babosa 1.0.2 → 2.0.0

data/lib/babosa/transliterator/base.rb

@@ -1,29 +1,11 @@
-# encoding: utf-8
+# frozen_string_literal: true
 
-require 'singleton'
+require "singleton"
 
 module Babosa
-
   module Transliterator
-
-    autoload :Bulgarian,  "babosa/transliterator/bulgarian"
-    autoload :Cyrillic,   "babosa/transliterator/cyrillic"
-    autoload :Danish,     "babosa/transliterator/danish"
-    autoload :German,     "babosa/transliterator/german"
-    autoload :Latin,      "babosa/transliterator/latin"
-    autoload :Macedonian, "babosa/transliterator/macedonian"
-    autoload :Norwegian,  "babosa/transliterator/norwegian"
-    autoload :Romanian,   "babosa/transliterator/romanian"
-    autoload :Russian,    "babosa/transliterator/russian"
-    autoload :Serbian,    "babosa/transliterator/serbian"
-    autoload :Spanish,    "babosa/transliterator/spanish"
-    autoload :Swedish,    "babosa/transliterator/swedish"
-    autoload :Ukrainian,  "babosa/transliterator/ukrainian"
-    autoload :Greek,      "babosa/transliterator/greek"
-    autoload :Vietnamese, "babosa/transliterator/vietnamese"
-
     def self.get(symbol)
-      class_name = symbol.to_s.split("_").map {|a| a.gsub(/\b('?[a-z])/) { $1.upcase }}.join
+      class_name = symbol.to_s.split("_").map { |a| a.gsub(/\b('?[a-z])/) { Regexp.last_match(1).upcase } }.join
       const_get(class_name)
     end
 
@@ -45,36 +27,39 @@ module Babosa
         "”" => '"',
         "„" => '"',
         "‟" => '"',
-        '’' => "'",
-        '，' => ",",
-        '。' => ".",
-        '！' => "!",
-        '？' => '?',
-        '、' => ',',
-        '（' => '(',
-        '）' => ')',
-        '【' => '[',
-        '】' => ']',
-        '；' => ';',
-        '：' => ':',
-        '《' => '<',
-        '》' => '>',
-        # various kinds of space characters
-        "\xc2\xa0"     => " ",
-        "\xe2\x80\x80" => " ",
-        "\xe2\x80\x81" => " ",
-        "\xe2\x80\x82" => " ",
-        "\xe2\x80\x83" => " ",
-        "\xe2\x80\x84" => " ",
-        "\xe2\x80\x85" => " ",
-        "\xe2\x80\x86" => " ",
-        "\xe2\x80\x87" => " ",
-        "\xe2\x80\x88" => " ",
-        "\xe2\x80\x89" => " ",
-        "\xe2\x80\x8a" => " ",
-        "\xe2\x81\x9f" => " ",
-        "\xe3\x80\x80" => " ",
-      }.freeze
+        "’" => "'",
+        "，" => ",",
+        "。" => ".",
+        "！" => "!",
+        "？" => "?",
+        "、" => ",",
+        "（" => "(",
+        "）" => ")",
+        "【" => "[",
+        "】" => "]",
+        "；" => ";",
+        "：" => ":",
+        "《" => "<",
+        "》" => ">"
+      }.merge(
+        {
+          # various kinds of space characters
+          "\xc2\xa0"     => " ",
+          "\xe2\x80\x80" => " ",
+          "\xe2\x80\x81" => " ",
+          "\xe2\x80\x82" => " ",
+          "\xe2\x80\x83" => " ",
+          "\xe2\x80\x84" => " ",
+          "\xe2\x80\x85" => " ",
+          "\xe2\x80\x86" => " ",
+          "\xe2\x80\x87" => " ",
+          "\xe2\x80\x88" => " ",
+          "\xe2\x80\x89" => " ",
+          "\xe2\x80\x8a" => " ",
+          "\xe2\x81\x9f" => " ",
+          "\xe3\x80\x80" => " "
+        }
+      ).freeze
 
       attr_reader :approximations
 
@@ -85,8 +70,8 @@ module Babosa
           @approximations = {}
         end
         self.class.const_get(:APPROXIMATIONS).inject(@approximations) do |memo, object|
-          index       = object[0].unpack("U").shift
-          value       = object[1].unpack("C*")
+          index       = object[0].codepoints.shift
+          value       = object[1].codepoints
           memo[index] = value.length == 1 ? value[0] : value
           memo
         end
@@ -101,8 +86,26 @@ module Babosa
 
