RubyGems - kebab - Versions diffs - 1.0.2 - Mend

kebab 1.0.2

Files changed (55) hide show

checksums.yaml +7 -0
data/.gemtest +0 -0
data/Changelog.md +99 -0
data/MIT-LICENSE +19 -0
data/README.md +26 -0
data/Rakefile +34 -0
data/lib/kebab.rb +18 -0
data/lib/kebab/identifier.rb +294 -0
data/lib/kebab/transliterator/base.rb +110 -0
data/lib/kebab/transliterator/bulgarian.rb +27 -0
data/lib/kebab/transliterator/cyrillic.rb +108 -0
data/lib/kebab/transliterator/danish.rb +15 -0
data/lib/kebab/transliterator/german.rb +15 -0
data/lib/kebab/transliterator/greek.rb +77 -0
data/lib/kebab/transliterator/hindi.rb +137 -0
data/lib/kebab/transliterator/latin.rb +199 -0
data/lib/kebab/transliterator/macedonian.rb +29 -0
data/lib/kebab/transliterator/norwegian.rb +14 -0
data/lib/kebab/transliterator/romanian.rb +13 -0
data/lib/kebab/transliterator/russian.rb +22 -0
data/lib/kebab/transliterator/serbian.rb +34 -0
data/lib/kebab/transliterator/spanish.rb +9 -0
data/lib/kebab/transliterator/swedish.rb +16 -0
data/lib/kebab/transliterator/turkish.rb +8 -0
data/lib/kebab/transliterator/ukrainian.rb +30 -0
data/lib/kebab/transliterator/vietnamese.rb +143 -0
data/lib/kebab/utf8/active_support_proxy.rb +26 -0
data/lib/kebab/utf8/dumb_proxy.rb +49 -0
data/lib/kebab/utf8/java_proxy.rb +22 -0
data/lib/kebab/utf8/mappings.rb +193 -0
data/lib/kebab/utf8/proxy.rb +125 -0
data/lib/kebab/utf8/unicode_proxy.rb +23 -0
data/lib/kebab/version.rb +5 -0
data/spec/kebab_spec.rb +155 -0
data/spec/spec_helper.rb +45 -0
data/spec/transliterators/base_spec.rb +16 -0
data/spec/transliterators/bulgarian_spec.rb +20 -0
data/spec/transliterators/danish_spec.rb +17 -0
data/spec/transliterators/german_spec.rb +17 -0
data/spec/transliterators/greek_spec.rb +17 -0
data/spec/transliterators/hindi_spec.rb +17 -0
data/spec/transliterators/latin_spec.rb +9 -0
data/spec/transliterators/macedonian_spec.rb +9 -0
data/spec/transliterators/norwegian_spec.rb +18 -0
data/spec/transliterators/polish_spec.rb +14 -0
data/spec/transliterators/romanian_spec.rb +19 -0
data/spec/transliterators/russian_spec.rb +9 -0
data/spec/transliterators/serbian_spec.rb +25 -0
data/spec/transliterators/spanish_spec.rb +13 -0
data/spec/transliterators/swedish_spec.rb +18 -0
data/spec/transliterators/turkish_spec.rb +24 -0
data/spec/transliterators/ukrainian_spec.rb +88 -0
data/spec/transliterators/vietnamese_spec.rb +18 -0
data/spec/utf8_proxy_spec.rb +53 -0
metadata +167 -0

