Class: CompletionKit::Run

Inherits:
ApplicationRecord show all
Includes:
Turbo::Broadcastable
Defined in:
app/models/completion_kit/run.rb

Constant Summary collapse

STATUSES = 
%w[pending running completed failed].freeze

Constants inherited from ApplicationRecord

ApplicationRecord::TenantScopedUniquenessValidator

Instance Method Summary collapse

Instance Method Details

#as_json(options = {}) ⇒ Object 



139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
 
# File 'app/models/completion_kit/run.rb', line 139

def as_json(options = {})
  snap = progress_snapshot
  {
    id: id, name: name, status: status, prompt_id: prompt_id,
    dataset_id: dataset_id, judge_model: judge_model, temperature: temperature,
    created_at: created_at, updated_at: updated_at,
    responses_count: responses.count, avg_score: avg_score,
    progress_current: snap[:generated_done],
    progress_total: snap[:generated_total],
    progress: {
      generated: { done: snap[:generated_done], total: snap[:generated_total], failed: snap[:generated_failed] },
      judged:    { done: snap[:judged_done],    total: snap[:judged_total],    failed: snap[:judged_failed] }
    },
    failed_response_ids: responses.where(status: "failed").pluck(:id),
    failure_summary: failure_summary,
    error_message: error_message,
    metric_ids: metric_ids
  }
end

#avg_score ⇒ Object 



56
57
58
59
60
61
62
 
# File 'app/models/completion_kit/run.rb', line 56

def avg_score
  all_reviews = responses.flat_map(&:reviews)
  scores = all_reviews.map(&:ai_score).compact.map(&:to_f)
  return nil if scores.empty?

  (scores.sum / scores.length).round(2)
end

#generate_responses! ⇒ Object 



112
113
114
 
# File 'app/models/completion_kit/run.rb', line 112

def generate_responses!
  start!
end

#judge_configured? ⇒ Boolean

Returns:

  • (Boolean) 


44
45
46
 
# File 'app/models/completion_kit/run.rb', line 44

def judge_configured?
  judge_model.present? && metrics.any? && ApiConfig.valid_for_model?(judge_model)
end

#mark_completed! ⇒ Object 



20
21
22
23
 
# File 'app/models/completion_kit/run.rb', line 20

def mark_completed!
  update!(status: "completed")
  broadcast_ui
end

#metric_averages ⇒ Object 



64
65
66
67
68
69
70
 
# File 'app/models/completion_kit/run.rb', line 64

def metric_averages
  all_reviews = responses.flat_map(&:reviews).select { |r| r.ai_score.present? }
  all_reviews.group_by(&:metric_name).map do |name, reviews|
    scores = reviews.map { |r| r.ai_score.to_f }
    { name: name, avg: (scores.sum / scores.length).round(1) }
  end
end

#outstanding_work_zero? ⇒ Boolean

Returns:

  • (Boolean) 


25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
 
# File 'app/models/completion_kit/run.rb', line 25

def outstanding_work_zero?
  return false if responses.where.not(status: Response::TERMINAL_STATUSES).exists?

  metric_ids = metrics.pluck(:id)
  return true if metric_ids.empty?

  succeeded_response_ids = responses.where(status: "succeeded").pluck(:id)
  expected_reviews = succeeded_response_ids.size * metric_ids.size
  return true if expected_reviews.zero?

  terminal_review_count = Review.where(
    response_id: succeeded_response_ids,
    metric_id: metric_ids,
    status: Review::TERMINAL_STATUSES
  ).count

  terminal_review_count >= expected_reviews
end

#progress_snapshot ⇒ Object 



116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
 
# File 'app/models/completion_kit/run.rb', line 116

def progress_snapshot
  generated_done = responses.where(status: "succeeded").count
  generated_failed = responses.where(status: "failed").count
  generated_total = progress_total

  metric_count = metrics.count
  succeeded_count = generated_done
  judged_total = succeeded_count * metric_count
  judged_done = Review.joins(:response)
    .where(completion_kit_responses: { run_id: id }, status: "succeeded").count
  judged_failed = Review.joins(:response)
    .where(completion_kit_responses: { run_id: id }, status: "failed").count

  {
    generated_done: generated_done,
    generated_total: generated_total,
    generated_failed: generated_failed,
    judged_done: judged_done,
    judged_total: judged_total,
    judged_failed: judged_failed
  }
end

#replace_metrics!(metric_ids) ⇒ Object 



48
49
50
51
52
53
54
 
# File 'app/models/completion_kit/run.rb', line 48

def replace_metrics!(metric_ids)
  return unless metric_ids
  run_metrics.delete_all
  Array(metric_ids).reject(&:blank?).each_with_index do |metric_id, index|
    run_metrics.create!(metric_id: metric_id, position: index + 1)
  end
end

#start! ⇒ Object 



72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
 
# File 'app/models/completion_kit/run.rb', line 72

def start!
  rows = if dataset
           CsvProcessor.process_self(self)
         else
           [{}]
         end

  return fail_with_summary!("Dataset has no rows") if rows.empty?

  client = LlmClient.for_model(prompt.llm_model, ApiConfig.for_model(prompt.llm_model))
  unless client.configured?
    return fail_with_summary!("LLM API not configured: #{client.configuration_errors.join(', ')}")
  end

  transaction do
    responses.destroy_all
    update!(
      status: "running",
      progress_current: 0,
      progress_total: rows.length,
      failure_summary: nil,
      error_message: nil
    )
    rows.each_with_index do |row, index|
      input = row.empty? ? nil : row.to_json
      response = responses.create!(
        status: "pending",
        row_index: index,
        input_data: input,
        expected_output: row["expected_output"]
      )
      GenerateRowJob.perform_later(id, response.id)
    end
  end

  broadcast_ui
  broadcast_clear_responses
  true
end