kdrant-core/dev.kdrant.dsl/RelevanceFeedbackBuilder

RelevanceFeedbackBuilder

class RelevanceFeedbackBuilder

DSL for Qdrant's relevance-feedback query.

Constructors

RelevanceFeedbackBuilder

constructor()

Functions

fun feedback(example: VectorInput, score: Float)

Add one result and its relevance score from the feedback provider.

fun naive(a: Float, b: Float, c: Float)

Use Qdrant's built-in linear strategy and its trained coefficients.

fun target(id: PointId)

The stored vector used for the original query.

fun target(input: VectorInput)

The original query as any supported vector input.

fun target(values: List<Float>)

The dense vector used for the original query.