secretAccessKey
}
}
switch (this.serviceType) {
case 'r2':
if (!this.accountId) {
throw new Error('accountId is required for Cloudflare R2')
}
clientConfig.region = 'auto' // R2 uses 'auto' as the region
clientConfig.endpoint = `https://${this.accountId}.r2.cloudflarestorage.com`
break
case 's3':
if (!this.region) {
throw new Error('region is required for Amazon S3')
}
clientConfig.region = this.region
// No endpoint needed for standard S3
break
case 'gcs':
if (!this.endpoint) {
// Default GCS endpoint if not provided
this.endpoint = 'https://storage.googleapis.com'
}
clientConfig.endpoint = this.endpoint
clientConfig.region = this.region || 'auto'
break
case 'custom':
if (!this.endpoint) {
throw new Error(
'endpoint is required for custom S3-compatible services'
)
}
clientConfig.endpoint = this.endpoint
if (this.region) {
clientConfig.region = this.region
}
break
default:
throw new Error(`Unsupported service type: ${this.serviceType}`)
}
// Initialize the S3 client with the configured options
this.s3Client = new S3Client(clientConfig)
this.isInitialized = true
} catch (importError) {
throw new Error(
`Failed to import AWS SDK: ${importError}. Make sure @aws-sdk/client-s3 is installed.`
)
}
} catch (error) {
console.error(`Failed to initialize ${this.serviceType} storage:`, error)
throw new Error(
`Failed to initialize ${this.serviceType} storage: ${error}`
)
}
}
/**
* Save a node to storage
*/
public async saveNode(node: HNSWNode): Promise<void> {
await this.ensureInitialized()
try {
// Convert connections Map to a serializable format
const serializableNode = {
...node,
connections: this.mapToObject(node.connections, (set) =>
Array.from(set as Set<string>)
)
}
// Get the appropriate prefix based on the node's metadata
const nodePrefix = await this.getNodePrefix(node.id)
// Import the PutObjectCommand only when needed
const { PutObjectCommand } = await import('@aws-sdk/client-s3')
// Save the node to S3-compatible storage
await this.s3Client.send(
new PutObjectCommand({
Bucket: this.bucketName,
Key: `${nodePrefix}${node.id}.json`,
Body: JSON.stringify(serializableNode, null, 2),
ContentType: 'application/json'
})
)
} catch (error) {
console.error(`Failed to save node ${node.id}:`, error)
throw new Error(`Failed to save node ${node.id}: ${error}`)
}
}
/**
* Get a node from storage
*/
public async getNode(id: string): Promise<HNSWNode | null> {
await this.ensureInitialized()
try {
// Import the GetObjectCommand only when needed
const { GetObjectCommand } = await import('@aws-sdk/client-s3')
// Try to get the node from the consolidated nouns directory
const response = await this.s3Client.send(
new GetObjectCommand({
Bucket: this.bucketName,
Key: `${NOUN_PREFIX}${id}.json`
})
)
// Convert the response body to a string
const bodyContents = await response.Body.transformToString()
const parsedNode = JSON.parse(bodyContents)
// Convert serialized connections back to Map<number, Set<string>>
const connections = new Map<number, Set<string>>()
for (const [level, nodeIds] of Object.entries(parsedNode.connections)) {
connections.set(Number(level), new Set(nodeIds as string[]))
}
return {
id: parsedNode.id,
vector: parsedNode.vector,
connections
}
} catch (error) {
// Node not found or other error
return null
}
}
/**
* Get nodes by noun type
* @param nounType The noun type to filter by
* @returns Promise that resolves to an array of nodes of the specified noun type
*/
public async getNodesByNounType(nounType: string): Promise<HNSWNode[]> {
await this.ensureInitialized()
try {
// Import the ListObjectsV2Command and GetObjectCommand only when needed
const { ListObjectsV2Command, GetObjectCommand } = await import(
'@aws-sdk/client-s3'
)
// List all objects in the consolidated nouns directory
const listResponse = await this.s3Client.send(
new ListObjectsV2Command({
Bucket: this.bucketName,
Prefix: NOUN_PREFIX
})
)
const nodes: HNSWNode[] = []
// If there are no objects, return an empty array
if (!listResponse.Contents || listResponse.Contents.length === 0) {
return nodes
}
// Get each node and filter by noun type
const nodePromises = listResponse.Contents.map(
async (object: { Key: string }) => {
try {
// Extract node ID from the key (remove prefix and .json extension)
