i18n_data 0.14.0 → 0.17.0

data/cache/file_data_provider/languages-ZH_CN.txt

@@ -24,7 +24,7 @@ CA;;加泰罗尼亚语
 CS;;捷克语
 CH;;查莫罗语
 CE;;车臣语
-CU;;Church Slavic; Old Slavonic; Church Slavonic; Old Bulgarian; Old Church Slavonic
+CU;;古教会斯拉夫语
 CV;;楚瓦什语
 KW;;康沃尔语
 CO;;科西嘉语

data/cache/file_data_provider/languages-ZH_HK.txt

@@ -13,7 +13,7 @@ AZ;;阿塞拜疆語
 BA;;巴什基爾語
 BM;;班巴拉語
 BE;;白俄羅斯語
-BN;;孟加拉語
+BN;;孟加拉文
 BH;;比哈爾語系
 BI;;比斯拉瑪語
 BO;;藏語
@@ -24,7 +24,7 @@ CA;;加泰羅尼亞語
 CS;;捷克語
 CH;;查莫羅語
 CE;;車臣語
-CU;;Church Slavic; Old Slavonic; Church Slavonic; Old Bulgarian; Old Church Slavonic
+CU;;古教會斯拉夫語
 CV;;楚瓦什語
 KW;;康沃爾語
 CO;;科西嘉語

data/cache/file_data_provider/languages-ZH_TW.txt

@@ -24,7 +24,7 @@ CA;;加泰羅尼亞語
 CS;;捷克語
 CH;;查莫羅語
 CE;;車臣語
-CU;;Church Slavic; Old Slavonic; Church Slavonic; Old Bulgarian; Old Church Slavonic
+CU;;古教會斯拉夫語
 CV;;楚瓦什語
 KW;;康沃爾語
 CO;;科西嘉語