       # Transliterates a string.
       def transliterate(string)
-        string.unpack("U*").map {|char| self[char] || char}.flatten.pack("U*")
+        string.codepoints.map { |char| self[char] || char }.flatten.pack("U*")
       end
     end
   end
 end
+
+require "babosa/transliterator/cyrillic"
+require "babosa/transliterator/latin"
+require "babosa/transliterator/bulgarian"
+require "babosa/transliterator/danish"
+require "babosa/transliterator/german"
+require "babosa/transliterator/hindi"
+require "babosa/transliterator/macedonian"
+require "babosa/transliterator/norwegian"
+require "babosa/transliterator/romanian"
+require "babosa/transliterator/russian"
+require "babosa/transliterator/serbian"
+require "babosa/transliterator/spanish"
+require "babosa/transliterator/swedish"
+require "babosa/transliterator/ukrainian"
+require "babosa/transliterator/greek"
+require "babosa/transliterator/vietnamese"
+require "babosa/transliterator/turkish"

data/lib/babosa/transliterator/bulgarian.rb

@@ -1,4 +1,5 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class Bulgarian < Cyrillic
@@ -21,7 +22,7 @@ module Babosa
         "ь" => "i",
         "ю" => "iu",
         "я" => "ia"
-      }
+      }.freeze
     end
   end
 end

data/lib/babosa/transliterator/cyrillic.rb

@@ -1,7 +1,7 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
-
     # Approximations are based on GOST 7.79, System B:
     # http://en.wikipedia.org/wiki/ISO_9#GOST_7.79
     class Cyrillic < Base
@@ -97,11 +97,11 @@ module Babosa
         "Ѵ" => "Yh",
         "ѵ" => "yh",
         "Ґ" => "G",
-        "ґ" => "g",
-      }
+        "ґ" => "g"
+      }.freeze
 
       def transliterate(string)
-        super.gsub(/(c)z([ieyj])/) { "#{$1}#{$2}" }
+        super.gsub(/(c)z([ieyj])/) { "#{Regexp.last_match(1)}#{Regexp.last_match(2)}" }
       end
     end
   end

data/lib/babosa/transliterator/danish.rb

@@ -1,4 +1,5 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class Danish < Latin
@@ -8,8 +9,7 @@ module Babosa
         "å" => "aa",
         "Ø" => "Oe",
         "Å" => "Aa"
-      }
+      }.freeze
     end
   end
 end
-

data/lib/babosa/transliterator/german.rb

@@ -1,4 +1,5 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class German < Latin
@@ -9,7 +10,7 @@ module Babosa
         "Ä" => "Ae",
         "Ö" => "Oe",
         "Ü" => "Ue"
-      }
+      }.freeze
     end
   end
 end

data/lib/babosa/transliterator/greek.rb

@@ -1,4 +1,5 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class Greek < Base
@@ -71,7 +72,7 @@ module Babosa
         "Ώ" => "O",
         "ω" => "o",
         "ώ" => "o"
-      }
+      }.freeze
     end
   end
-end
+end

