Sha256: c26b29d5c357d377a09722776ee2e8e82135141c03dafd827934d8979c711ff6
Contents?: true
Size: 1.52 KB
Versions: 11
Compression:
Stored size: 1.52 KB
Contents
# frozen_string_literal: true
module Spandx
module Core
class Content
attr_reader :raw
def initialize(raw)
@raw = raw
end
def tokens
@tokens ||= tokenize(canonicalize(raw)).to_set
end
def similar?(other, algorithm: :dice_coefficient)
case algorithm
when :dice_coefficient
similarity_score(other, algorithm: algorithm) > 89.0
when :levenshtein
similarity_score(other, algorithm: algorithm) < 3
when :jaro_winkler
similarity_score(other, algorithm: algorithm) > 89.0
end
end
def similarity_score(other, algorithm: :dice_coefficient)
case algorithm
when :dice_coefficient
dice_coefficient(other)
when :levenshtein
require 'text'
Text::Levenshtein.distance(raw, other.raw, 100)
when :jaro_winkler
require 'jaro_winkler'
JaroWinkler.distance(raw, other.raw) * 100.0
end
end
private
def canonicalize(content)
content&.downcase
end
def tokenize(content)
content.to_s.scan(/[a-zA-Z]+/)
end
def blank?(content)
content.nil? || content.chomp.strip.empty?
end
# https://en.wikibooks.org/wiki/Algorithm_Implementation/Strings/Dice%27s_coefficient#Ruby
def dice_coefficient(other)
overlap = (tokens & other.tokens).size
total = tokens.size + other.tokens.size
100.0 * (overlap * 2.0 / total)
end
end
end
end
Version data entries
11 entries across 11 versions & 1 rubygems