lib/fuzzily/trigram.rb in fuzzily-0.0.1 vs lib/fuzzily/trigram.rb in fuzzily-0.0.2

@@ -1,25 +1,23 @@
-require 'iconv'
+require 'active_support/core_ext/string/multibyte'
 
 module Fuzzily
   module String
     def trigrams
-      normalized_words.map do |word|
-        (0..(word.length - 3)).map { |index| word[index,3] }
-      end.flatten.uniq
+      normalized = self.normalize
+      (0..(normalized.length - 3)).map { |index| normalized[index,3] }.uniq
     end
 
-    private
+    protected
 
     # Remove accents, downcase, replace spaces and word start with '*',
     # return list of normalized words
-    def normalized_words
-      self.split(/\s+/).map { |word|
-        Iconv.iconv('ascii//translit//ignore', 'utf-8', word).first.downcase.gsub(/\W/,'')
-      }.
-      delete_if(&:empty?).
-      map { |word|
-        "**#{word}"
-      }
+    def normalize
+      # Iconv.iconv('ascii//translit//ignore', 'utf-8', self).first.
+      ActiveSupport::Multibyte::Chars.new(self).
+        mb_chars.normalize(:kd).gsub(/[^\x00-\x7F]/,'').downcase.to_s.
+        gsub(/\W/,' ').
+        gsub(/\s+/,'*').
+        gsub(/^/,'**')
     end
   end
 end