data/cache/file_data_provider/languages-ZU.txt

@@ -1,70 +1,70 @@
 AA;;Afar
-AB;;Isi-Azerbaijani
+AB;;Abkhazian
 AF;;Isibhunu
-AK;;Isi-Azerbaijani
-AM;;isi-Arabhu
+AK;;Akan
+AM;;Amharic
 AR;;isi-Arabhu
 AN;;Aragonese
 AS;;Assamese
-AV;;isi-Arabhu
+AV;;Avaric
 AE;;Avestan
 AY;;Aymara
 AZ;;Isi-Azerbaijani
 BA;;Bashkir
 BM;;Bambara
-BE;;Isi-Buligariyana
+BE;;Belarusian
 BN;;Bengali
-BH;;Isi-Koreyani
+BH;;Bihari languages
 BI;;Bislama
 BO;;Tibetan
-BS;;Isi-Romayinani
+BS;;Bosnian
 BR;;Isi-Bretoni
 BG;;Isi-Buligariyana
 CA;;Catalan; Valencian
 CS;;Isi Czechiya
 CH;;Chamorro
-CE;;Isi Czechiya
+CE;;Chechen
 CU;;Church Slavic; Old Slavonic; Church Slavonic; Old Bulgarian; Old Church Slavonic
 CV;;Chuvash
-KW;;Isi-Irishi
-CO;;Isi-Bosniyani
-CR;;isi-Grike
+KW;;Cornish
+CO;;Corsican
+CR;;Cree
 CY;;Isi Welsh
 DA;;Isi-Danishi
 DE;;isi-Jalimani
 DV;;Divehi; Dhivehi; Maldivian
 DZ;;Dzongkha
 EL;;Greek, Modern (1453-)
-EN;;Isi-Ngisi (UK)
+EN;;English
 EO;;Isi-Esperanto
 ET;;Isi-Estoniyani
 EU;;Isi-Basque
 EE;;Ewe
 FO;;Faroese
-FA;;Isi-Sebhiyani
+FA;;Persian
 FJ;;Fijian
 FI;;Isi-Finishi
 FR;;Isi-Frentshi
-FY;;Isi-Ukraniyani
+FY;;Western Frisian
 FF;;Fulah
 GD;;Gaelic; Scottish Gaelic
 GA;;Isi-Irishi
-GL;;Isi-Galashiyani
+GL;;Galician
 GV;;Manx
-GN;;Isi-Buligariyana
+GN;;Guarani
 GU;;Gujarati
 HT;;Haitian; Haitian Creole
 HA;;Hausa
 HE;;Isi-Hebheru
-HZ;;Isi-Hebheru
+HZ;;Herero
 HI;;Hindi
-HO;;Isi-Thekishi
+HO;;Hiri Motu
 HR;;Isi-Croshiyani
 HU;;Isi-Hangariyani
-HY;;Isi-Romayinani
+HY;;Armenian
 IG;;Igbo
 IO;;Ido
-II;;Isi-Lithuwaniya
+II;;Sichuan Yi; Nuosu
 IU;;Inuktitut
 IE;;Interlingue; Occidental
 IA;;Interlingua (International Auxiliary Language Association)
@@ -72,13 +72,13 @@ ID;;Indoneshiya
 IK;;Inupiaq
 IS;;isi-Icilendikhi
 IT;;Isi-Taliyane
-JV;;Isi-Jaliphani
+JV;;Javanese
 JA;;Isi-Jaliphani
 KL;;Kalaallisut; Greenlandic
 KN;;Kannada
 KS;;Kashmiri
-KA;;isi-Jalimani
-KR;;isi-Kedishi
+KA;;Georgian
+KR;;Kanuri
 KK;;Kazakh
 KM;;Central Khmer
 KI;;Kikuyu; Gikuyu
@@ -90,7 +90,7 @@ KO;;Isi-Koreyani
 KJ;;Kuanyama; Kwanyama
 KU;;isi-Kedishi
 LO;;Lao
-LA;;Isi-Lavithiyani
+LA;;Latin
 LV;;Isi-Lavithiyani
 LI;;Limburgan; Limburger; Limburgish
 LN;;Lingala
@@ -98,13 +98,13 @@ LT;;Isi-Lithuwaniya
 LB;;Luxembourgish; Letzeburgesch
 LU;;Luba-Katanga
 LG;;Ganda
-MH;;Isi-Malithalise
+MH;;Marshallese
 ML;;Malayalam
-MR;;Isi-Mawori
+MR;;Marathi
 MK;;Isi-Makhedoniya
 MG;;Malagasy
 MT;;Isi-Malithalise
-MN;;Isi-Makhedoniya
+MN;;Mongolian
 MI;;Isi-Mawori
 MS;;Malay
 MY;;Burmese
@@ -112,55 +112,55 @@ NA;;Nauru
 NV;;Navajo; Navaho
 NR;;Ndebele, South; South Ndebele
 ND;;Ndebele, North; North Ndebele
-NG;;Isi-Bosniyani
+NG;;Ndonga
 NE;;Nepali
 NL;;Dutch; Flemish
-NN;;Isi-Nowejiyani (Nynorsk)
+NN;;Norwegian Nynorsk; Nynorsk, Norwegian
 NB;;Bokmål, Norwegian; Norwegian Bokmål
-NO;;Isi-Koreyani
+NO;;Norwegian
 NY;;Chichewa; Chewa; Nyanja
 OC;;Occitan (post 1500); Provençal
 OJ;;Ojibwa
 OR;;Oriya
 OM;;Oromo
-OS;;Isi-Rashiyani
+OS;;Ossetian; Ossetic
 PA;;Panjabi; Punjabi
-PI;;Isi-Pholishi
+PI;;Pali
 PL;;Isi-Pholishi
 PT;;Isi-Phuthukesi
 PS;;Pushto; Pashto
 QU;;Quechua
-RM;;Isi-Romayinani
-RO;;Isi-Lavithiyani
+RM;;Romansh
+RO;;Romanian; Moldavian; Moldovan
 RN;;Rundi
 RU;;Isi-Rashiyani
 SG;;Sango
 SA;;Sanskrit
-SI;;Isi-Vitheyinamise
+SI;;Sinhala; Sinhalese
 SK;;Isi-Silovaki
 SL;;Isi-Siloveniyani
 SE;;Northern Sami
 SM;;Samoan
-SN;;Isi-Xhosa
+SN;;Shona
 SD;;Sindhi
 SO;;Somali
 ST;;Sotho, Southern
 ES;;Spanish; Castilian
-SQ;;Isi-Azerbaijani
-SC;;Isi-Makhedoniya
+SQ;;Albanian
+SC;;Sardinian
 SR;;Isi-Sebhiyani
 SS;;Swati
 SU;;Sundanese
 SW;;Swahili
 SV;;Isi-Swedishi
-TY;;Isi-Lavithiyani
+TY;;Tahitian
 TA;;Isi-Thamili
 TT;;Tatar
 TE;;Telugu
-TG;;Isi-Thayi
+TG;;Tajik
 TL;;Tagalog
 TH;;Isi-Thayi
-TI;;Isi-Thekishi
+TI;;Tigrinya
 TO;;Tonga (Tonga Islands)
 TN;;Tswana
 TS;;Tsonga
@@ -168,7 +168,7 @@ TK;;Turkmen
 TR;;Isi-Thekishi
 TW;;Twi
 UG;;Uighur; Uyghur
-UK;;Isi-Ukraniyani
+UK;;Ukrainian
 UR;;Urdu
 UZ;;Uzbek
 VE;;Venda

