David Snelling
54c7c39669
An untyped (JS) caller that smuggles a Brainy-reserved field (confidence, weight, subtype, visibility, service, createdBy, noun/verb, data, createdAt, updatedAt, _rev) inside a write-path metadata bag previously got a silent remap-or-drop — a class of bug where confidence-evolution writes no-oped for weeks before being caught on read-back. 8.0 closes this with no silent failures. - New BrainyConfig.reservedFieldPolicy: 'throw' | 'warn' | 'remap' (default 'throw'). 'throw' rejects the write naming every offending key + its correct write path; 'warn' remaps with a one-shot per-key warning; 'remap' is the legacy silent path. - Central enforceReservedPolicy gate wired into all four remap methods (add, update, relate, updateRelation) so live calls AND their transact()/with() mirrors honor it. Single-source reservedWritePath guidance shared by throw+warn. - 'warn' now warns for EVERY reserved key (closes the gap where only system-managed fields warned). Dead warnDropped* helpers removed. - Import pipeline migrated to route reserved values (confidence/weight/subtype) through dedicated params and strip reserved keys from extractor/customMetadata bags via the canonical split*MetadataRecord helpers — imports no longer trip the default throw. - Tests: new reservedFieldPolicy matrix (throw/warn/remap across every write path + transact); remap-correctness suite reframed as opt-in 'remap'; shared test-factory no longer emits reserved keys in custom metadata.
1688 lines
No EOL
62 KiB
TypeScript
1688 lines
No EOL
62 KiB
TypeScript
/**
|
|
* 🧠 Brainy 3.0 Type Definitions
|
|
*
|
|
* Beautiful, consistent, type-safe interfaces for the future of neural databases
|
|
*/
|
|
|
|
import { StorageAdapter, Vector } from '../coreTypes.js'
|
|
import type { EntityVisibility } from '../coreTypes.js'
|
|
import { NounType, VerbType } from './graphTypes.js'
|
|
import type {
|
|
EntityMetadataInput,
|
|
EntityMetadataPatch,
|
|
RelationMetadataInput,
|
|
RelationMetadataPatch
|
|
} from './reservedFields.js'
|
|
|
|
// ============= Core Types =============
|
|
|
|
/**
|
|
* Entity (Noun) — the fundamental data unit in Brainy
|
|
*
|
|
* **Data vs Metadata:**
|
|
* - `data`: Content used for vector embeddings. Searchable via **semantic similarity**
|
|
* (HNSW index) and hybrid text+semantic search. NOT queryable via `where` filters.
|
|
* - `metadata`: Structured fields indexed by MetadataIndex. Queryable via `where`
|
|
* filters in `find()`. Standard system fields (noun, createdAt, etc.) are stored
|
|
* alongside user metadata but extracted to top-level Entity fields on read.
|
|
*/
|
|
export interface Entity<T = any> {
|
|
/** Unique identifier (UUID v4) */
|
|
id: string
|
|
/** Embedding vector (384-dim by default). Empty array when loaded without includeVectors. */
|
|
vector: Vector
|
|
/** Entity type classification (NounType enum) */
|
|
type: NounType
|
|
/**
|
|
* Per-product sub-classification within the NounType (e.g. a `Person` entity might
|
|
* have `subtype: 'employee'` or `'customer'`; a `Document` might have `subtype: 'invoice'`).
|
|
* Flat string, no hierarchy. Top-level standard field — indexed on the fast path and
|
|
* rolled into per-NounType statistics.
|
|
*/
|
|
subtype?: string
|
|
/**
|
|
* Visibility tier (reserved top-level field). Absent === `'public'` (counted and
|
|
* returned everywhere). `'internal'` hides the entity from default `find()` / counts /
|
|
* `stats()` (opt back in with `find({ includeInternal: true })`); `'system'` is Brainy
|
|
* plumbing, hidden unless `find({ includeSystem: true })`. See {@link EntityVisibility}.
|
|
*/
|
|
visibility?: EntityVisibility
|
|
/** Opaque content — used for embeddings and semantic search. Not indexed by MetadataIndex. */
|
|
data?: any
|
|
/** User-defined structured fields — indexed and queryable via `where` filters. */
|
|
metadata?: T
|
|
/** Multi-tenancy service identifier */
|
|
service?: string
|
|
/** Creation timestamp (ms since epoch) */
|
|
createdAt: number
|
|
/** Last update timestamp (ms since epoch) */
|
|
updatedAt?: number
|
|
/** Source that created this entity (e.g., augmentation info) */
|
|
createdBy?: string
|
|
/** Type classification confidence (0-1) */
|
|
confidence?: number
|
|
/** Entity importance/salience (0-1) */
|
|
weight?: number
|
|
/**
|
|
* Monotonic revision counter, auto-bumped on every successful `update()`.
|
|
* Starts at `1` on `add()`. Pre-7.31.0 entities without `_rev` are read as `1`.
|
|
* Pass back as `update({ id, ..., ifRev: _rev })` for optimistic-concurrency CAS —
|
|
* throws `RevisionConflictError` if the persisted rev moved since read.
|
|
*/
|
|
_rev?: number
|
|
}
|
|
|
|
/**
|
|
* Relation (Verb) — a typed edge connecting two entities
|
|
*
|
|
* **Data vs Metadata (on relationships):**
|
|
* - `data`: Opaque content stored on the relationship. If provided during relate(),
|
|
* overrides the auto-computed vector (default: average of source+target vectors).
|
|
* - `metadata`: Structured queryable fields on the edge (e.g., role, startDate).
|
|
*/
|
|
export interface Relation<T = any> {
|
|
/** Unique identifier (UUID v4) */
|
|
id: string
|
|
/** Source entity ID */
|
|
from: string
|
|
/** Target entity ID */
|
|
to: string
|
|
/** Relationship type classification (VerbType enum) */
|
|
type: VerbType
|
|
/**
|
|
* Per-product sub-classification within the VerbType (e.g. a `ReportsTo`
|
|
* relationship might have `subtype: 'direct'` vs `'dotted-line'`; a `RelatedTo`
|
|
* edge might carry `'spouse'` / `'sibling'` / `'colleague'`). Flat string, no
|
|
* hierarchy. Top-level standard field — indexed on the fast path and rolled into
|
|
* per-VerbType statistics so queries like `related({ verb, subtype })` hit
|
|
* the column-store directly.
|
|
*/
|
|
subtype?: string
|
|
/**
|
|
* Visibility tier (reserved top-level field), the verb mirror of `Entity.visibility`.
|
|
* Absent === `'public'` (counted and returned everywhere). `'internal'` hides the edge
|
|
* from default `related()` / counts / `stats()` (opt back in with
|
|
* `related({ includeInternal: true })`); `'system'` is Brainy plumbing. See
|
|
* {@link EntityVisibility}.
|
|
*/
|
|
visibility?: EntityVisibility
|
|
/** Connection strength (0-1, default: 1.0) */
|
|
weight?: number
|
|
/** Opaque content for the relationship (overrides auto-computed vector if provided) */
|
|
data?: any
|
|
/** User-defined structured fields on the edge */
|
|
metadata?: T
|
|
/** Multi-tenancy service identifier */
|
|
service?: string
|
|
/** Creation timestamp (ms since epoch) */
|
|
createdAt: number
|
|
/** Last update timestamp (ms since epoch) */
|
|
updatedAt?: number
|
|
/** Relationship certainty (0-1) */
|
|
confidence?: number
|
|
/** Evidence for why this relationship was detected */
|
|
evidence?: RelationEvidence
|
|
}
|
|
|
|
/**
|
|
* Evidence for why a relationship was detected
|
|
*/
|
|
export interface RelationEvidence {
|
|
sourceText?: string // Text that indicated this relationship
|
|
position?: { // Position in source text
|
|
start: number
|
|
end: number
|
|
}
|
|
method: 'neural' | 'pattern' | 'structural' | 'explicit' // How it was detected
|
|
reasoning?: string // Human-readable explanation
|
|
}
|
|
|
|
/**
|
|
* Search result with similarity score
|
|
*
|
|
* Flattens commonly-used entity fields to top level for convenience,
|
|
* while preserving full entity in 'entity' field for backward compatibility.
|
|
*/
|
|
export interface Result<T = any> {
|
|
// Search metadata
|
|
id: string
|
|
score: number
|
|
|
|
// Convenience: Common entity fields flattened to top level
|
|
type?: NounType // Entity type (from entity.type)
|
|
subtype?: string // Per-product sub-classification (from entity.subtype)
|
|
visibility?: EntityVisibility // Visibility tier (from entity.visibility); absent === 'public'
|
|
metadata?: T // Entity metadata (from entity.metadata)
|
|
data?: any // Entity data (from entity.data)
|
|
confidence?: number // Type classification confidence (from entity.confidence)
|
|
weight?: number // Entity importance (from entity.weight)
|
|
_rev?: number // Monotonic revision counter (from entity._rev) — pass back as update({ ifRev }) for CAS
|
|
|
|
// Full entity (preserved for backward compatibility)
|
|
entity: Entity<T>
|
|
|
|
// Score transparency
|
|
explanation?: ScoreExplanation
|
|
|
|
// Match visibility - shows what matched in hybrid search
|
|
textMatches?: string[] // Query words found in entity (e.g., ["david", "warrior"])
|
|
textScore?: number // Normalized text match score (0-1)
|
|
semanticScore?: number // Semantic similarity score (0-1)
|
|
matchSource?: 'text' | 'semantic' | 'both' // Where this result came from
|
|
rrfScore?: number // Raw RRF fusion score (for advanced ranking analysis)
|
|
|
|
// Aggregation: present only on rows returned by find({ aggregate }). These mirror the
|
|
// documented AggregateResult shape so callers can read group/metric data at the top level
|
|
// instead of digging into `metadata`. (The same values are also flattened into `metadata`
|
|
// for backward compatibility.)
