hybrid

Documentation

Index

• Constants
• func FusionRanked(weights []float64, resultSets [][]*search.Result, setNames []string) []*search.Result
• func FusionRelativeScore(weights []float64, resultSets [][]*search.Result, names []string, ...) []*search.Result
• func HybridCombiner(ctx context.Context, params *Params, resultSet [][]*search.Result, ...) ([]*search.Result, error)
• func Search(ctx context.Context, params *Params, logger logrus.FieldLogger, ...) ([]*search.Result, error)
• type Params
• type Results

Constants

const DefaultLimit = 100

Functions

func FusionRanked

func FusionRanked(weights []float64, resultSets [][]*search.Result, setNames []string) []*search.Result

func FusionRelativeScore

func FusionRelativeScore(weights []float64, resultSets [][]*search.Result, names []string, descending bool) []*search.Result

FusionRelativeScore uses the relative differences in the scores from keyword and vector search to combine the results. This method retains more information than ranked fusion and should result in better results.

The scores from each result are normalized between 0 and 1, e.g. the maximum score becomes 1 and the minimum 0 and the other scores are in between, keeping their relative distance to the other scores. Example:

Input score = [1, 8, 6, 11] => [0, 0.7, 0.5, 1]

The normalized scores are then combined using their respective weight and the combined scores are sorted

func HybridCombiner

func HybridCombiner(ctx context.Context, params *Params, resultSet [][]*search.Result, weights []float64, names []string, logger logrus.FieldLogger, postProc postProcFunc) ([]*search.Result, error)

Search combines the result sets using Reciprocal Rank Fusion or Relative Score Fusion

func Search

func Search(ctx context.Context, params *Params, logger logrus.FieldLogger, sparseSearch sparseSearchFunc, denseSearch denseSearchFunc, postProc postProcFunc, modules modulesProvider, schemaGetter uc.SchemaGetter, targetVectorParamHelper targetVectorParamHelper) ([]*search.Result, error)

Search executes sparse and dense searches and combines the result sets using Reciprocal Rank Fusion

Types

type Params

type Params struct {
	*searchparams.HybridSearch
	Keyword *searchparams.KeywordRanking
	Class   string
	Autocut int
}

type Results

type Results []*search.Result