judge

Documentation

Overview

Package judge evaluates arbitrary subjects with a structured-output chat model. Subject validation and prompt construction remain owned by callers.

Index

• type Config
• type Evaluator
  • func NewEvaluator[T any](config Config[T]) (*Evaluator[T], error)
  • func (e *Evaluator[T]) Evaluate(ctx context.Context, subject T) (eval.Report, error)
• type Prompt

Types

type Config

type Config[T any] struct {
	Model   chat.Model
	Metric  eval.Metric
	Prompt  Prompt[T]
	Options chat.Options
	// Threshold is optional. Without one, evaluation produces a score without
	// inventing a pass/fail decision.
	Threshold *eval.Score
	Samples   int
}

type Evaluator

type Evaluator[T any] struct {
	// contains filtered or unexported fields
}

func NewEvaluator

func NewEvaluator[T any](config Config[T]) (*Evaluator[T], error)

func (*Evaluator[T]) Evaluate

func (e *Evaluator[T]) Evaluate(ctx context.Context, subject T) (eval.Report, error)

type Prompt

type Prompt[T any] func(T) (chat.Message, error)