Multiple scores from a single Scorer

[David Elner (delner)]

Scorers can now return multiple named metrics in one call by returning an array of score hashes. This is useful when several dimensions of quality can be computed together — sharing one LLM call or one pass over the output — rather than running separate scorers for each metric.

There are no breaking changes: existing scorers that return a single value or {score:, metadata:} hash continue to work.

Usage

Return an Array of hashes instead of a single value. Each hash requires :name and :score; :metadata is optional:

Braintrust::Scorer.new("summary_quality") do |output:, expected:|
  words = output.downcase.split
  key_terms = expected[:key_terms]
  covered = key_terms.count { |t| words.include?(t) }

  [
    {name: "coverage",    score: covered.to_f / key_terms.size, metadata: {missing: key_terms - words}},
    {name: "conciseness", score: words.size <= expected[:max_words] ? 1.0 : 0.0}
  ]
end

Each metric appears as its own score column in the Braintrust UI.