const nodeId = object.Key.replace(NOUN_PREFIX, '').replace('.json', '')
// Get the metadata to check the noun type
const metadata = await this.getMetadata(nodeId)
// Skip if metadata doesn't exist or noun type doesn't match
if (!metadata || metadata.noun !== nounType) {
return null
}
const response = await this.s3Client.send(
new GetObjectCommand({
Bucket: this.bucketName,
Key: object.Key
})
)
const bodyContents = await response.Body.transformToString()
const parsedNode = JSON.parse(bodyContents)
// Convert serialized connections back to Map<number, Set<string>>
const connections = new Map<number, Set<string>>()
for (const [level, nodeIds] of Object.entries(
parsedNode.connections
)) {
connections.set(Number(level), new Set(nodeIds as string[]))
}
return {
id: parsedNode.id,
vector: parsedNode.vector,
connections
}
} catch (error) {
console.error(`Failed to get node from ${object.Key}:`, error)
return null
}
}
)
const nodeResults = await Promise.all(nodePromises)
return nodeResults.filter((node): node is HNSWNode => node !== null)
} catch (error) {
console.error(`Failed to get nodes for noun type ${nounType}:`, error)
throw new Error(`Failed to get nodes for noun type ${nounType}: ${error}`)
}
}
/**
* Get all nodes from storage
*/
public async getAllNodes(): Promise<HNSWNode[]> {
await this.ensureInitialized()
try {
// Import the ListObjectsV2Command and GetObjectCommand only when needed
const { ListObjectsV2Command, GetObjectCommand } = await import(
'@aws-sdk/client-s3'
)
// List all objects in the consolidated nouns directory
const listResponse = await this.s3Client.send(
new ListObjectsV2Command({
Bucket: this.bucketName,
Prefix: NOUN_PREFIX
})
)
const nodes: HNSWNode[] = []
// If there are no objects, return an empty array
if (!listResponse.Contents || listResponse.Contents.length === 0) {
return nodes
}
// Get each node
const nodePromises = listResponse.Contents.map(
async (object: { Key: string }) => {
try {
const response = await this.s3Client.send(
new GetObjectCommand({
Bucket: this.bucketName,
Key: object.Key
})
)
const bodyContents = await response.Body.transformToString()
const parsedNode = JSON.parse(bodyContents)
// Convert serialized connections back to Map<number, Set<string>>
const connections = new Map<number, Set<string>>()
for (const [level, nodeIds] of Object.entries(
parsedNode.connections
)) {
connections.set(Number(level), new Set(nodeIds as string[]))
}
return {
id: parsedNode.id,
vector: parsedNode.vector,
connections
}
} catch (error) {
console.error(`Failed to get node from ${object.Key}:`, error)
return null
}
}
)
const nodeResults = await Promise.all(nodePromises)
return nodeResults.filter((node): node is HNSWNode => node !== null)
} catch (error) {
console.error('Failed to get all nodes:', error)
throw new Error(`Failed to get all nodes: ${error}`)
}
}
/**
* Delete a node from storage
*/
public async deleteNode(id: string): Promise<void> {
await this.ensureInitialized()
try {
// Import the DeleteObjectCommand only when needed
const { DeleteObjectCommand } = await import('@aws-sdk/client-s3')
// Delete the node from the consolidated nouns directory
await this.s3Client.send(
new DeleteObjectCommand({
Bucket: this.bucketName,
Key: `${NOUN_PREFIX}${id}.json`
})
)
} catch (error) {
console.error(`Failed to delete node ${id}:`, error)
throw new Error(`Failed to delete node ${id}: ${error}`)
}
}
/**
* Save an edge to storage
*/
public async saveEdge(edge: Edge): Promise<void> {
await this.ensureInitialized()
try {
// Convert connections Map to a serializable format
const serializableEdge = {
...edge,
connections: this.mapToObject(edge.connections, (set) =>
Array.from(set as Set<string>)
)
}
// Import the PutObjectCommand only when needed
const { PutObjectCommand } = await import('@aws-sdk/client-s3')
// Save the edge to S3-compatible storage
await this.s3Client.send(
new PutObjectCommand({
Bucket: this.bucketName,
Key: `${VERBS_PREFIX}${edge.id}.json`,
Body: JSON.stringify(serializableEdge, null, 2),
ContentType: 'application/json'
})
)
} catch (error) {
console.error(`Failed to save edge ${edge.id}:`, error)
throw new Error(`Failed to save edge ${edge.id}: ${error}`)