|
|
groupKey?: Record<string, string | number> // Group-by key values for this aggregate row
|
|
metrics?: Record<string, number> // Computed metric values (sum/avg/min/max/count)
|
|
count?: number // Total entity count in this group
|
|
}
|
|
|
|
/**
|
|
* Score explanation for transparency
|
|
*/
|
|
export interface ScoreExplanation {
|
|
vectorScore?: number
|
|
metadataScore?: number
|
|
graphScore?: number
|
|
boosts?: Record<string, number>
|
|
penalties?: Record<string, number>
|
|
}
|
|
|
|
// ============= Operation Parameters =============
|
|
|
|
/**
|
|
* Parameters for adding entities
|
|
*
|
|
* **Data vs Metadata:**
|
|
* - `data` is embedded into a vector and searchable via semantic similarity (HNSW).
|
|
* It is NOT indexed by MetadataIndex and NOT queryable via `where` filters.
|
|
* - `metadata` is indexed by MetadataIndex and queryable via `where` filters in `find()`.
|
|
*/
|
|
/**
|
|
* Declaration-merging registry for per-`(type, subtype)` metadata shapes
|
|
* (Brainy 8.0, per BRAINY-8.0-SUBTYPE-CONTRACT § C-3).
|
|
*
|
|
* Consumers extend this interface in their own code to tell TypeScript what
|
|
* shape `metadata` takes for a given `(NounType, subtype)` pair. Brainy
|
|
* doesn't ship any entries — every entry is a consumer concern.
|
|
*
|
|
* @example
|
|
* ```ts
|
|
* declare module '@soulcraft/brainy' {
|
|
* interface SubtypeRegistry {
|
|
* // For NounType.Person, subtype 'employee':
|
|
* 'person:employee': { employeeId: string; department: string }
|
|
* // For NounType.Document, subtype 'invoice':
|
|
* 'document:invoice': { invoiceNumber: string; amount: number }
|
|
* }
|
|
* }
|
|
* ```
|
|
*
|
|
* Keys are `${NounType}:${subtype}`. Values describe the metadata shape.
|
|
* Brainy reads this registry (when present) to give typed `metadata`
|
|
* autocomplete + type-checking on `add()` / `update()` / `find()`. Consumers
|
|
* with no registry entries see no type-level change — the metadata bag
|
|
* stays `T = any`.
|
|
*/
|
|
export interface SubtypeRegistry {
|
|
// Intentionally empty. Consumers extend via declaration merging.
|
|
}
|
|
|
|
/**
|
|
* A single back-fill rule for `brain.fillSubtypes()`.
|
|
*
|
|
* - A **literal string** assigns that subtype to every matching entry that
|
|
* lacks one (`'general'` — a blanket default).
|
|
* - A **function** receives the full entry and returns the subtype to assign,
|
|
* or `undefined` to leave the entry untouched (it is counted as `skipped`
|
|
* so a later run with a stricter rule can pick it up). Functions can derive
|
|
* the subtype from existing fields, e.g. `(e) => e.metadata?.kind`.
|
|
*
|
|
* @typeParam E - The entry shape the rule sees: `Entity<T>` for NounType keys,
|
|
* `Relation<T>` for VerbType keys.
|
|
*/
|
|
export type FillSubtypeRule<E> = string | ((entry: E) => string | undefined)
|
|
|
|
/**
|
|
* Rule map for `brain.fillSubtypes()` — the 8.0 subtype migration helper.
|
|
*
|
|
* Keys are `NounType` values (entity rules) and/or `VerbType` values
|
|
* (relationship rules); the two vocabularies don't overlap, so a single map
|
|
* covers both sides. Each value is a {@link FillSubtypeRule}: a literal
|
|
* subtype string or a function deriving one from the entry.
|
|
*
|
|
* @example
|
|
* ```ts
|
|
* await brain.fillSubtypes({
|
|
* [NounType.Person]: (e) => e.metadata?.kind ?? 'unspecified',
|
|
* [NounType.Thing]: 'general', // literal default
|
|
* [VerbType.RelatedTo]: 'unspecified' // relationship rule
|
|
* })
|
|
* ```
|
|
*/
|
|
export type FillSubtypeRules<T = any> = {
|
|
[K in NounType]?: FillSubtypeRule<Entity<T>>
|
|
} & {
|
|
[K in VerbType]?: FillSubtypeRule<Relation<T>>
|
|
}
|
|
|
|
/**
|
|
* Summary returned by `brain.fillSubtypes()`.
|
|
*
|
|
* After a run, `skipped` is exactly the remaining migration debt — re-running
|
|
* `brain.audit()` reports the same entries. Entries that already carry a
|
|
* subtype count toward `scanned` only (they are not debt, so they are neither
|
|
* `filled` nor `skipped`).
|
|
*/
|
|
export interface FillSubtypesResult {
|
|
/** Total entries examined (entities + relationships). */
|
|
scanned: number
|
|
/** Entries that received a subtype during this run. */
|
|
filled: number
|
|
/**
|
|
* Entries still missing a subtype after the run — either their type has no
|
|
* rule in the map, or their rule function returned `undefined`/empty.
|
|
*/
|
|
skipped: number
|
|
/** Per-entry write failures (`fillSubtypes` continues past individual errors). */
|
|
errors: Array<{ id: string; error: string }>
|
|
/** Fill counts grouped by NounType/VerbType key (only types with fills appear). */
|
|
byType: Record<string, number>
|
|
}
|
|
|
|
export interface AddParams<T = any> {
|
|
/** Content to embed and store. Strings are auto-embedded; objects are JSON-stringified for embedding. */
|
|
data: any | Vector
|
|
/** Entity type classification (required) */
|
|
type: NounType
|
|
/**
|
|
* Per-product sub-classification within the NounType (e.g. a `Person` entity might have
|
|
* `subtype: 'employee'` or `'customer'`; a `Document` might have `subtype: 'invoice'`).
|
|
* Flat string, no hierarchy — consumers choose the vocabulary. Indexed and rolled up into
|
|
* per-NounType statistics for fast `find({ type, subtype })` filtering and `groupBy:['subtype']`
|
|
* aggregation on the standard-field fast path.
|
|
*/
|
|
subtype?: string
|
|
/**
|
|
* Visibility tier (reserved top-level field). Omit (or pass `'public'`) for normal
|
|
* data that is counted and returned everywhere. Pass `'internal'` for app-internal
|
|
* data that should be hidden from default `find()` / counts / `stats()` but stay
|
|
* retrievable via `find({ includeInternal: true })`. The `'system'` tier is reserved
|
|
* for Brainy's own plumbing and is intentionally not accepted here. Stored only when
|
|
* not `'public'`.
|
|
*/
|
|
visibility?: 'public' | 'internal'
|
|
/**
|
|
* Structured queryable fields — indexed by MetadataIndex, used in `where` filters.
|
|
*
|
|
* Reserved entity fields (`RESERVED_ENTITY_FIELDS` — `noun`, `subtype`, `visibility`,
|
|
* `createdAt`, `updatedAt`, `confidence`, `weight`, `service`, `data`, `createdBy`,
|
|
* `_rev`) may NOT appear here — they have dedicated top-level params and the type makes
|
|
* a literal reserved key a compile error. Untyped (JavaScript) callers that pass one
|
|
* anyway are normalized at write time: user-settable fields remap to their top-level
|
|
* param (top-level wins when both are supplied), system-managed fields are dropped with
|
|
* a one-shot warning.
|
|
*/
|
|
metadata?: EntityMetadataInput<T>
|
|
/** Custom entity ID (auto-generated UUID v4 if not provided) */
|
|
id?: string
|
|
/** Pre-computed embedding vector (skips auto-embedding when provided) */
|
|
vector?: Vector
|
|
/** Multi-tenancy service identifier */
|
|
service?: string
|
|
/** Type classification confidence (0-1) */
|
|
confidence?: number
|
|
/** Entity importance/salience (0-1) */
|
|
weight?: number
|
|
/** Track which augmentation created this entity */
|
|
createdBy?: { augmentation: string; version: string }
|
|
/**
|
|
* Conditional insert. When `true` AND a custom `id` is supplied AND an entity with
|
|
* that `id` already exists, `add()` returns the existing `id` without writing — no
|
|
* throw, no overwrite. Used for idempotent bootstrap and create-or-noop patterns.
|
|
* Ignored when `id` is omitted (a freshly generated UUID can never collide).
|
|
*/
|
|
ifAbsent?: boolean
|
|
}
|
|
|
|
/**
|
|
* Parameters for updating entities
|
|
*/
|
|
export interface UpdateParams<T = any> {
|
|
id: string // Entity to update
|
|
data?: any // New content to re-embed
|
|
type?: NounType // Change type
|
|
subtype?: string // Change subtype (set to '' or null-equivalent via dedicated unset is future work)
|
|
/**
|
|
* Change the visibility tier. Omit to preserve the existing value. Toggling between
|
|
* `'public'` and `'internal'` moves the entity in/out of the default-visible counts and
|
|
* `find()` results. The `'system'` tier is Brainy-internal (e.g. re-asserted on the VFS
|
|
* root during repair) — consumer code should only ever use `'public'` / `'internal'`.
