Use jaro winkler to analyze brand name data in OSM

gistfile1.txt

#!/usr/local/opt/ruby/bin/ruby
require 'json'
require 'jaro_winkler'

# parse out all the brands files
files = Dir.glob('brands/**/*.json')
brand_data = {}
files.each { |f| brand_data.merge!(JSON.parse(File.read(f))) }
brands = brand_data.keys

# parse out all known names
name_data = JSON.parse(File.read('dist/names_all.json'))

# iterate over all the brands and compare them to all the names then print if > .95 score
brands.each do |b|
  b_type,b_name = b.split('|')
  name_data.keys.each do |n|
    n_type,n_name = n.split('|')

    # don't compare the data to itself or if the entry is in the nomatch field of the brands data
    next if n_name == b_name
    next if brand_data[b].key?('nomatch') && brand_data[b]['nomatch'].include?(n)

    score = JaroWinkler.distance(b_name, n_name, ignore_case: true)
    puts "#{b} -> #{n}: #{name_data[n]} at #{score}" if score > 0.95
  end
end