data/lib/i18n_data/file_data_provider.rb

@@ -13,7 +13,7 @@ module I18nData
     end
 
     def write_cache(provider)
-      languages = provider.codes(:languages, 'EN').keys + ['zh_CN', 'zh_TW', 'zh_HK','bn_IN','pt_BR']
+      languages = provider.codes(:languages, 'EN').keys + ['zh_CN', 'zh_TW', 'zh_HK', 'bn_IN', 'pt_BR', 'sr@latin']
       languages.map do |language_code|
         [:languages, :countries].each do |type|
           begin

data/lib/i18n_data/live_data_provider.rb

@@ -1,4 +1,5 @@
 require 'json'
+require 'simple_po_parser'
 
 module I18nData
   # fetches data online from debian git
@@ -42,20 +43,21 @@ module I18nData
       end
     end
 
-    def translate(type, language, to_language_code)
-      translated = translations(type, to_language_code)[language]
-      translated.to_s.empty? ? nil : translated
-    end
-
     def translated(type, language_code)
       @translated ||= {}
+
       @translated["#{type}_#{language_code}"] ||= begin
-        Hash[send("english_#{type}").map do |code, name|
-          [code, translate(type, name, language_code) || name]
+        Hash[send("alpha_codes_for_#{type}").map do |alpha2, alpha3|
+          [alpha2, translate(type, alpha3, language_code) || fallback_name(type, alpha3)]
         end]
       end
     end
 
+    def translate(type, alpha3, to_language_code)
+      translated = translations(type, to_language_code)[alpha3]
+      translated.to_s.empty? ? nil : translated
+    end
+
     def translations(type, language_code)
       @translations ||= {}
       @translations["#{type}_#{language_code}"] ||= begin
@@ -63,16 +65,50 @@ module I18nData
         code[0].downcase!
         code = code.join("_")
 
-        url = TRANSLATIONS[type]+"#{code}.po"
         begin
-          data = get(url)
+          file_path = "#{CLONE_DEST}/#{TRANSLATIONS[type]}#{code}.po"
+          data = SimplePoParser.parse(file_path)
+          data = data[1..-1] # Remove the initial info block in the .po file
+
+          # Prefer the "Common name for" blocks, but fallback to "Name for" blocks
+          common_names = get_po_data(data, 'Common name for')
+          fallback_names = get_po_data(data, 'Name for')
+
+          fallback_names.merge(common_names)
         rescue Errno::ENOENT
           raise NoTranslationAvailable, "for #{type} and language code = #{code} (#{$!})"
         end
+      end
+    end
+
+    def get_po_data(data, extracted_comment_string)
+      # Ignores the 'fuzzy' entries
+      po_entries = data.select do |t|
+        t[:extracted_comment].start_with?(extracted_comment_string) && t[:flag] != 'fuzzy'
+      end
+
+      # Maps over the alpha3 country code in the 'extracted_comment'
+      # Eg: "Name for GBR"
+      po_entries.map.with_object({}) do |t, translations|
+        alpha3 = t[:extracted_comment][-3..-1].upcase
+        translation = t[:msgstr]
+        translations[alpha3] = translation.is_a?(Array) ? translation.join : translation
+      end
+    end
 