|
|
*/
|
|
visibility?: EntityVisibility
|
|
/**
|
|
* Metadata fields to merge (or replace when `merge: false`). Reserved entity
|
|
* fields (`RESERVED_ENTITY_FIELDS`) may NOT appear here — `confidence` /
|
|
* `weight` / `subtype` / `visibility` have dedicated params on this call, and the rest
|
|
* are system-managed. A literal reserved key is a compile error; untyped callers
|
|
* are normalized at write time (remap user-settable, drop system-managed
|
|
* with a one-shot warning).
|
|
*/
|
|
metadata?: EntityMetadataPatch<T>
|
|
merge?: boolean // Merge or replace metadata (default: true)
|
|
vector?: Vector // New pre-computed vector
|
|
confidence?: number // Update type classification confidence
|
|
weight?: number // Update entity importance/salience
|
|
/**
|
|
* Optimistic concurrency check. When provided, the update fails with
|
|
* `RevisionConflictError` if the persisted entity's `_rev` does not equal `ifRev`.
|
|
* `_rev` is auto-bumped on every successful update — read it from the entity returned
|
|
* by `get()` / `find()` / `search()`. Omit to skip the check (unconditional update).
|
|
*/
|
|
ifRev?: number
|
|
}
|
|
|
|
/**
|
|
* Parameters for creating relationships
|
|
*
|
|
* **Data vs Metadata (on relationships):**
|
|
* - `data`: Opaque content for the edge. If provided, overrides the auto-computed
|
|
* vector (default: average of source+target entity vectors).
|
|
* - `metadata`: Structured queryable fields on the edge.
|
|
*/
|
|
export interface RelateParams<T = any> {
|
|
/** Source entity ID (required — must exist) */
|
|
from: string
|
|
/** Target entity ID (required — must exist) */
|
|
to: string
|
|
/** Relationship type classification (required) */
|
|
type: VerbType
|
|
/**
|
|
* Per-product sub-classification within the VerbType (e.g. a `ReportsTo`
|
|
* relationship might have `subtype: 'direct'` or `'dotted-line'`; a `RelatedTo`
|
|
* edge might carry `'spouse'` or `'colleague'`). Flat string, no hierarchy.
|
|
* Indexed and rolled up into per-VerbType statistics for fast filtering
|
|
* (`related({ verb, subtype })`) and aggregation (`groupBy:['subtype']`).
|
|
*/
|
|
subtype?: string
|
|
/**
|
|
* Visibility tier (reserved top-level field), the verb mirror of `AddParams.visibility`.
|
|
* Omit (or pass `'public'`) for normal edges that are counted and returned everywhere.
|
|
* Pass `'internal'` for app-internal edges hidden from default `related()` / counts /
|
|
* `stats()` but retrievable via `related({ includeInternal: true })`. The `'system'`
|
|
* tier is reserved for Brainy's own plumbing and is not accepted here. Stored only when
|
|
* not `'public'`.
|
|
*/
|
|
visibility?: 'public' | 'internal'
|
|
/** Connection strength (0-1, default: 1.0) */
|
|
weight?: number
|
|
/** Content for the relationship (optional — overrides auto-computed vector) */
|
|
data?: any
|
|
/**
|
|
* Structured queryable fields on the edge. Reserved relationship fields
|
|
* (`RESERVED_RELATION_FIELDS` — `verb`, `subtype`, `visibility`, `createdAt`,
|
|
* `updatedAt`, `confidence`, `weight`, `service`, `data`, `createdBy`, `_rev`) may NOT
|
|
* appear here — they have dedicated params. A literal reserved key is a
|
|
* compile error; untyped callers are normalized at write time.
|
|
*/
|
|
metadata?: RelationMetadataInput<T>
|
|
/** Create reverse edge too (default: false) */
|
|
bidirectional?: boolean
|
|
/** Multi-tenancy service identifier */
|
|
service?: string
|
|
/** Relationship certainty (0-1) */
|
|
confidence?: number
|
|
/** Evidence for why this relationship exists */
|
|
evidence?: RelationEvidence
|
|
}
|
|
|
|
/**
|
|
* Parameters for updating relationships
|
|
*/
|
|
export interface UpdateRelationParams<T = any> {
|
|
id: string // Relation to update
|
|
type?: VerbType // Change verb type
|
|
subtype?: string // Change sub-classification (omit to preserve existing)
|
|
/**
|
|
* Change the visibility tier. Omit to preserve the existing value. The verb mirror of
|
|
* `UpdateParams.visibility`; `'system'` is Brainy-internal — consumer code should only
|
|
* use `'public'` / `'internal'`.
|
|
*/
|
|
visibility?: EntityVisibility
|
|
weight?: number // New weight
|
|
confidence?: number // New confidence (0-1)
|
|
data?: any // New content
|
|
/**
|
|
* Metadata fields to merge (or replace when `merge: false`). Reserved
|
|
* relationship fields (`RESERVED_RELATION_FIELDS`) may NOT appear here —
|
|
* a literal reserved key is a compile error; untyped callers are
|
|
* normalized at write time.
|
|
*/
|
|
metadata?: RelationMetadataPatch<T>
|
|
merge?: boolean // Merge or replace metadata
|
|
}
|
|
|
|
// ============= Query Parameters =============
|
|
|
|
/**
|
|
* Unified find parameters — Triple Intelligence search
|
|
*
|
|
* Combines three search dimensions in one query:
|
|
* - **Vector:** `query` or `vector` for semantic/hybrid similarity search (searches `data`)
|
|
* - **Metadata:** `where` for structured field filters (queries `metadata` via MetadataIndex)
|
|
* - **Graph:** `connected` for relationship traversal (via GraphAdjacencyIndex)
|
|
*
|
|
* See also: [Query Operators](../../docs/QUERY_OPERATORS.md) for all `where` operators.
|
|
*/
|
|
export interface FindParams<T = any> {
|
|
// Vector Intelligence
|
|
/** Natural language or semantic search query (embedded and matched via HNSW + text index) */
|
|
query?: string
|
|
/** Direct vector search (pre-computed embedding) */
|
|
vector?: Vector
|
|
|
|
// Metadata Intelligence
|
|
/** Filter by entity type(s). Alias for `where.noun`. */
|
|
type?: NounType | NounType[]
|
|
/**
|
|
* Filter by per-product subtype (top-level standard field — uses the fast path, not the
|
|
* metadata fallback). Pass a single string for equality, or an array for set membership
|
|
* (e.g. `subtype: ['employee', 'contractor']`). For operator-form predicates (e.g.
|
|
* `{ exists: true }`, `{ missing: true }`) use `where: { subtype: { …operators… } }`.
|
|
*/
|
|
subtype?: string | string[]
|
|
/** Metadata filters using BFO operators (e.g., `{ year: { greaterThan: 2020 } }`) */
|
|
where?: Partial<T>
|
|
|
|
// Visibility
|
|
/**
|
|
* Also return `visibility: 'internal'` entities. By default `find()` returns ONLY
|
|
* public entities (those with no `visibility` field, or `visibility: 'public'`);
|
|
* app-internal entities are hidden. Set `true` to include them. Applied as a hard
|
|
* candidate filter, so `limit` / `offset` stay correct. Does NOT include `'system'`
|
|
* entities — use `includeSystem` for those.
|
|
*/
|
|
includeInternal?: boolean
|
|
/**
|
|
* Also return `visibility: 'system'` entities (Brainy's own plumbing, e.g. the VFS
|
|
* root). Hidden by default and even when `includeInternal` is set. Set `true` only
|
|
* when you specifically need to see system entities. Applied as a hard candidate
|
|
* filter, so `limit` / `offset` stay correct.
|
|
*/
|
|
includeSystem?: boolean
|
|
|
|
// Graph Intelligence
|
|
connected?: GraphConstraints
|
|
|
|
// Proximity search
|
|
near?: {
|
|
id: string // Find near this entity
|
|
threshold?: number // Min similarity (0-1)
|
|
}
|
|
|
|
// Control options
|
|
limit?: number // Max results (default: 10)
|
|
offset?: number // Skip N results
|
|
cursor?: string // Cursor-based pagination
|
|
|
|
// Sorting
|
|
orderBy?: string // Field to sort by (e.g., 'createdAt', 'title', 'metadata.priority')
|
|
order?: 'asc' | 'desc' // Sort direction: 'asc' (default) or 'desc'
|
|
|
|
// Advanced options
|
|
mode?: SearchMode // Search strategy
|
|
explain?: boolean // Return scoring explanation
|
|
includeRelations?: boolean // Include entity relationships
|
|
excludeVFS?: boolean // Exclude VFS entities from results (default: false - VFS included)
|
|
service?: string // Multi-tenancy filter
|
|
|
|
// Hybrid search options
|
|
searchMode?: 'auto' | 'text' | 'semantic' | 'vector' | 'hybrid' // Search strategy: auto (default), text-only, semantic/vector-only, or explicit hybrid
|
|
hybridAlpha?: number // Weight between text (0.0) and semantic (1.0) search, default: auto-detected by query length
|
|
|
|
// Triple Intelligence Fusion
|
|
fusion?: {
|
|
strategy?: 'adaptive' | 'weighted' | 'progressive'
|
|
weights?: {
|
|
vector?: number
|
|
graph?: number
|
|
field?: number
|
|
}
|
|
}
|
|
|
|
// Performance options
|
|
writeOnly?: boolean // Skip validation for high-speed ingestion
|
|
|
|
// Aggregation
|
|
/** Query a named aggregate definition. String shorthand or full query params. */
|
|
aggregate?: string | AggregateQueryParams
|
|
}
|
|
|
|
/**
|
|
* Graph constraints for search
|
|
*/
|
|
export interface GraphConstraints {
|
|
to?: string // Connected to this entity
|
|
from?: string // Connected from this entity
|
|
via?: VerbType | VerbType[] // Via these relationship types
|
|
type?: VerbType | VerbType[] // Alias for via
|
|
/**
|
|
* Filter traversal edges by VerbType subtype. Single string for equality,
|
|
* array for set membership. Composes with `via` — `{ via: 'manages', subtype:
|
|
* 'direct' }` traverses only direct-management edges, not dotted-line. Edges
|
|
* without a subtype value are excluded when this filter is set.
