class RubyLLM::Judgment

The typed answers to a set of questions, with model identity and accounting. Obtain one through RubyLLM.judge or Judge.judge.

judgment[:urgent].probability
judgment[:department].choice
judgment[:frustration].score

Attributes

answers [R]

Returns the answers keyed by their declared String or Symbol names.

model [R]

Returns the actual model ID reported by the provider.

raw [R]

Returns the original provider response.

Public Instance Methods

[] (name)

Source

# File lib/ruby_llm/judgment.rb, line 34
def [](name)
  answers[@answer_keys[name.to_s]]
end

Returns the named answer, accepting a String or Symbol, or nil if absent.

cost ()

Source

# File lib/ruby_llm/judgment.rb, line 54
def cost
  return ruby_llm_usage_cost unless ruby_llm_usage_entries.empty?

  Cost.new(tokens:, model: @model_info)
end

Returns cost across all provider attempts, preserving unknown amounts.

each (&)

Source

# File lib/ruby_llm/judgment.rb, line 44
def each(&)
  answers.each(&)
end

Yields each question name and answer, or returns an Enumerator.

fetch (name)

Source

# File lib/ruby_llm/judgment.rb, line 39
def fetch(name)
  answers.fetch(@answer_keys.fetch(name.to_s))
end

Returns the named answer or raises KeyError if it is absent.

to_h ()

Source

# File lib/ruby_llm/judgment.rb, line 61
def to_h
  { model:, answers: answers.transform_values(&:to_h), tokens: tokens.to_h, cost: cost.to_h }
end

Returns the model, typed answers, token usage, and cost as a Hash.

tokens ()

Source

# File lib/ruby_llm/judgment.rb, line 49
def tokens
  ruby_llm_usage_entries.empty? ? @tokens : ruby_llm_usage_tokens
end

Returns token usage across all provider attempts.