data/lib/babosa/transliterator/hindi.rb

@@ -0,0 +1,138 @@
+# frozen_string_literal: true
+
+module Babosa
+  module Transliterator
+    class Hindi < Base
+      APPROXIMATIONS = {
+        "ऀ" => "n",
+        "ँ" => "n",
+        "ं" => "n",
+        "ः" => "h",
+        "ऄ" => "a",
+        "अ" => "a",
+        "आ" => "aa",
+        "इ" => "i",
+        "ई" => "ii",
+        "उ" => "u",
+        "ऊ" => "uu",
+        "ऋ" => "ri",
+        "ऌ" => "lri",
+        "ऍ" => "e",
+        "ऎ" => "e",
+        "ए" => "e",
+        "ऐ" => "ei",
+        "ऑ" => "o",
+        "ऒ" => "o",
+        "ओ" => "o",
+        "औ" => "ou",
+        "क" => "k",
+        "ख" => "kh",
+        "ग" => "g",
+        "घ" => "gh",
+        "ङ" => "d",
+        "च" => "ch",
+        "छ" => "chh",
+        "ज" => "j",
+        "झ" => "jh",
+        "ञ" => "ny",
+        "ट" => "tt",
+        "ठ" => "tth",
+        "ड" => "dd",
+        "ढ" => "ddh",
+        "ण" => "nn",
+        "त" => "t",
+        "थ" => "th",
+        "द" => "d",
+        "ध" => "dh",
+        "न" => "n",
+        "ऩ" => "nnn",
+        "प" => "p",
+        "फ" => "ph",
+        "ब" => "b",
+        "भ" => "bh",
+        "म" => "m",
+        "य" => "y",
+        "र" => "r",
+        "ऱ" => "rr",
+        "ल" => "l",
+        "ळ" => "ll",
+        "ऴ" => "ll",
+        "व" => "v",
+        "श" => "sh",
+        "ष" => "ss",
+        "स" => "s",
+        "ह" => "h",
+        "ऺ" => "oe",
+        "ऻ" => "ooe",
+        "़" => "",
+        "ऽ" => "-",
+        "ा" => "aa",
+        "ि" => "i",
+        "ी" => "ii",
+        "ु" => "u",
+        "ू" => "uu",
+        "ृ" => "r",
+        "ॄ" => "rr",
+        "ॅ" => "e",
+        "ॆ" => "e",
+        "े" => "e",
+        "ै" => "ai",
+        "ॉ" => "o",
+        "ॊ" => "o",
+        "ो" => "o",
+        "ौ" => "au",
+        "्" => "",
+        "ॎ" => "e",
+        "ॏ" => "aw",
+        "ॐ" => "om",
+        "॑" => "",
+        "॒" => "_",
+        "॓" => "",
+        "॔" => "",
+        "ॕ" => "ee",
+        "ॖ" => "ue",
+        "ॗ" => "uue",
+        "क़" => "q",
+        "ख़" => "khh",
+        "ग़" => "ghh",
+        "ज़" => "za",
+        "ड़" => "dddh",
+        "ढ़" => "rh",
+        "फ़" => "f",
+        "य़" => "yy",
+        "ॠ" => "rri",
+        "ॡ" => "lr",
+        "ॢ" => "l",
+        "ॣ" => "l",
+        "।" => ".",
+        "॥" => "..",
+        "०" => "0",
+        "१" => "1",
+        "२" => "2",
+        "३" => "3",
+        "४" => "4",
+        "५" => "5",
+        "६" => "6",
+        "७" => "7",
+        "८" => "8",
+        "९" => "9",
+        "॰" => ".",
+        "ॱ" => ".",
+        "ॲ" => "a",
+        "ॳ" => "oe",
+        "ॴ" => "ooe",
+        "ॵ" => "aw",
+        "ॶ" => "ue",
+        "ॷ" => "uue",
+        "ॸ" => "dd",
+        "ॹ" => "zh",
+        "ॺ" => "y",
+        "ॻ" => "gg",
+        "ॼ" => "jj",
+        "ॽ" => "?",
+        "ॾ" => "ddd",
+        "ॿ" => "bb"
+      }.freeze
+    end
+  end
+end

data/lib/babosa/transliterator/latin.rb

@@ -1,8 +1,8 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class Latin < Base
-
       APPROXIMATIONS = {
         "À" => "A",
         "Á" => "A",
@@ -35,14 +35,14 @@ module Babosa
         "Ý" => "Y",
         "Þ" => "Th",
         "ß" => "ss",
-        "à" => "a" ,
+        "à" => "a",
         "á" => "a",
         "â" => "a",
         "ã" => "a",
         "ä" => "a",
         "å" => "a",
         "æ" => "ae",
-        "ç" => "c" ,
+        "ç" => "c",
         "è" => "e",
         "é" => "e",
         "ê" => "e",
@@ -193,7 +193,7 @@ module Babosa
         "ž" => "z",
         "ź" => "z",
         "ż" => "z"
-      }
+      }.freeze
     end
   end
 end

data/lib/babosa/transliterator/macedonian.rb

@@ -1,4 +1,5 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class Macedonian < Cyrillic
@@ -23,7 +24,7 @@ module Babosa
         "ѕ" => "z",
         "ј" => "j",
         "х" => "h"
-      }
+      }.freeze
     end
   end
 end

data/lib/babosa/transliterator/norwegian.rb

@@ -1,4 +1,5 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class Norwegian < Latin
@@ -7,8 +8,7 @@ module Babosa
         "å" => "aa",
         "Ø" => "Oe",
         "Å" => "Aa"
-      }
+      }.freeze
     end
   end
 end
-

data/lib/babosa/transliterator/romanian.rb

@@ -1,4 +1,5 @@
-# encoding: utf-8
+# frozen_string_literal: true
+
 module Babosa
   module Transliterator
     class Romanian < Latin
@@ -7,7 +8,7 @@ module Babosa
         "ț" => "t",
         "Ș" => "S",
         "Ț" => "T"
-      }
+      }.freeze
     end
   end
 end