|
|
*/
|
|
subtype?: string | string[]
|
|
depth?: number // Max traversal depth (default: 1)
|
|
direction?: 'in' | 'out' | 'both' // Direction of traversal (default: 'both')
|
|
bidirectional?: boolean // Consider both directions
|
|
}
|
|
|
|
/**
|
|
* Search modes
|
|
*/
|
|
export type SearchMode =
|
|
| 'auto' // Automatically choose best mode (default - enables hybrid text+semantic)
|
|
| 'vector' // Pure vector search (semantic only)
|
|
| 'semantic' // Alias for vector search
|
|
| 'text' // Pure text/keyword search
|
|
| 'metadata' // Pure metadata filtering
|
|
| 'graph' // Pure graph traversal
|
|
| 'hybrid' // Combine all intelligences (explicit hybrid mode)
|
|
|
|
/**
|
|
* Parameters for similarity search
|
|
*/
|
|
export interface SimilarParams<T = any> {
|
|
to: string | Entity<T> | Vector // Find similar to this
|
|
limit?: number // Max results (default: 10)
|
|
threshold?: number // Min similarity score
|
|
type?: NounType | NounType[] // Restrict to types
|
|
where?: Partial<T> // Additional filters
|
|
service?: string // Multi-tenancy
|
|
excludeVFS?: boolean // Exclude VFS entities (default: false - VFS included)
|
|
}
|
|
|
|
/**
|
|
* Parameters for `brain.related()` / `db.related()`
|
|
*
|
|
* All parameters are optional. When called without parameters, returns all relationships
|
|
* with pagination (default limit: 100).
|
|
*
|
|
* @example
|
|
* ```typescript
|
|
* // Get all relationships (default limit: 100)
|
|
* const all = await brain.related()
|
|
*
|
|
* // Get relationships from a specific entity (string shorthand)
|
|
* const fromEntity = await brain.related(entityId)
|
|
*
|
|
* // Equivalent to:
|
|
* const fromEntity2 = await brain.related({ from: entityId })
|
|
*
|
|
* // Get relationships to a specific entity
|
|
* const toEntity = await brain.related({ to: entityId })
|
|
*
|
|
* // Filter by relationship type
|
|
* const friends = await brain.related({ type: VerbType.FriendOf })
|
|
*
|
|
* // Pagination
|
|
* const page2 = await brain.related({ offset: 100, limit: 50 })
|
|
*
|
|
* // Combined filters
|
|
* const filtered = await brain.related({
|
|
* from: entityId,
|
|
* type: VerbType.WorksWith,
|
|
* limit: 20
|
|
* })
|
|
* ```
|
|
*
|
|
* Fixed bug where calling without parameters returned empty array
|
|
* Added string ID shorthand syntax
|
|
*/
|
|
export interface RelatedParams {
|
|
/**
|
|
* Filter by source entity ID
|
|
*
|
|
* Returns all relationships originating from this entity.
|
|
*/
|
|
from?: string
|
|
|
|
/**
|
|
* Filter by target entity ID
|
|
*
|
|
* Returns all relationships pointing to this entity.
|
|
*/
|
|
to?: string
|
|
|
|
/**
|
|
* Filter by relationship type(s)
|
|
*
|
|
* Can be a single VerbType or array of VerbTypes.
|
|
*/
|
|
type?: VerbType | VerbType[]
|
|
|
|
/**
|
|
* Filter by VerbType subtype
|
|
*
|
|
* Top-level standard field — uses the fast path, not the metadata fallback.
|
|
* Pass a single string for equality (e.g. `subtype: 'direct'`), or an array
|
|
* for set membership (e.g. `subtype: ['direct', 'dotted-line']`).
|
|
*/
|
|
subtype?: string | string[]
|
|
|
|
/**
|
|
* Also return `visibility: 'internal'` relationships.
|
|
*
|
|
* By default `related()` returns ONLY public relationships (those with no
|
|
* `visibility` field, or `visibility: 'public'`); app-internal edges are hidden.
|
|
* Set `true` to include them. Applied as a hard candidate filter so `limit` /
|
|
* `offset` stay correct. Does NOT include `'system'` edges — use `includeSystem`.
|
|
*/
|
|
includeInternal?: boolean
|
|
|
|
/**
|
|
* Also return `visibility: 'system'` relationships (Brainy's own plumbing).
|
|
*
|
|
* Hidden by default and even when `includeInternal` is set. Set `true` only when
|
|
* you specifically need system edges. Applied as a hard candidate filter so
|
|
* `limit` / `offset` stay correct.
|
|
*/
|
|
includeSystem?: boolean
|
|
|
|
/**
|
|
* Maximum number of results to return
|
|
*
|
|
* @default 100
|
|
*/
|
|
limit?: number
|
|
|
|
/**
|
|
* Number of results to skip (offset-based pagination)
|
|
*
|
|
* @default 0
|
|
*/
|
|
offset?: number
|
|
|
|
/**
|
|
* Cursor for cursor-based pagination
|
|
*
|
|
* More efficient than offset for large result sets.
|
|
*/
|
|
cursor?: string
|
|
|
|
/**
|
|
* Filter by service (multi-tenancy)
|
|
*
|
|
* Only return relationships belonging to this service.
|
|
*/
|
|
service?: string
|
|
}
|
|
|
|
// ============= Batch Operations =============
|
|
|
|
/**
|
|
* Batch add parameters
|
|
*/
|
|
export interface AddManyParams<T = any> {
|
|
items: AddParams<T>[] // Items to add
|
|
parallel?: boolean // Process in parallel (default: true)
|
|
chunkSize?: number // Batch size (default: 100)
|
|
onProgress?: (done: number, total: number) => void
|
|
continueOnError?: boolean // Continue if some fail
|
|
/**
|
|
* Conditional insert applied to every item. Equivalent to setting `ifAbsent: true`
|
|
* on each item individually. Item-level `ifAbsent` overrides the batch flag.
|
|
*/
|
|
ifAbsent?: boolean
|
|
}
|
|
|
|
/**
|
|
* Batch update parameters
|
|
*/
|
|
export interface UpdateManyParams<T = any> {
|
|
items: UpdateParams<T>[] // Items to update
|
|
parallel?: boolean
|
|
chunkSize?: number
|
|
onProgress?: (done: number, total: number) => void
|
|
continueOnError?: boolean
|
|
}
|
|
|
|
/**
|
|
* Batch remove parameters
|
|
*/
|
|
export interface RemoveManyParams {
|
|
ids?: string[] // Specific IDs to remove
|
|
type?: NounType // Remove all of type
|
|
where?: any // Remove by metadata
|
|
limit?: number // Max to remove (safety)
|
|
onProgress?: (done: number, total: number) => void
|
|
continueOnError?: boolean // Continue processing if a removal fails
|
|
}
|
|
|
|
/**
|
|
* Batch relate parameters
|
|
*/
|
|
export interface RelateManyParams<T = any> {
|
|
items: RelateParams<T>[] // Relations to create
|
|
parallel?: boolean
|
|
chunkSize?: number
|
|
onProgress?: (done: number, total: number) => void
|
|
continueOnError?: boolean
|
|
}
|
|
|
|
/**
|
|
* Batch result
|
|
*/
|
|
export interface BatchResult<T = any> {
|
|
successful: T[] // Successfully processed items
|
|
failed: Array<{ // Failed items with errors
|
|
item: any
|
|
error: string
|
|
}>
|
|
total: number // Total attempted
|
|
duration: number // Time taken in ms
|
|
}
|
|
|
|
// ============= Import Progress =============
|
|
|
|
/**
|
|
* Import stage enumeration
|
|
*/
|
|
export type ImportStage =
|
|
| 'detecting' // Detecting file format
|
|
| 'reading' // Reading file from disk/network
|
|
| 'parsing' // Parsing file structure (CSV rows, PDF pages, Excel sheets)
|
|
| 'extracting' // Extracting entities using AI
|
|
| 'indexing' // Creating graph nodes and relationships
|
|
| 'completing' // Final cleanup and stats
|
|
|
|
/**
|
|
* Overall import status
|
|
*/
|
|
export type ImportStatus =
|
|
| 'starting' // Initializing import
|
|
| 'processing' // Actively importing
|
|
| 'completing' // Finalizing
|
|
| 'done' // Complete
|
|
|
|
/**
|
|
* Comprehensive import progress information
|
|
*
|
|
* Provides multi-dimensional progress tracking:
|
|
* - Bytes processed (always deterministic)
|
|
* - Entities extracted and indexed
|
|
* - Stage-specific progress
|
|
* - Time estimates
|
|
* - Performance metrics
|
|
*
|
|
*/
|
|
export interface ImportProgress {
|
|
// Overall Progress
|
|
overall_progress: number // 0-100 weighted estimate across all stages
|
|
overall_status: ImportStatus // High-level status
|
|
|
|
// Current Stage
|
|
stage: ImportStage // What's happening now
|
|
stage_progress: number // 0-100 within current stage (0 if unknown)
|
|
stage_message: string // Human-readable: "Extracting entities from PDF..."