}
}
/**
* Get an edge from storage
*/
public async getEdge(id: string): Promise<Edge | null> {
await this.ensureInitialized()
try {
// Import the GetObjectCommand only when needed
const { GetObjectCommand } = await import('@aws-sdk/client-s3')
try {
// Try to get the edge from S3-compatible storage
const response = await this.s3Client.send(
new GetObjectCommand({
Bucket: this.bucketName,
Key: `${VERBS_PREFIX}${id}.json`
})
)
// Convert the response body to a string
const bodyContents = await response.Body.transformToString()
const parsedEdge = JSON.parse(bodyContents)
// Convert serialized connections back to Map<number, Set<string>>
const connections = new Map<number, Set<string>>()
for (const [level, nodeIds] of Object.entries(parsedEdge.connections)) {
connections.set(Number(level), new Set(nodeIds as string[]))
}
return {
id: parsedEdge.id,
vector: parsedEdge.vector,
connections,
sourceId: parsedEdge.sourceId,
targetId: parsedEdge.targetId,
type: parsedEdge.type,
weight: parsedEdge.weight,
metadata: parsedEdge.metadata
}
} catch {
return null // Edge not found
}
} catch (error) {
console.error(`Failed to get edge ${id}:`, error)
return null
}
}
/**
* Get all edges from storage
*/
public async getAllEdges(): Promise<Edge[]> {
await this.ensureInitialized()
try {
// Import the ListObjectsV2Command and GetObjectCommand only when needed
const { ListObjectsV2Command, GetObjectCommand } = await import(
'@aws-sdk/client-s3'
)
// List all objects with the edges prefix
const listResponse = await this.s3Client.send(
new ListObjectsV2Command({
Bucket: this.bucketName,
Prefix: VERBS_PREFIX
})
)
const edges: Edge[] = []
// If there are no objects, return an empty array
if (!listResponse.Contents || listResponse.Contents.length === 0) {
return edges
}
// Get each edge
const edgePromises = listResponse.Contents.map(
async (object: { Key: string }) => {
try {
const response = await this.s3Client.send(
new GetObjectCommand({
Bucket: this.bucketName,
Key: object.Key
})
)
const bodyContents = await response.Body.transformToString()
const parsedEdge = JSON.parse(bodyContents)
// Convert serialized connections back to Map<number, Set<string>>
const connections = new Map<number, Set<string>>()
for (const [level, nodeIds] of Object.entries(
parsedEdge.connections
)) {
connections.set(Number(level), new Set(nodeIds as string[]))
}
return {
id: parsedEdge.id,
vector: parsedEdge.vector,
connections,
sourceId: parsedEdge.sourceId,
targetId: parsedEdge.targetId,
type: parsedEdge.type,
weight: parsedEdge.weight,
metadata: parsedEdge.metadata
}
} catch (error) {
console.error(`Failed to get edge from ${object.Key}:`, error)
return null
}
}
)
const edgeResults = await Promise.all(edgePromises)
return edgeResults.filter((edge): edge is Edge => edge !== null)
} catch (error) {
console.error('Failed to get all edges:', error)
throw new Error(`Failed to get all edges: ${error}`)
}
}
/**
* Get edges by source node ID
*/
public async getEdgesBySource(sourceId: string): Promise<Edge[]> {
await this.ensureInitialized()
try {
const allEdges = await this.getAllEdges()
return allEdges.filter((edge) => edge.sourceId === sourceId)
} catch (error) {
console.error(`Failed to get edges by source ${sourceId}:`, error)
throw new Error(`Failed to get edges by source ${sourceId}: ${error}`)
}
}
/**
* Get edges by target node ID
*/
public async getEdgesByTarget(targetId: string): Promise<Edge[]> {
await this.ensureInitialized()
try {
const allEdges = await this.getAllEdges()
return allEdges.filter((edge) => edge.targetId === targetId)
} catch (error) {
console.error(`Failed to get edges by target ${targetId}:`, error)
throw new Error(`Failed to get edges by target ${targetId}: ${error}`)
}
}
/**
* Get edges by type
*/
public async getEdgesByType(type: string): Promise<Edge[]> {
await this.ensureInitialized()
try {
const allEdges = await this.getAllEdges()
return allEdges.filter((edge) => edge.type === type)
} catch (error) {
console.error(`Failed to get edges by type ${type}:`, error)
throw new Error(`Failed to get edges by type ${type}: ${error}`)
}
}
/**
* Delete an edge from storage
*/
public async deleteEdge(id: string): Promise<void> {
await this.ensureInitialized()