data/lib/kebab/transliterator/latin.rb ADDED

@@ -0,0 +1,199 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Latin < Base
+      APPROXIMATIONS = {
+        "À" => "A",
+        "Á" => "A",
+        "Â" => "A",
+        "Ã" => "A",
+        "Ä" => "A",
+        "Å" => "A",
+        "Æ" => "Ae",
+        "Ç" => "C",
+        "È" => "E",
+        "É" => "E",
+        "Ê" => "E",
+        "Ë" => "E",
+        "Ì" => "I",
+        "Í" => "I",
+        "Î" => "I",
+        "Ï" => "I",
+        "Ð" => "D",
+        "Ñ" => "N",
+        "Ò" => "O",
+        "Ó" => "O",
+        "Ô" => "O",
+        "Õ" => "O",
+        "Ö" => "O",
+        "Ø" => "O",
+        "Ù" => "U",
+        "Ú" => "U",
+        "Û" => "U",
+        "Ü" => "U",
+        "Ý" => "Y",
+        "Þ" => "Th",
+        "ß" => "ss",
+        "à" => "a" ,
+        "á" => "a",
+        "â" => "a",
+        "ã" => "a",
+        "ä" => "a",
+        "å" => "a",
+        "æ" => "ae",
+        "ç" => "c" ,
+        "è" => "e",
+        "é" => "e",
+        "ê" => "e",
+        "ë" => "e",
+        "ì" => "i",
+        "í" => "i",
+        "î" => "i",
+        "ï" => "i",
+        "ð" => "d",
+        "ñ" => "n",
+        "ò" => "o",
+        "ó" => "o",
+        "ô" => "o",
+        "õ" => "o",
+        "ö" => "o",
+        "ø" => "o",
+        "ù" => "u",
+        "ú" => "u",
+        "û" => "u",
+        "ü" => "u",
+        "ý" => "y",
+        "þ" => "th",
+        "ÿ" => "y",
+        "Ā" => "A",
+        "Ă" => "A",
+        "Ą" => "A",
+        "Ć" => "C",
+        "Ĉ" => "C",
+        "Ċ" => "C",
+        "Č" => "C",
+        "Ď" => "D",
+        "Đ" => "D",
+        "Ē" => "E",
+        "Ĕ" => "E",
+        "Ė" => "E",
+        "Ę" => "E",
+        "Ě" => "E",
+        "Ĝ" => "G",
+        "Ğ" => "G",
+        "Ġ" => "G",
+        "Ģ" => "G",
+        "Ĥ" => "H",
+        "Ħ" => "H",
+        "Ĩ" => "I",
+        "Ī" => "I",
+        "Ĭ" => "I",
+        "Į" => "I",
+        "İ" => "I",
+        "Ĳ" => "Ij",
+        "Ĵ" => "J",
+        "Ķ" => "K",
+        "Ĺ" => "L",
+        "Ļ" => "L",
+        "Ľ" => "L",
+        "Ŀ" => "L",
+        "Ł" => "L",
+        "Ń" => "N",
+        "Ņ" => "N",
+        "Ň" => "N",
+        "Ŋ" => "Ng",
+        "Ō" => "O",
+        "Ŏ" => "O",
+        "Ő" => "O",
+        "Œ" => "OE",
+        "Ŕ" => "R",
+        "Ŗ" => "R",
+        "Ř" => "R",
+        "Ś" => "S",
+        "Ŝ" => "S",
+        "Ş" => "S",
+        "Š" => "S",
+        "Ţ" => "T",
+        "Ť" => "T",
+        "Ŧ" => "T",
+        "Ũ" => "U",
+        "Ū" => "U",
+        "Ŭ" => "U",
+        "Ů" => "U",
+        "Ű" => "U",
+        "Ų" => "U",
+        "Ŵ" => "W",
+        "Ŷ" => "Y",
+        "Ÿ" => "Y",
+        "Ź" => "Z",
+        "Ż" => "Z",
+        "Ž" => "Z",
+        "ā" => "a",
+        "ă" => "a",
+        "ą" => "a",
+        "ć" => "c",
+        "ĉ" => "c",
+        "ċ" => "c",
+        "č" => "c",
+        "ď" => "d",
+        "đ" => "d",
+        "ē" => "e",
+        "ĕ" => "e",
+        "ė" => "e",
+        "ę" => "e",
+        "ě" => "e",
+        "ĝ" => "g",
+        "ğ" => "g",
+        "ġ" => "g",
+        "ģ" => "g",
+        "ĥ" => "h",
+        "ħ" => "h",
+        "ĩ" => "i",
+        "ī" => "i",
+        "ĭ" => "i",
+        "į" => "i",
+        "ı" => "i",
+        "ĳ" => "ij",
+        "ĵ" => "j",
+        "ķ" => "k",
+        "ĸ" => "k",
+        "ĺ" => "l",
+        "ļ" => "l",
+        "ľ" => "l",
+        "ŀ" => "l",
+        "ł" => "l",
+        "ń" => "n",
+        "ņ" => "n",
+        "ň" => "n",
+        "ŉ" => "n",
+        "ŋ" => "ng",
+        "ō" => "o",
+        "ŏ" => "o",
+        "ő" => "o",
+        "œ" => "oe",
+        "ŕ" => "r",
+        "ŗ" => "r",
+        "ř" => "r",
+        "ś" => "s",
+        "ŝ" => "s",
+        "ş" => "s",
+        "š" => "s",
+        "ţ" => "t",
+        "ť" => "t",
+        "ŧ" => "t",
+        "ũ" => "u",
+        "ū" => "u",
+        "ŭ" => "u",
+        "ů" => "u",
+        "ű" => "u",
+        "ų" => "u",
+        "ŵ" => "w",
+        "ŷ" => "y",
+        "ž" => "z",
+        "ź" => "z",
+        "ż" => "z"
+      }
+    end
+  end
+end