|
|
|
|
// Bytes (Always Available - most deterministic metric)
|
|
bytes_processed: number // Bytes read/processed so far
|
|
total_bytes: number // Total file size (0 if streaming/unknown)
|
|
bytes_percentage: number // Convenience: bytes_processed / total_bytes * 100
|
|
bytes_per_second?: number // Processing rate (relevant during parsing)
|
|
|
|
// Entities (Available when extraction starts)
|
|
entities_extracted: number // Entities found during AI extraction
|
|
entities_indexed: number // Entities added to Brainy graph
|
|
entities_per_second?: number // Entities/sec (relevant during extraction/indexing)
|
|
estimated_total_entities?: number // Estimated final count
|
|
estimation_confidence?: number // 0-1 confidence in estimation
|
|
|
|
// Timing
|
|
elapsed_ms: number // Time since import started
|
|
estimated_remaining_ms?: number // Estimated time remaining
|
|
estimated_total_ms?: number // Estimated total time
|
|
|
|
// Context (helps users understand what's happening)
|
|
current_item?: string // "Processing page 5 of 23"
|
|
current_file?: string // "Sheet: Q2 Sales Data"
|
|
file_number?: number // 3 (when importing multiple files)
|
|
total_files?: number // 10
|
|
|
|
// Performance Metrics (for debugging/optimization)
|
|
metrics?: {
|
|
parsing_rate_mbps?: number // MB/s during parsing
|
|
extraction_rate_entities_per_sec?: number // Entities/s during extraction
|
|
indexing_rate_entities_per_sec?: number // Entities/s during indexing
|
|
memory_usage_mb?: number // Current memory usage
|
|
peak_memory_mb?: number // Peak memory usage
|
|
}
|
|
|
|
// Backwards Compatibility (for legacy code)
|
|
current: number // Alias for entities_indexed
|
|
total: number // Alias for estimated_total_entities or 0
|
|
}
|
|
|
|
/**
|
|
* Import progress callback - backwards compatible
|
|
*
|
|
* Supports both legacy (current, total) and new (ImportProgress object) signatures
|
|
*/
|
|
export type ImportProgressCallback =
|
|
| ((progress: ImportProgress) => void)
|
|
| ((current: number, total: number) => void)
|
|
|
|
/**
|
|
* Stage weight configuration for overall progress calculation
|
|
*
|
|
* These weights reflect the typical time distribution across stages.
|
|
* Extraction is typically the slowest stage (60% of time).
|
|
*/
|
|
export interface StageWeights {
|
|
detecting: number // Default: 0.01 (1%)
|
|
reading: number // Default: 0.05 (5%)
|
|
parsing: number // Default: 0.10 (10%)
|
|
extracting: number // Default: 0.60 (60% - slowest!)
|
|
indexing: number // Default: 0.20 (20%)
|
|
completing: number // Default: 0.04 (4%)
|
|
}
|
|
|
|
/**
|
|
* Import result statistics
|
|
*/
|
|
export interface ImportStats {
|
|
graphNodesCreated: number // Entities added to graph
|
|
graphEdgesCreated: number // Relationships created
|
|
vfsFilesCreated: number // VFS files created
|
|
duration: number // Total time in ms
|
|
bytesProcessed: number // Total bytes read
|
|
averageRate: number // Average entities/sec
|
|
peakMemoryMB?: number // Peak memory usage
|
|
}
|
|
|
|
/**
|
|
* Import operation result
|
|
*/
|
|
export interface ImportResult {
|
|
success: boolean
|
|
stats: ImportStats
|
|
errors?: Array<{
|
|
stage: ImportStage
|
|
message: string
|
|
error?: any
|
|
}>
|
|
}
|
|
|
|
// ============= Advanced Operations =============
|
|
|
|
/**
|
|
* Options for brain.get() entity retrieval
|
|
*
|
|
* **Performance Optimization**:
|
|
* By default, brain.get() loads ONLY metadata (not vectors), resulting in:
|
|
* - **76-81% faster** reads (10ms vs 43ms for metadata-only)
|
|
* - **95% less bandwidth** (300 bytes vs 6KB per entity)
|
|
* - **87% less memory** (optimal for VFS and large-scale operations)
|
|
*
|
|
* **When to use includeVectors**:
|
|
* - Computing similarity on a specific entity (not search): `brain.similar({ to: entity.vector })`
|
|
* - Manual vector operations: `cosineSimilarity(entity.vector, otherVector)`
|
|
* - Inspecting embeddings for debugging
|
|
*
|
|
* **When NOT to use includeVectors** (metadata-only is sufficient):
|
|
* - VFS operations (readFile, stat, readdir) - 100% of cases
|
|
* - Existence checks: `if (await brain.get(id))`
|
|
* - Metadata inspection: `entity.metadata`, `entity.data`, `entity.type`
|
|
* - Relationship traversal: `brain.related({ from: id })`
|
|
* - Search operations: `brain.find()` generates embeddings automatically
|
|
*
|
|
* @example
|
|
* ```typescript
|
|
* // ✅ FAST (default): Metadata-only - 10ms, 300 bytes
|
|
* const entity = await brain.get(id)
|
|
* console.log(entity.data, entity.metadata) // ✅ Available
|
|
* console.log(entity.vector) // Empty Float32Array (stub)
|
|
*
|
|
* // ✅ FULL: Load vectors when needed - 43ms, 6KB
|
|
* const fullEntity = await brain.get(id, { includeVectors: true })
|
|
* const similarity = cosineSimilarity(fullEntity.vector, otherVector)
|
|
*
|
|
* // ✅ VFS automatically uses fast path (no change needed)
|
|
* await vfs.readFile('/file.txt') // 53ms → 10ms (81% faster)
|
|
* ```
|
|
*
|
|
*/
|
|
export interface GetOptions {
|
|
/**
|
|
* Include 384-dimensional vector embeddings in the response
|
|
*
|
|
* **Default: false** (metadata-only for 76-81% speedup)
|
|
*
|
|
* Set to `true` when you need to:
|
|
* - Compute similarity on this specific entity's vector
|
|
* - Perform manual vector operations
|
|
* - Inspect embeddings for debugging
|
|
*
|
|
* **Note**: Search operations (`brain.find()`) generate vectors automatically,
|
|
* so you don't need this flag for search. Only for direct vector operations
|
|
* on a retrieved entity.
|
|
*
|
|
* @default false
|
|
*/
|
|
includeVectors?: boolean
|
|
}
|
|
|
|
/**
|
|
* Graph traversal parameters
|
|
*/
|
|
export interface TraverseParams {
|
|
from: string | string[] // Starting node(s)
|
|
direction?: 'out' | 'in' | 'both' // Traversal direction
|
|
types?: VerbType[] // Edge types to follow
|
|
depth?: number // Max depth (default: 2)
|
|
strategy?: 'bfs' | 'dfs' // Breadth or depth first
|
|
filter?: (entity: Entity, depth: number, path: string[]) => boolean
|
|
limit?: number // Max nodes to visit
|
|
}
|
|
|
|
// ============= Aggregation Engine Types =============
|
|
|
|
/**
|
|
* Supported aggregation operations
|
|
*/
|
|
export type AggregationOp =
|
|
| 'sum'
|
|
| 'count'
|
|
| 'avg'
|
|
| 'min'
|
|
| 'max'
|
|
| 'stddev'
|
|
| 'variance'
|
|
/** Exact percentile (requires `p` in `[0,1]` on the metric def) — value-multiset, delete-safe. */
|
|
| 'percentile'
|
|
/** Exact count of distinct values — value-multiset, delete-safe. */
|
|
| 'distinctCount'
|
|
|
|
/**
|
|
* Time window granularity for GROUP BY time dimensions
|
|
*/
|
|
export type TimeWindowGranularity =
|
|
| 'hour' | 'day' | 'week' | 'month' | 'quarter' | 'year'
|
|
| { seconds: number }
|
|
|
|
/**
|
|
* A GROUP BY dimension — one of:
|
|
* - a plain metadata field name (string),
|
|
* - a time-windowed field (`{ field, window }`), or
|
|
* - an **unnest** field (`{ field, unnest: true }`): the field holds an array, and the entity
|
|
* contributes once to a group per distinct element (e.g. `tags: string[]` → tag frequency).
|
|
* An entity whose unnest field is missing or an empty array contributes to no group.