try {
// Import the DeleteObjectCommand and GetObjectCommand only when needed
const { DeleteObjectCommand, GetObjectCommand } = await import(
'@aws-sdk/client-s3'
)
try {
// Check if the edge exists before deleting
await this.s3Client.send(
new GetObjectCommand({
Bucket: this.bucketName,
Key: `${EDGES_PREFIX}${id}.json`
})
)
// Delete the edge
await this.s3Client.send(
new DeleteObjectCommand({
Bucket: this.bucketName,
Key: `${EDGES_PREFIX}${id}.json`
})
)
} catch {
// Edge not found, nothing to delete
return
}
} catch (error) {
console.error(`Failed to delete edge ${id}:`, error)
throw new Error(`Failed to delete edge ${id}: ${error}`)
}
}
/**
* Save metadata to storage
*/
public async saveMetadata(id: string, metadata: any): Promise<void> {
await this.ensureInitialized()
try {
// Import the PutObjectCommand only when needed
const { PutObjectCommand } = await import('@aws-sdk/client-s3')
// Save the metadata to S3-compatible storage
await this.s3Client.send(
new PutObjectCommand({
Bucket: this.bucketName,
Key: `${METADATA_PREFIX}${id}.json`,
Body: JSON.stringify(metadata, null, 2),
ContentType: 'application/json'
})
)
} catch (error) {
console.error(`Failed to save metadata for ${id}:`, error)
throw new Error(`Failed to save metadata for ${id}: ${error}`)
}
}
/**
* Get metadata from storage
*/
public async getMetadata(id: string): Promise<any | null> {
await this.ensureInitialized()
try {
// Import the GetObjectCommand only when needed
const { GetObjectCommand } = await import('@aws-sdk/client-s3')
try {
// Try to get the metadata from S3-compatible storage
const response = await this.s3Client.send(
new GetObjectCommand({
Bucket: this.bucketName,
Key: `${METADATA_PREFIX}${id}.json`
})
)
// Convert the response body to a string
const bodyContents = await response.Body.transformToString()
return JSON.parse(bodyContents)
} catch {
return null // Metadata not found
}
} catch (error) {
console.error(`Failed to get metadata for ${id}:`, error)
return null
}
}
/**
* Clear all data from storage
*/
public async clear(): Promise<void> {
await this.ensureInitialized()
try {
// Import the ListObjectsV2Command and DeleteObjectCommand only when needed
const { ListObjectsV2Command, DeleteObjectCommand } = await import(
'@aws-sdk/client-s3'
)
// List all objects in the bucket
const listResponse = await this.s3Client.send(
new ListObjectsV2Command({
Bucket: this.bucketName
})
)
// If there are no objects, return
if (!listResponse.Contents || listResponse.Contents.length === 0) {
return
}
// Delete each object
const deletePromises = listResponse.Contents.map(
async (object: { Key: string }) => {
try {
await this.s3Client.send(
new DeleteObjectCommand({
Bucket: this.bucketName,
Key: object.Key
})
)
} catch (error) {
console.error(`Failed to delete object ${object.Key}:`, error)
}
}
)
await Promise.all(deletePromises)
} catch (error) {
console.error('Failed to clear storage:', error)
throw new Error(`Failed to clear storage: ${error}`)
}
}
/**
* Get information about storage usage and capacity
*/
public async getStorageStatus(): Promise<{
type: string
used: number
quota: number | null
details?: Record<string, any>
}> {
await this.ensureInitialized()
try {
// Import the ListObjectsV2Command only when needed
const { ListObjectsV2Command } = await import('@aws-sdk/client-s3')
// List all objects in the bucket to calculate total size
const listResponse = await this.s3Client.send(
new ListObjectsV2Command({
Bucket: this.bucketName
})
)
let totalSize = 0
let nodeCount = 0
let edgeCount = 0
let metadataCount = 0
// Calculate total size and counts
if (listResponse.Contents) {
for (const object of listResponse.Contents) {
totalSize += object.Size || 0
const key = object.Key || ''
if (key.startsWith(NOUNS_PREFIX)) {
nodeCount++
} else if (key.startsWith(VERBS_PREFIX)) {
edgeCount++
} else if (key.startsWith(METADATA_PREFIX)) {
metadataCount++
}
}
}
// Count nodes by noun type by examining metadata
const nounTypeCounts: Record<string, number> = {
person: 0,
place: 0,
thing: 0,
event: 0,
concept: 0,
content: 0,
group: 0,
list: 0,
category: 0,
default: 0
}
// List all noun objects and count by type using metadata
const nounsListResponse = await this.s3Client.send(
