Class: Licensee::Matchers::Dice

Inherits:
Object
  • Object
show all
Defined in:
lib/licensee/matchers/dice_matcher.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(file) ⇒ Dice

Returns a new instance of Dice.



6
7
8
 
# File 'lib/licensee/matchers/dice_matcher.rb', line 6

def initialize(file)
  @file = file
end

Instance Attribute Details

#file ⇒ Object (readonly)

Returns the value of attribute file.



4
5
6
 
# File 'lib/licensee/matchers/dice_matcher.rb', line 4

def file
  @file
end

Instance Method Details

#confidence ⇒ Object

Confidence that the matched license is a match



55
56
57
 
# File 'lib/licensee/matchers/dice_matcher.rb', line 55

def confidence
  @confidence ||= match ? file.similarity(match) : 0
end

#licenses_by_similiarity ⇒ Object 



39
40
41
42
43
44
45
46
 
# File 'lib/licensee/matchers/dice_matcher.rb', line 39

def licenses_by_similiarity
  @licenses_by_similiarity ||= begin
    licenses = potential_licenses.map do |license|
      [license, license.similarity(file)]
    end
    licenses.sort_by { |_, similarity| similarity }.reverse
  end
end

#match ⇒ Object

Return the first potential license that is more similar than the confidence threshold



12
13
14
15
16
17
18
 
# File 'lib/licensee/matchers/dice_matcher.rb', line 12

def match
  @match ||= if matches.empty?
    nil
  else
    matches.first[0]
  end
end

#matches ⇒ Object 



48
49
50
51
52
 
# File 'lib/licensee/matchers/dice_matcher.rb', line 48

def matches
  @matches ||= licenses_by_similiarity.select do |_, similarity|
    similarity >= Licensee.confidence_threshold
  end
end

#potential_licenses ⇒ Object

Licenses that may be a match for this file. To avoid false positives:

  1. Creative commons licenses cannot be matched against license files that begin with the title of a non-open source CC license variant

  2. The percentage change in file length may not exceed the inverse of the confidence threshold



27
28
29
30
31
32
33
34
35
36
37
 
# File 'lib/licensee/matchers/dice_matcher.rb', line 27

def potential_licenses
  @potential_licenses ||= begin
    Licensee.licenses(hidden: true).select do |license|
      if license.creative_commons? && file.potential_false_positive?
        false
      else
        license.wordset && license.length_delta(file) <= license.max_delta
      end
    end
  end
end