data/lib/kebab/transliterator/macedonian.rb ADDED

@@ -0,0 +1,29 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Macedonian < Cyrillic
+      APPROXIMATIONS = {
+        "Ѓ" => "Gj",
+        "Љ" => "Lj",
+        "Њ" => "Nj",
+        "Ќ" => "Kj",
+        "Џ" => "Dzh",
+        "Ж" => "Zh",
+        "Ц" => "C",
+        "Ѕ" => "Z",
+        "Ј" => "J",
+        "Х" => "H",
+        "ѓ" => "gj",
+        "љ" => "lj",
+        "њ" => "nj",
+        "ќ" => "kj",
+        "џ" => "dzh",
+        "ж" => "zh",
+        "ц" => "c",
+        "ѕ" => "z",
+        "ј" => "j",
+        "х" => "h"
+      }
+    end
+  end
+end

data/lib/kebab/transliterator/norwegian.rb ADDED

@@ -0,0 +1,14 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Norwegian < Latin
+      APPROXIMATIONS = {
+        "ø" => "oe",
+        "å" => "aa",
+        "Ø" => "Oe",
+        "Å" => "Aa"
+      }
+    end
+  end
+end

data/lib/kebab/transliterator/romanian.rb ADDED

@@ -0,0 +1,13 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Romanian < Latin
+      APPROXIMATIONS = {
+        "ș" => "s",
+        "ț" => "t",
+        "Ș" => "S",
+        "Ț" => "T"
+      }
+    end
+  end
+end

data/lib/kebab/transliterator/russian.rb ADDED

@@ -0,0 +1,22 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Russian < Cyrillic
+      APPROXIMATIONS = {
+        "Й" => "I",
+        "М" => "M",
+        "Х" => "H",
+        "Ц" => "Ts",
+        "Ш" => "Sh",
+        "Щ" => "Sch",
+        "Ю" => "U",
+        "Я" => "Ya",
+        "й" => "i",
+        "х" => "h",
+        "ц" => "ts",
+        "щ" => "sch",
+        "ю" => "u"
+      }
+    end
+  end
+end

data/lib/kebab/transliterator/serbian.rb ADDED

@@ -0,0 +1,34 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Serbian < Latin
+      APPROXIMATIONS = Cyrillic.const_get(:APPROXIMATIONS).merge({
+        "Ð" => "Dj",
+        "Č" => "Ch",
+        "Š" => "Sh",
+        "č" => "ch",
+        "đ" => "dj",
+        "š" => "sh",
+        "Ћ" => "C",
+        "Ц" => "C",
+        "Ч" => "Ch",
+        "Ђ" => "Dj",
+        "Џ" => "Dz",
+        "Х" => "H",
+        "Ј" => "J",
+        "Љ" => "Lj",
+        "Њ" => "Nj",
+        "ц" => "c",
+        "ћ" => "c",
+        "ч" => "ch",
+        "ђ" => "dj",
+        "џ" => "dz",
+        "х" => "h",
+        "ј" => "j",
+        "љ" => "lj",
+        "њ" => "nj"
+      })
+    end
+  end
+end

data/lib/kebab/transliterator/spanish.rb ADDED

@@ -0,0 +1,9 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Spanish < Latin
+      APPROXIMATIONS = {"ñ" => "ni", "Ñ" => "Ni"}
+    end
+  end
+end

data/lib/kebab/transliterator/swedish.rb ADDED

@@ -0,0 +1,16 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Swedish < Latin
+      APPROXIMATIONS = {
+        "å" => "aa",
+        "ä" => "ae",
+        "ö" => "oe",
+        "Å" => "Aa",
+        "Ä" => "Ae",
+        "Ö" => "Oe"
+      }
+    end
+  end
+end

data/lib/kebab/transliterator/turkish.rb ADDED

@@ -0,0 +1,8 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Turkish < Latin
+    end
+  end
+end

data/lib/kebab/transliterator/ukrainian.rb ADDED

@@ -0,0 +1,30 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Ukrainian < Cyrillic
+      APPROXIMATIONS = {
+        "Г" => "H",
+        "г" => "h",
+        "Ґ" => "G",
+        "ґ" => "g",
+        "є" => "ie",
+        "И" => "Y",
+        "и" => "y",
+        "І" => "I",
+        "і" => "i",
+        "ї" => "i",
+        "Й" => "Y",
+        "й" => "i",
+        "Х" => "Kh",
+        "х" => "kh",
+        "Ц" => "Ts",
+        "ц" => 'ts',
+        "Щ" => "Shch",
+        "щ" => "shch",
+        "ю" => "iu",
+        "я" => "ia",
+        "'" => ""
+      }
+    end
+  end
+end

