Class: RubyLLM::Contract::Eval::RetryOptimizer

Inherits:
Object
  • Object
show all
Defined in:
lib/ruby_llm/contract/eval/retry_optimizer.rb

Overview

Runs compare_models on ALL evals for a step, builds a score matrix, identifies the constraining eval, and suggests an escalation chain.

optimizer = RetryOptimizer.new(step: MyStep, candidates: [...], context: {})
result = optimizer.call
result.print_summary
result.to_dsl  # => copy-paste retry_policy

Defined Under Namespace

Classes: Result

Instance Method Summary collapse

Constructor Details

#initialize(step:, candidates:, context: {}, min_score: DEFAULT_MIN_SCORE, runs: 1, production_mode: nil) ⇒ RetryOptimizer

Returns a new instance of RetryOptimizer.



101
102
103
104
105
106
107
108
109
110
 
# File 'lib/ruby_llm/contract/eval/retry_optimizer.rb', line 101

def initialize(step:, candidates:, context: {},
               min_score: DEFAULT_MIN_SCORE,
               runs: 1, production_mode: nil)
  @step = step
  @candidates = candidates
  @context = context
  @min_score = min_score
  @runs = runs
  @production_mode = production_mode
end

Instance Method Details

#call ⇒ Object 



112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
 
# File 'lib/ruby_llm/contract/eval/retry_optimizer.rb', line 112

def call
  evals = @step.eval_names
  return empty_result(evals) if evals.empty?

  score_matrix = {}
  evals.each do |eval_name|
    # `retry_policy_override: nil` in context disables the step's
    # class-level retry policy for this comparison run — see
    # step/base.rb#runtime_settings, which honours the key when
    # present (even when value is nil). Replaces the prior
    # `define_singleton_method(:retry_policy)` mutation, which was
    # not thread-safe across concurrent optimizer calls.
    comparison = @step.compare_models(
      eval_name,
      candidates: @candidates,
      context: @context.merge(retry_policy_override: nil),
      runs: @runs,
      production_mode: @production_mode
    )
    score_matrix[eval_name] = extract_scores(comparison)
  end

  labels = score_matrix.values.flat_map(&:keys).uniq
  constraining = find_constraining_eval(score_matrix, labels)
  chain, details = build_chain(score_matrix, labels, evals)

  Result.new(
    step_name: @step.name || @step.to_s,
    eval_names: evals,
    candidate_labels: labels,
    score_matrix: score_matrix,
    constraining_eval: constraining,
    chain: chain,
    chain_details: details
  )
end