batch

Documentation

Index

• Constants
• func VectorizeBatch[T []float32](ctx context.Context, objs []*models.Object, skipObject []bool, ...) ([]T, []models.AdditionalProperties, map[int]error)
• type Batch
  • func NewBatchVectorizer[T dto.Embedding](client BatchClient[T], maxBatchTime time.Duration, settings Settings, ...) *Batch[T]
  • func (b *Batch[T]) SubmitBatchAndWait(ctx context.Context, cfg moduletools.ClassConfig, skipObject []bool, ...) ([]T, map[int]error)
• type BatchClient
• type BatchJob
• type Settings
• type TokenizerFuncType
  • func ReturnBatchTokenizer(multiplier float32, moduleName string, lowerCaseInput bool) TokenizerFuncType

Constants

const BatchChannelSize = 100

Functions

func VectorizeBatch

func VectorizeBatch[T []float32](ctx context.Context, objs []*models.Object, skipObject []bool, cfg moduletools.ClassConfig, logger logrus.FieldLogger, objectVectorizer objectVectorizer) ([]T, []models.AdditionalProperties, map[int]error)

Types

type Batch

type Batch[T dto.Embedding] struct {
	// contains filtered or unexported fields
}

func NewBatchVectorizer

func NewBatchVectorizer[T dto.Embedding](client BatchClient[T], maxBatchTime time.Duration, settings Settings, logger logrus.FieldLogger, label string) *Batch[T]

func (*Batch[T]) SubmitBatchAndWait

func (b *Batch[T]) SubmitBatchAndWait(ctx context.Context, cfg moduletools.ClassConfig, skipObject []bool, tokenCounts []int, texts []string) ([]T, map[int]error)

type BatchClient

type BatchClient[T dto.Embedding] interface {
	Vectorize(ctx context.Context, input []string,
		config moduletools.ClassConfig) (*modulecomponents.VectorizationResult[T], *modulecomponents.RateLimits, int, error)
	GetVectorizerRateLimit(ctx context.Context, config moduletools.ClassConfig) *modulecomponents.RateLimits
	GetApiKeyHash(ctx context.Context, config moduletools.ClassConfig) [32]byte
}

type BatchJob

type BatchJob[T dto.Embedding] struct {
	// contains filtered or unexported fields
}

type Settings

type Settings struct {
	TokenMultiplier    float32
	MaxTimePerBatch    float64
	MaxObjectsPerBatch int
	MaxTokensPerBatch  func(cfg moduletools.ClassConfig) int
	HasTokenLimit      bool
	ReturnsRateLimit   bool
}

type TokenizerFuncType

type TokenizerFuncType func(ctx context.Context, objects []*models.Object, skipObject []bool, cfg moduletools.ClassConfig, objectVectorizer *objectsvectorizer.ObjectVectorizer) ([]string, []int, bool, error)

func ReturnBatchTokenizer

func ReturnBatchTokenizer(multiplier float32, moduleName string, lowerCaseInput bool) TokenizerFuncType