data/lib/kebab/transliterator/vietnamese.rb ADDED

@@ -0,0 +1,143 @@
+# encoding: utf-8
+module Kebab
+  module Transliterator
+    class Vietnamese < Latin
+      APPROXIMATIONS = {
+        "à" => "a",
+        "á" => "a",
+        "ạ" => "a",
+        "ả" => "a",
+        "ã" => "a",
+        "â" => "a",
+        "ầ" => "a",
+        "ấ" => "a",
+        "ậ" => "a",
+        "ẩ" => "a",
+        "ẫ" => "a",
+        "ă" => "a",
+        "ằ" => "a",
+        "ắ" => "a",
+        "ặ" => "a",
+        "ẳ" => "a",
+        "ẵ" => "a",
+        "À" => "A",
+        "Á" => "A",
+        "Ạ" => "A",
+        "Ả" => "A",
+        "Ã" => "A",
+        "Â" => "A",
+        "Ầ" => "A",
+        "Ấ" => "A",
+        "Ậ" => "A",
+        "Ẩ" => "A",
+        "Ẫ" => "A",
+        "Ă" => "A",
+        "Ằ" => "A",
+        "Ắ" => "A",
+        "Ặ" => "A",
+        "Ẳ" => "A",
+        "Ẵ" => "A",
+        "ì" => "i",
+        "í" => "i",
+        "ị" => "i",
+        "ỉ" => "i",
+        "ĩ" => "i",
+        "Ì" => "I",
+        "Í" => "I",
+        "Ị" => "I",
+        "Ỉ" => "I",
+        "Ĩ" => "I",
+        "ù" => "u",
+        "ú" => "u",
+        "ụ" => "u",
+        "ủ" => "u",
+        "ũ" => "u",
+        "ư" => "u",
+        "ừ" => "u",
+        "ứ" => "u",
+        "ự" => "u",
+        "ử" => "u",
+        "ữ" => "u",
+        "Ù" => "U",
+        "Ú" => "U",
+        "Ụ" => "U",
+        "Ủ" => "U",
+        "Ũ" => "U",
+        "Ư" => "U",
+        "Ừ" => "U",
+        "Ứ" => "U",
+        "Ự" => "U",
+        "Ử" => "U",
+        "Ữ" => "U",
+        "è" => "e",
+        "é" => "e",
+        "ẹ" => "e",
+        "ẻ" => "e",
+        "ẽ" => "e",
+        "ê" => "e",
+        "ề" => "e",
+        "ế" => "e",
+        "ệ" => "e",
+        "ể" => "e",
+        "ễ" => "e",
+        "È" => "E",
+        "É" => "E",
+        "Ẹ" => "E",
+        "Ẻ" => "E",
+        "Ẽ" => "E",
+        "Ê" => "E",
+        "Ề" => "E",
+        "Ế" => "E",
+        "Ệ" => "E",
+        "Ể" => "E",
+        "Ễ" => "E",
+        "ò" => "o",
+        "ó" => "o",
+        "ọ" => "o",
+        "ỏ" => "o",
+        "õ" => "o",
+        "ô" => "o",
+        "ồ" => "o",
+        "ố" => "o",
+        "ộ" => "o",
+        "ổ" => "o",
+        "ỗ" => "o",
+        "ơ" => "o",
+        "ờ" => "o",
+        "ớ" => "o",
+        "ợ" => "o",
+        "ở" => "o",
+        "ỡ" => "o",
+        "Ò" => "O",
+        "Ó" => "O",
+        "Ọ" => "O",
+        "Ỏ" => "O",
+        "Õ" => "O",
+        "Ô" => "O",
+        "Ồ" => "O",
+        "Ố" => "O",
+        "Ộ" => "O",
+        "Ổ" => "O",
+        "Ỗ" => "O",
+        "Ơ" => "O",
+        "Ờ" => "O",
+        "Ớ" => "O",
+        "Ợ" => "O",
+        "Ở" => "O",
+        "Ỡ" => "O",
+        "ỳ" => "y",
+        "ý" => "y",
+        "ỵ" => "y",
+        "ỷ" => "y",
+        "ỹ" => "y",
+        "Ỳ" => "Y",
+        "Ý" => "Y",
+        "Ỵ" => "Y",
+        "Ỷ" => "Y",
+        "Ỹ" => "Y",
+        "đ" => "d",
+        "Đ" => "D"
+      }
+    end
+  end
+end