|
|
*/
|
|
export type GroupByDimension =
|
|
| string
|
|
| { field: string; window: TimeWindowGranularity }
|
|
| { field: string; unnest: true }
|
|
|
|
/**
|
|
* Source filter for which entities feed into an aggregate
|
|
*/
|
|
export interface AggregateSource {
|
|
/** Filter by entity type(s) */
|
|
type?: NounType | NounType[]
|
|
/** Metadata filter (same syntax as find({ where })) */
|
|
where?: Record<string, unknown>
|
|
/** Multi-tenancy service filter */
|
|
service?: string
|
|
}
|
|
|
|
/**
|
|
* Full aggregate definition — registered via brain.defineAggregate()
|
|
*/
|
|
export interface AggregateDefinition {
|
|
/** Unique name for this aggregate (used in queries) */
|
|
name: string
|
|
/** Which entities contribute to this aggregate */
|
|
source: AggregateSource
|
|
/** Dimensions to group by */
|
|
groupBy: GroupByDimension[]
|
|
/** Named metrics to compute */
|
|
metrics: Record<string, AggregateMetricDef>
|
|
/** Control materialization of results as NounType.Measurement entities */
|
|
materialize?: boolean | { debounceMs?: number; trackSources?: boolean }
|
|
}
|
|
|
|
/**
|
|
* Single metric definition within an aggregate
|
|
*/
|
|
export interface AggregateMetricDef {
|
|
/** Aggregation operation */
|
|
op: AggregationOp
|
|
/** Metadata field to aggregate (required for all ops except count) */
|
|
field?: string
|
|
/** Percentile fraction in `[0, 1]` — required when `op === 'percentile'`, ignored otherwise. */
|
|
p?: number
|
|
}
|
|
|
|
/**
|
|
* Internal running state for a single metric.
|
|
* Tracks enough to compute all operations incrementally.
|
|
*/
|
|
export interface MetricState {
|
|
sum: number
|
|
count: number
|
|
min: number
|
|
max: number
|
|
/** Running M2 for Welford's online variance (sum of squared differences from mean) */
|
|
m2?: number
|
|
/**
|
|
* Value multiset (String(value) → occurrence count) for exact percentile + distinctCount.
|
|
* Maintained only for `percentile`/`distinctCount` metrics. JSON-serializable; mirrors
|
|
* Cortex's `value_counts` so JS and native agree bit-for-bit.
|
|
*/
|
|
valueCounts?: Record<string, number>
|
|
}
|
|
|
|
/**
|
|
* Internal state for one aggregate group (one combination of group key values)
|
|
*/
|
|
export interface AggregateGroupState {
|
|
/** The group key values (e.g., { category: 'food', period: '2024-01' }) */
|
|
groupKey: Record<string, string | number>
|
|
/** Running metric states keyed by metric name */
|
|
metrics: Record<string, MetricState>
|
|
/** Entity ID of the materialized Measurement entity (if materialized) */
|
|
materializedEntityId?: string
|
|
/** Timestamp of last update */
|
|
lastUpdated: number
|
|
}
|
|
|
|
/**
|
|
* Query parameters for reading aggregate results
|
|
*/
|
|
export interface AggregateQueryParams {
|
|
/** Name of the aggregate to query */
|
|
name: string
|
|
/** Filter aggregate groups by their key values */
|
|
where?: Record<string, unknown>
|
|
/**
|
|
* Filter groups by their computed METRIC values (SQL HAVING). Same BFO operators as
|
|
* `where`, but applied to the derived metric results plus `count`, e.g.
|
|
* `{ revenue: { greaterThan: 1000 } }`. Evaluated per group (O(groups), independent of
|
|
* entity count), before sort/pagination.
|
|
*/
|
|
having?: Record<string, unknown>
|
|
/** Sort by metric name or group key field */
|
|
orderBy?: string
|
|
/** Sort direction */
|
|
order?: 'asc' | 'desc'
|
|
/** Max results */
|
|
limit?: number
|
|
/** Skip N results */
|
|
offset?: number
|
|
}
|
|
|
|
/**
|
|
* A single aggregate result row
|
|
*/
|
|
export interface AggregateResult {
|
|
/** Group key values for this row */
|
|
groupKey: Record<string, string | number>
|
|
/** Computed metric values (derived from MetricState based on op) */
|
|
metrics: Record<string, number>
|
|
/** Total entity count in this group */
|
|
count: number
|
|
/** Entity ID of materialized Measurement (if materialized) */
|
|
entityId?: string
|
|
}
|
|
|
|
/**
|
|
* Provider interface for Cortex-accelerated aggregation.
|
|
* When registered as 'aggregation' provider, Brainy delegates to this.
|
|
*/
|
|
export interface AggregationProvider {
|
|
/** Register an aggregate definition (caches compiled definition for hot path) */
|
|
defineAggregate?(def: AggregateDefinition): void
|
|
|
|
/** Remove a registered aggregate definition */
|
|
removeAggregate?(name: string): void
|
|
|
|
/** Incrementally update aggregation state when an entity changes */
|
|
incrementalUpdate(
|
|
name: string,
|
|
def: AggregateDefinition,
|
|
entity: Record<string, unknown>,
|
|
op: 'add' | 'update' | 'delete',
|
|
prev?: Record<string, unknown>
|
|
): AggregateGroupState[]
|
|
|
|
/** Compute the group key for a given entity */
|
|
computeGroupKey(
|
|
entity: Record<string, unknown>,
|
|
groupBy: GroupByDimension[]
|
|
): Record<string, string | number>
|
|
|
|
/** Rebuild an entire aggregate from scratch */
|
|
rebuildAggregate(
|
|
def: AggregateDefinition,
|
|
entities: Array<Record<string, unknown>>
|
|
): Map<string, AggregateGroupState>
|
|
|
|
/** Query aggregate state with filtering/sorting/pagination */
|
|
queryAggregate(
|
|
state: Map<string, AggregateGroupState>,
|
|
params: AggregateQueryParams
|
|
): AggregateResult[]
|
|
|
|
/** Restore previously serialized state (called during init) */
|
|
restoreState?(data: string): void
|
|
|
|
/** Serialize internal state for persistence (called during flush) */
|
|
serializeState?(): string
|
|
}
|
|
|
|
// ============= Configuration =============
|
|
|
|
/**
|
|
* Integration Hub configuration
|
|
*
|
|
* Enables external tool integrations: Excel, Power BI, Google Sheets, etc.
|
|
* Works in all environments with zero external dependencies.
|
|
*
|
|
* @example
|
|
* ```typescript
|
|
* // Enable all integrations with defaults
|
|
* new Brainy({ integrations: true })
|
|
*
|
|
* // Custom configuration
|
|
* new Brainy({
|
|
* integrations: {
|
|
* basePath: '/api/v1',
|
|
* enable: ['odata', 'sheets']
|
|
* }
|
|
* })
|
|
* ```
|
|
*/
|
|
export interface IntegrationsConfig {
|
|
/** Base path for all integration endpoints (default: '') */
|
|
basePath?: string
|
|
|
|
/** Which integrations to enable (default: all) */
|
|
enable?: ('odata' | 'sheets' | 'sse' | 'webhooks')[] | 'all'
|
|
|
|
/** Per-integration config overrides */
|
|
config?: {
|
|
odata?: { basePath?: string }
|
|
sheets?: { basePath?: string }
|
|
sse?: { basePath?: string; heartbeatInterval?: number }
|
|
webhooks?: { maxRetries?: number }
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Operator-facing summary returned by `brain.stats()`. Designed to fit in a
|
|
* terminal screen so an incident responder can read it at a glance: counts,
|
|
* mode, lock owner, indexed fields, and index health flags.
|
|
*/
|
|
export interface BrainyStats {
|
|
/** Whether this instance can mutate. `reader` is set by `openReadOnly()` and `asOf()`. */
|
|
mode: 'writer' | 'reader'
|
|
/** Total entity (noun) count across all types. */
|
|
entityCount: number
|
|
/** Breakdown by NounType name (`person`, `event`, ...). Zero-count types omitted. */
|
|
entitiesByType: Record<string, number>
|
|
/** Total relationship (verb) count across all types. */
|
|
relationCount: number
|
|
/** Breakdown by VerbType name. Zero-count types omitted. */
|
|
relationsByType: Record<string, number>
|
|
/** Indexed metadata field names known to the metadata index. */
|
|
fieldRegistry: string[]
|
|
/** Per-index health flags. `true` = index has entries OR no entities exist yet. */
|
|
indexHealth: {
|
|
/** Vector index health (8.0 — open-core JS HNSW path or a native acceleration provider). */
|
|
vector: boolean
|
|
metadata: boolean
|
|
graph: boolean
|
|
}
|
|
/** Storage backend info — `backend` is the adapter class name. */
|
|
storage: {
|
|
backend: string
|
|
rootDir?: string
|
|
}
|
|
/**
|
|
* Writer lock metadata if one is currently held on this directory. Present
|
|
* for both writer instances (their own lock) and readers inspecting a
|
|
* directory with a live writer.
|
|
*/
|
|
writerLock?: {
|
|
pid: number
|
|
hostname: string
|
|
startedAt: string
|
|
lastHeartbeat: string
|
|
version: string
|
|
rootDir?: string
|
|
}
|
|
/** The Brainy library version this instance was built against. */
|
|
version: string
|
|
}
|
|
|
|
/**
|
|
* Brainy configuration
|
|
*/
|
|
export interface BrainyConfig {
|
|
/**
|
|
* Storage configuration: either a config object resolved through the
|
|
* storage factory, or a pre-constructed adapter instance (used directly —
|
|
* e.g. `storage: new MemoryStorage()`; Brainy's historical-query
|
|
* materializer hands a pre-populated instance through this path).
|
|
*/
|
|
storage?:
|
|
| {
|
|
/**
|
|
* Storage backend. Optional — a top-level `path` implies `'filesystem'`,
|
|
* so `storage: { path: '/data' }` works without it. `'auto'` (the
|
|
* default) picks filesystem on Node, memory in a browser.
