Skip to main content

@elizaos/core v1.0.0-beta.51 / BM25Options

Interface: BM25Options

BM25 Options Interface. Extends TokenizerOptions and adds BM25 specific parameters.

Extends​

  • TokenizerOptions

Properties​

stopWords?​

optional stopWords: Set<string>

A set of words to be ignored during tokenization. Defaults to an empty set.

Inherited from​

TokenizerOptions.stopWords

Defined in​

packages/core/src/search.ts:770

minLength?​

optional minLength: number

The minimum length for a token to be kept. Defaults to 2. Numeric tokens are always kept.

Inherited from​

TokenizerOptions.minLength

Defined in​

packages/core/src/search.ts:772

stemming?​

optional stemming: boolean

Whether to apply stemming to tokens. Defaults to false.

Inherited from​

TokenizerOptions.stemming

Defined in​

packages/core/src/search.ts:774

stemmingRules?​

optional stemmingRules: StemmingRule[]

Custom stemming rules to apply before the default Porter2 stemmer. Defaults to an empty array.

Inherited from​

TokenizerOptions.stemmingRules

Defined in​

packages/core/src/search.ts:776

k1?​

optional k1: number

Term frequency saturation parameter (k1). Controls how quickly term frequency saturates. Higher values mean TF contributes more significantly even for high counts. Typical values are between 1.2 and 2.0. Default: 1.2.

Defined in​

packages/core/src/search.ts:1000

b?​

optional b: number

Document length normalization parameter (b). Controls the influence of document length. 0 means no length normalization, 1 means full normalization. Typical values are around 0.75. Default: 0.75.

Defined in​

packages/core/src/search.ts:1006

fieldBoosts?​

optional fieldBoosts: object

A dictionary defining boost factors for specific document fields. Terms found in fields with higher boost factors will contribute more to the score. Example: { title: 2, body: 1 }. Default: {} (no boosts).

Index Signature​

[key: string]: number

Defined in​

packages/core/src/search.ts:1012