new ListObjectsV2Command({
Bucket: this.bucketName,
Prefix: NOUN_PREFIX
})
)
if (nounsListResponse.Contents) {
for (const object of nounsListResponse.Contents) {
try {
// Extract node ID from the key
const nodeId = object.Key?.replace(NOUN_PREFIX, '').replace('.json', '')
if (nodeId) {
// Get metadata to determine noun type
const metadata = await this.getMetadata(nodeId)
const nounType = metadata?.noun || 'default'
if (nounType in nounTypeCounts) {
nounTypeCounts[nounType]++
} else {
nounTypeCounts.default++
}
}
} catch (error) {
// If we can't get metadata, count as default
nounTypeCounts.default++
}
}
}
return {
type: this.serviceType,
used: totalSize,
quota: null, // S3-compatible services typically don't provide quota information through the API
details: {
nodeCount,
edgeCount,
metadataCount,
nounTypes: {
person: { count: nounTypeCounts.person },
place: { count: nounTypeCounts.place },
thing: { count: nounTypeCounts.thing },
event: { count: nounTypeCounts.event },
concept: { count: nounTypeCounts.concept },
content: { count: nounTypeCounts.content },
group: { count: nounTypeCounts.group },
list: { count: nounTypeCounts.list },
category: { count: nounTypeCounts.category },
default: { count: nounTypeCounts.default }
}
}
}
} catch (error) {
console.error('Failed to get storage status:', error)
return {
type: this.serviceType,
used: 0,
quota: null,
details: { error: String(error) }
}
}
}
/**
* Ensure the storage adapter is initialized
*/
private async ensureInitialized(): Promise<void> {
if (!this.isInitialized) {
await this.init()
}
}
/**
* Get the appropriate prefix for a node - now all nouns use the same prefix
*/
private async getNodePrefix(id: string): Promise<string> {
// All nouns now use the same prefix regardless of type
return NOUN_PREFIX
}
/**
* Convert a Map to a plain object for serialization
*/
private mapToObject<K extends string | number, V>(
map: Map<K, V>,
valueTransformer: (value: V) => any = (v) => v
): Record<string, any> {
const obj: Record<string, any> = {}
for (const [key, value] of map.entries()) {
obj[key.toString()] = valueTransformer(value)
}
return obj
}
}

View file

@ -34,7 +34,7 @@ export interface GraphNoun {
}
/**
* Base interface for edges (verbs) in the graph
* Base interface for verbs in the graph
* Represents relationships between nouns
*/
export interface GraphVerb {
@ -45,6 +45,7 @@ export interface GraphVerb {
verb: VerbType // Type of relationship
createdAt: Timestamp // When the verb was created
updatedAt: Timestamp // When the verb was last updated
createdBy: CreatorMetadata // Information about what created this verb
data?: Record<string, any> // Additional flexible data storage
embedding?: number[] // Vector representation of the relationship
confidence?: number // Confidence score (0-1)

View file

@ -1,3 +1,4 @@
export * from './distance.js'
export * from './embedding.js'
export * from './workerUtils.js'
export * from './statistics.js'

44
src/utils/statistics.ts Normal file
View file

@ -0,0 +1,44 @@
/**
* Utility functions for retrieving statistics from Brainy
*/
import { BrainyData } from '../brainyData.js'
/**
* Get statistics about the current state of a BrainyData instance
* This function provides access to statistics at the root level of the library
*
* @param instance A BrainyData instance to get statistics from
* @param options Additional options for retrieving statistics
* @returns Object containing counts of nouns, verbs, metadata entries, and HNSW index size
* @throws Error if the instance is not provided or if statistics retrieval fails
*/
export async function getStatistics(
instance: BrainyData,
options: {
service?: string | string[] // Filter statistics by service(s)
} = {}
): Promise<{
nounCount: number
verbCount: number
metadataCount: number
hnswIndexSize: number
serviceBreakdown?: {
[service: string]: {
nounCount: number
verbCount: number
metadataCount: number
}
}
}> {
if (!instance) {
throw new Error('BrainyData instance must be provided to getStatistics')
}
try {
return await instance.getStatistics(options)
} catch (error) {
console.error('Failed to get statistics:', error)
throw new Error(`Failed to get statistics: ${error}`)
}
}