|
|
*/
|
|
type?: 'auto' | 'memory' | 'filesystem'
|
|
/**
|
|
* **Canonical** directory for filesystem storage. The rest of the API
|
|
* already speaks `path` (`persist(path)`, `Brainy.load(path)`,
|
|
* `asOf(path)`, `restore(path)`). Specifying it implies
|
|
* `type: 'filesystem'`. Passed through to plugin-provided storage
|
|
* factories (e.g. native mmap providers) so they resolve the same root.
|
|
* @example
|
|
* new Brainy({ storage: { path: '/var/lib/app/data' } })
|
|
*/
|
|
path?: string
|
|
/**
|
|
* @deprecated REMOVED in 8.0 — passing `rootDirectory` THROWS. Use the
|
|
* canonical {@link path}.
|
|
*/
|
|
rootDirectory?: string
|
|
/**
|
|
* @deprecated REMOVED in 8.0 — a nested `options.path` / `options.rootDirectory`
|
|
* THROWS. Use the top-level {@link path}.
|
|
*/
|
|
options?: any
|
|
/**
|
|
* @deprecated REMOVED in 8.0 — `fileSystemStorage.path` / `.rootDirectory`
|
|
* THROWS. Use the top-level {@link path}.
|
|
*/
|
|
fileSystemStorage?: { path?: string; rootDirectory?: string; [key: string]: any }
|
|
}
|
|
| StorageAdapter
|
|
|
|
/**
|
|
* Disable the automatic index rebuild check during `init()`. By default
|
|
* Brainy auto-decides from dataset size: small datasets rebuild missing
|
|
* indexes inline, large datasets rebuild lazily on first query. Set `true`
|
|
* only when an operator wants full manual control via `repairIndex()`.
|
|
*/
|
|
disableAutoRebuild?: boolean
|
|
|
|
/**
|
|
* Vector index configuration (Brainy 8.0).
|
|
*
|
|
* Two knobs. No escape hatch. The algorithm-internal HNSW knobs
|
|
* (`M`, `efConstruction`, `efSearch`, `ml`, …) and DiskANN knobs
|
|
* (`pqM`, `searchListSize`, `alpha`, …) are deliberately not exposed —
|
|
* the `recall` preset covers the legitimate quality/latency tradeoff
|
|
* range, and the open-core defaults match Brainy 7.x's defaults
|
|
* byte-for-byte so a `'balanced'`-default upgrade is a no-op.
|
|
*
|
|
* **Closed-form contract** locked in handoff thread
|
|
* BRAINY-8.0-RENAME-COORDINATION § A.2 + § G.2 (cortex-confirmed 2026-06-09).
|
|
*/
|
|
vector?: {
|
|
/**
|
|
* Recall preset.
|
|
* - `'fast'` — minimum-latency search, accepts lower recall
|
|
* - `'balanced'` (default) — Brainy 7.x defaults, the right pick for almost everyone
|
|
* - `'accurate'` — maximum recall, accepts higher latency
|
|
*
|
|
* Means the same thing whether the open-core JS HNSW path is in play
|
|
* or a native vector provider has taken over the `'vector'` provider
|
|
* key — Brainy translates to HNSW knobs; native providers translate
|
|
* to their own (e.g. DiskANN's `defaultLSearch` / `defaultPaddingFactor`).
|
|
*/
|
|
recall?: 'fast' | 'balanced' | 'accurate'
|
|
|
|
/**
|
|
* Vector persistence mode. `'immediate'` writes per-noun graph state on
|
|
* every `add()`; durable but slower. `'deferred'` writes only on
|
|
* `flush()` / `close()`; faster bulk ingest.
|
|
*
|
|
* **Auto-selected from the storage adapter when omitted:** `'immediate'`
|
|
* on filesystem storage (durability is the point of a persistent
|
|
* backend), `'deferred'` on memory storage (nothing survives the process
|
|
* anyway, so per-add persistence writes are pure overhead).
|
|
*
|
|
* Brainy 7.x called this `hnswPersistMode` at the top level. The 8.0
|
|
* surface folds it under `config.vector` alongside `recall`.
|
|
*/
|
|
persistMode?: 'immediate' | 'deferred'
|
|
}
|
|
|
|
/**
|
|
* Generational-history retention policy (8.0 MVCC).
|
|
*
|
|
* Every `transact()` produces an immutable generation record-set that
|
|
* serves historical reads (`asOf()`, pinned `Db` values). Without
|
|
* compaction those record-sets accumulate forever, so Brainy
|
|
* **auto-compacts on every `flush()` and `close()`** with this policy.
|
|
* A generation is reclaimed only when it is older than BOTH bounds
|
|
* (and never while a live `Db` pin protects it):
|
|
*
|
|
* - `retainGenerations` — keep at least the N most recent committed
|
|
* generations (default: `100`).
|
|
* - `retainMs` — keep everything committed within the window
|
|
* (default: 7 days).
|
|
*
|
|
* Escape hatches: raise either bound for longer time-travel windows,
|
|
* or set `autoCompact: false` to manage history manually via
|
|
* `brain.compactHistory()`. Long-term archives belong in
|
|
* `db.persist(path)` snapshots, which compaction never touches.
|
|
*/
|
|
history?: {
|
|
/** Keep at least this many recent committed generations (default: 100). */
|
|
retainGenerations?: number
|
|
/** Keep generations committed within this window in ms (default: 604800000 = 7 days). */
|
|
retainMs?: number
|
|
/** Run compaction automatically on flush()/close() (default: true). */
|
|
autoCompact?: boolean
|
|
}
|
|
|
|
// Memory management options
|
|
maxQueryLimit?: number // Override auto-detected query result limit (max: 100000)
|
|
reservedQueryMemory?: number // Memory reserved for queries in bytes (e.g., 1073741824 = 1GB)
|
|
|
|
/**
|
|
* Controls when the WASM embedding engine is initialized.
|
|
*
|
|
* **Adaptive default (8.0):** when omitted, the engine eagerly initializes
|
|
* during `init()` whenever the WASM embedder is the *active* one — i.e. no
|
|
* native `'embeddings'` provider is registered — and this instance is a
|
|
* writer (not `mode: 'reader'`) running outside unit tests. The WASM module
|
|
* (≈93MB with the embedded model) takes 90-140s to compile on throttled
|
|
* CPUs, so paying that during boot rather than on the first `embed()`-driven
|
|
* call is the right default for a single-process server.
|
|
*
|
|
* The adaptive path skips itself automatically when a native embeddings
|
|
* provider owns embeddings, in reader-mode (readers query existing vectors
|
|
* and never embed), and in unit-test mode (kept fast via the mock embedder).
|
|
*
|
|
* - `true` — force eager init during `init()` (the adaptive default already
|
|
* does this for the active-embedder writer case; set it explicitly to be
|
|
* unambiguous).
|
|
* - `false` — explicit override to force lazy init (first `embed()` call)
|
|
* even when this instance is the active embedder.
|
|
*/
|
|
eagerEmbeddings?: boolean
|
|
|
|
// Plugin configuration
|
|
// Controls which plugins are loaded during init()
|
|
// - undefined (default): Auto-detect installed plugins (@soulcraft/cor, etc.)
|
|
// - false: No plugins — skip auto-detection entirely
|
|
// - []: No plugins — skip auto-detection entirely
|
|
// - ['@soulcraft/cor']: Load only specified plugins, no auto-detection
|
|
plugins?: string[] | false
|
|
|
|
// Logging configuration
|
|
verbose?: boolean // Enable verbose logging
|
|
silent?: boolean // Suppress all logging output
|
|
|
|
// Integration Hub
|
|
// Enable external tool integrations: Excel, Power BI, Google Sheets, etc.
|
|
// - true: Enable all integrations with default paths
|
|
// - false/undefined: Disable integrations (default)
|
|
// - IntegrationsConfig: Custom configuration
|
|
integrations?: boolean | IntegrationsConfig
|
|
|
|
// Migration configuration
|
|
// - true/undefined (default): Automatically run pending data migrations during
|
|
// init() for small datasets (<10K entities). Larger datasets defer with a
|
|
// notice — call brain.migrate() explicitly (optionally with backupTo).
|
|
// - false: Never auto-run; log a notice when pending migrations exist.
|
|
autoMigrate?: boolean
|
|
|
|
/**
|
|
* Brain-wide subtype enforcement mode.
|
|
*
|
|
* **Default-on since 8.0.0** (`undefined` resolves to `true`; in 7.30.x this
|
|
* was an opt-in flag defaulting to `false`).
|
|
*
|
|
* - `true` / `undefined` (default): every `add()` / `addMany()` / `update()` /
|
|
* `relate()` / `relateMany()` / `updateRelation()` rejects writes where the
|
|
* entity's NounType (or relationship's VerbType) has no non-empty `subtype`
|
|
* value. Brainy's own VFS infrastructure writes (`metadata.isVFSEntity` /
|
|
* `metadata.isVFS`) bypass the check.
|
|
* - `{ except: [NounType.Thing, NounType.Custom] }`: same as `true`, but the
|
|
* listed types are allowed through without a subtype. Use for genuine
|
|
* catch-all types where no subtype makes sense.
|
|
* - `false`: disable the brain-wide check entirely. Last-resort escape hatch
|
|
* for opening pre-8.0 data — run `brain.audit()` to find the gaps, back-fill
|
|
* with `brain.fillSubtypes(rules)`, then remove the opt-out so the default
|
|
* enforcement protects new writes.
|
|
*
|
|
* Per-type registrations always compose with the brain-wide flag — a type
|
|
* registered with `requireSubtype(type, { required: true })` is always
|
|
* enforced regardless of this flag.
