Class: Hyrum::Validators::QualityValidator

Inherits:
Object
  • Object
show all
Defined in:
lib/hyrum/validators/quality_validator.rb

Constant Summary collapse

DIVERSITY_WEIGHT =
0.5
SIMILARITY_WEIGHT =
0.5
MIN_DIVERSITY_THRESHOLD =
30.0
MIN_SIMILARITY_THRESHOLD =
85.0

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(original_message, messages, options) ⇒ QualityValidator

Returns a new instance of QualityValidator.



13
14
15
16
17
# File 'lib/hyrum/validators/quality_validator.rb', line 13

def initialize(original_message, messages, options)
  @original_message = original_message
  @messages = messages
  @options = options
end

Instance Attribute Details

#messagesObject (readonly)

Returns the value of attribute messages.



11
12
13
# File 'lib/hyrum/validators/quality_validator.rb', line 11

def messages
  @messages
end

#optionsObject (readonly)

Returns the value of attribute options.



11
12
13
# File 'lib/hyrum/validators/quality_validator.rb', line 11

def options
  @options
end

#original_messageObject (readonly)

Returns the value of attribute original_message.



11
12
13
# File 'lib/hyrum/validators/quality_validator.rb', line 11

def original_message
  @original_message
end

Instance Method Details

#validateObject



19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
# File 'lib/hyrum/validators/quality_validator.rb', line 19

def validate
  return empty_result if messages.empty?

  all_variations = messages.values.flatten
  return single_variation_result if all_variations.size <= 1

  semantic_score = calculate_semantic_similarity(all_variations)
  lexical_score = calculate_lexical_diversity(all_variations)

  overall_score = (semantic_score * SIMILARITY_WEIGHT) + (lexical_score * DIVERSITY_WEIGHT)

  passed = overall_score >= options[:min_quality] &&
           lexical_score >= MIN_DIVERSITY_THRESHOLD &&
           semantic_score >= MIN_SIMILARITY_THRESHOLD

  warnings = build_warnings(semantic_score, lexical_score)

  ValidationResult.new(
    score: overall_score.round(2),
    semantic_similarity: semantic_score.round(2),
    lexical_diversity: lexical_score.round(2),
    passed: passed,
    details: {
      min_quality_threshold: options[:min_quality],
      variation_count: all_variations.size
    },
    warnings: warnings
  )
end