|
|
*/
|
|
requireSubtype?: boolean | { except: Array<import('../types/graphTypes.js').NounType | import('../types/graphTypes.js').VerbType> }
|
|
|
|
/**
|
|
* Process role for multi-process safety on filesystem storage.
|
|
*
|
|
* - `'writer'` (default): acquires an exclusive lock on the storage directory at
|
|
* `init()` and refuses to open if another live writer holds the lock. Required
|
|
* for any instance that calls `add`/`update`/`remove`/`relate` or any other
|
|
* mutation. Released on `close()` and on process exit/SIGINT/SIGTERM.
|
|
* - `'reader'`: opens without acquiring the writer lock. Coexists with a live
|
|
* writer and with other readers. All mutation methods throw
|
|
* `Cannot mutate a read-only Brainy instance`. Use `Brainy.openReadOnly()`
|
|
* as a convenience factory.
|
|
*
|
|
* See `docs/concepts/multi-process.md` for the full coordination model.
|
|
*/
|
|
mode?: 'writer' | 'reader'
|
|
|
|
/**
|
|
* Bypass the writer-lock check at init even when another writer holds the lock.
|
|
* Use only when you know the existing lock is stale and stale-detection
|
|
* (PID liveness + heartbeat) cannot prove it. Logs a warning regardless.
|
|
*/
|
|
force?: boolean
|
|
|
|
/**
|
|
* How write paths react when an untyped (JavaScript) caller smuggles a
|
|
* Brainy-reserved field (`RESERVED_ENTITY_FIELDS` / `RESERVED_RELATION_FIELDS`
|
|
* — `confidence`, `weight`, `subtype`, `visibility`, `service`, `createdBy`,
|
|
* `noun`/`verb`, `data`, `createdAt`, `updatedAt`, `_rev`) **inside the
|
|
* `metadata` bag** of `add()` / `update()` / `relate()` / `updateRelation()`
|
|
* (and their `transact()` / `with()` mirrors). TypeScript callers can't write
|
|
* these shapes at all — the compile-time guard on the metadata param types
|
|
* (`NoReservedEntityKeys` / `NoReservedRelationKeys`) rejects a literal
|
|
* reserved key — so this policy only governs untyped callers that slip one
|
|
* past the compiler.
|
|
*
|
|
* - `'throw'` (**default, 8.0**): a reserved key in the bag throws a clear
|
|
* `Error` naming the offending key(s) and the correct write path. No silent
|
|
* remap, no data loss, no surprise. This is the 8.0 "no silent failures"
|
|
* contract.
|
|
* - `'warn'`: legacy remapping with a loud, one-shot (per key, per process)
|
|
* warning for EVERY reserved key found — user-mutable fields are remapped to
|
|
* their dedicated top-level param (top-level wins when both are supplied),
|
|
* system-managed fields are dropped. Use while migrating untyped call sites.
|
|
* - `'remap'`: the pre-8.0 silent remapping, no warning. Last-resort
|
|
* compatibility hatch for code that intentionally relies on the bag path.
|
|
*
|
|
* @default 'throw'
|
|
*/
|
|
reservedFieldPolicy?: 'throw' | 'warn' | 'remap'
|
|
}
|
|
|
|
// ============= Neural API Types =============
|
|
|
|
/**
|
|
* Neural similarity parameters
|
|
*/
|
|
export interface NeuralSimilarityParams {
|
|
between?: [any, any] // Compare two items
|
|
items?: any[] // Compare multiple items
|
|
explain?: boolean // Return detailed breakdown
|
|
}
|
|
|
|
/**
|
|
* Neural clustering parameters
|
|
*/
|
|
export interface NeuralClusterParams {
|
|
items?: string[] | Entity[] // Items to cluster (or all)
|
|
algorithm?: 'hierarchical' | 'kmeans' | 'dbscan' | 'spectral'
|
|
params?: {
|
|
k?: number // Number of clusters (kmeans)
|
|
threshold?: number // Distance threshold (hierarchical)
|
|
epsilon?: number // DBSCAN epsilon
|
|
minPoints?: number // DBSCAN min points
|
|
}
|
|
visualize?: boolean // Return visualization data
|
|
}
|
|
|
|
/**
|
|
* Neural anomaly detection parameters
|
|
*/
|
|
export interface NeuralAnomalyParams {
|
|
threshold?: number // Standard deviations (default: 2.5)
|
|
type?: NounType // Check specific type
|
|
method?: 'isolation' | 'lof' | 'statistical' | 'autoencoder'
|
|
returnScores?: boolean // Return anomaly scores
|
|
}
|
|
|
|
// ============= Content Extraction Types =============
|
|
|
|
/**
|
|
* Detected content type for smart text extraction
|
|
*
|
|
*/
|
|
export type ContentType = 'plaintext' | 'richtext-json' | 'html' | 'markdown'
|
|
|
|
/**
|
|
* Content category for extracted segments
|
|
*
|
|
* Universal categories that work across documents, code, and UI:
|
|
* - 'title': Names the subject — headings, identifiers, labels, JSON keys
|
|
* - 'annotation': Human explanation — comments, docstrings, captions, alt text
|
|
* - 'content': Body substance — paragraphs, list items, flowing text
|
|
* - 'value': Data literals — strings, numbers, form values, error messages
|
|
* - 'code': Unparsed code blocks (custom parsers decompose into above)
|
|
* - 'structural': Boilerplate — keywords, operators, punctuation, formatting
|
|
*
|
|
* Built-in extractors produce: 'title', 'content', 'code'.
|
|
* All 6 categories are available for custom parsers.
|
|
*/
|
|
export type ContentCategory = 'title' | 'annotation' | 'content' | 'value' | 'code' | 'structural'
|
|
|
|
/**
|
|
* A segment of extracted text with its content category
|
|
*
|
|
*/
|
|
export interface ExtractedSegment {
|
|
/** The extracted text content */
|
|
text: string
|
|
/** What role this text plays: 'title' | 'annotation' | 'content' | 'value' | 'code' | 'structural' */
|
|
contentCategory: ContentCategory
|
|
}
|
|
|
|
// ============= Semantic Highlighting Types =============
|
|
|
|
/**
|
|
* Parameters for hybrid highlighting
|
|
*
|
|
* Zero-config highlighting that returns both text (exact) and semantic (concept) matches.
|
|
* Perfect for UI highlighting at different levels.
|
|
*
|
|
* Added contentType hint and contentExtractor callback for structured text
|
|
* (rich-text JSON, HTML, Markdown). Auto-detects format when not specified.
|
|
*
|
|
* @example
|
|
* ```typescript
|
|
* // Plain text
|
|
* const highlights = await brain.highlight({
|
|
* query: "david the warrior",
|
|
* text: "David Smith is a brave fighter who battles dragons"
|
|
* })
|
|
*
|
|
* // Rich-text JSON (auto-detected)
|
|
* const highlights = await brain.highlight({
|
|
* query: "david the warrior",
|
|
* text: JSON.stringify(tiptapDocument)
|
|
* })
|
|
*
|
|
* // Custom extractor (for proprietary formats)
|
|
* const highlights = await brain.highlight({
|
|
* query: "function",
|
|
* text: sourceCode,
|
|
* contentExtractor: (text) => treeSitterParse(text)
|
|
* })
|
|
* ```
|
|
*/
|
|
export interface HighlightParams {
|
|
/** The search query to match against */
|
|
query: string
|
|
|
|
/** The text to highlight (e.g., entity.data) */
|
|
text: string
|
|
|
|
/** Granularity of highlighting: 'word' (default), 'phrase', or 'sentence' */
|
|
granularity?: 'word' | 'phrase' | 'sentence'
|
|
|
|
/** Minimum semantic similarity score for semantic matches (default: 0.5) */
|
|
threshold?: number
|
|
|
|
/**
|
|
* Optional content type hint to skip auto-detection.
|
|
* When omitted, the content type is detected from the text content.
|
|
*/
|
|
contentType?: ContentType
|
|
|
|
/**
|
|
* Optional custom content extractor function.
|
|
* When provided, bypasses built-in detection and extraction entirely.
|
|
* Use this to plug in custom parsers (tree-sitter, Monaco, proprietary formats).
|
|
*/
|
|
contentExtractor?: (text: string) => ExtractedSegment[]
|
|
}
|
|
|
|
/**
|
|
* A highlight showing which text matched the query
|
|
*
|
|
* matchType tells the UI how to style the highlight:
|
|
* - 'text': Exact word match (strongest signal, highest confidence)
|
|
* - 'semantic': Conceptually similar match (may need softer highlight)
|
|
*/
|
|
export interface Highlight {
|
|
/** The text that matched */
|
|
text: string
|
|
|
|
/** Match score (0-1). For text matches, always 1.0. For semantic, varies by similarity. */
|
|
score: number
|
|
|
|
/** Position in original text [start, end] */
|
|
position: [number, number]
|
|
|
|
/** Match type: 'text' (exact word match) or 'semantic' (concept match) */
|
|
matchType: 'text' | 'semantic'
|
|
|
|
/**
|
|
* Content category of the source segment: 'title' | 'annotation' | 'content' | 'value' | 'code' | 'structural'.
|
|
* Present when the input text was structured (JSON, HTML, Markdown).
|
|
*/
|
|
contentCategory?: ContentCategory
|
|
}
|
|
|
|
// ============= Export all types =============
|
|
|
|
export * from './graphTypes.js' // Re-export NounType, VerbType, etc.
|