-        data = data.force_encoding('utf-8') if data.respond_to?(:force_encoding) # 1.9
-        data = data.split("\n")
-        po_to_hash data
+    def alpha_codes_for_countries
+      @alpha_codes_for_countries ||= json(:countries)['3166-1'].each_with_object({}) do |entry, codes|
+        alpha2 = entry['alpha_2']&.upcase
+        alpha3 = entry['alpha_3']&.upcase
+        codes[alpha2] = alpha3 unless alpha2.nil?
+      end
+    end
+
+    def alpha_codes_for_languages
+      @alpha_codes_for_languages ||= json(:languages)['639-2'].each_with_object({}) do |entry, codes|
+        alpha2 = entry['alpha_2']&.upcase
+        alpha3 = entry['alpha_3']&.upcase
+        codes[alpha2] = alpha3 unless alpha2.nil?
       end
     end
 
@@ -101,13 +137,8 @@ module I18nData
       end
     end
 
-    def po_to_hash(data)
-      names = data.select{|l| l =~ /^msgid/ }.map{|line| line.match(/^msgid "(.*?)"/)[1] }
-      translations = data.select{|l| l =~ /^msgstr/ }.map{|line| line.match(/^msgstr "(.*?)"/)[1] }
-
-      Hash[names.each_with_index.map do |name,index|
-        [name, translations[index]]
-      end]
+    def fallback_name(type, alpha3)
+      send("english_#{type}")[send("alpha_codes_for_#{type}").invert[alpha3]]
     end
 
     def get(url)

data/lib/i18n_data/version.rb

@@ -1,3 +1,3 @@
 module I18nData
-  VERSION = Version = "0.14.0"
+  VERSION = Version = "0.17.0"
 end

metadata

@@ -1,15 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: i18n_data
 version: !ruby/object:Gem::Version
-  version: 0.14.0
+  version: 0.17.0
 platform: ruby
 authors:
 - Michael Grosser
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2021-12-16 00:00:00.000000000 Z
-dependencies: []
+date: 2022-09-09 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: simple_po_parser
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.1'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - "~>"
+      - !ruby/object:Gem::Version
+        version: '1.1'
 description: 
 email: michael@grosser.it
 executables: []
@@ -121,6 +135,7 @@ files:
 - cache/file_data_provider/countries-SO.txt
 - cache/file_data_provider/countries-SQ.txt
 - cache/file_data_provider/countries-SR.txt
+- cache/file_data_provider/countries-SR@LATIN.txt
 - cache/file_data_provider/countries-SV.txt
 - cache/file_data_provider/countries-SW.txt
 - cache/file_data_provider/countries-TA.txt
@@ -181,6 +196,7 @@ files:
 - cache/file_data_provider/languages-IS.txt
 - cache/file_data_provider/languages-IT.txt
 - cache/file_data_provider/languages-JA.txt
+- cache/file_data_provider/languages-KA.txt
 - cache/file_data_provider/languages-KN.txt
 - cache/file_data_provider/languages-KO.txt
 - cache/file_data_provider/languages-LT.txt
@@ -210,6 +226,7 @@ files:
 - cache/file_data_provider/languages-SL.txt
 - cache/file_data_provider/languages-SQ.txt
 - cache/file_data_provider/languages-SR.txt
+- cache/file_data_provider/languages-SR@LATIN.txt
 - cache/file_data_provider/languages-SV.txt
 - cache/file_data_provider/languages-TA.txt
 - cache/file_data_provider/languages-TE.txt
@@ -250,7 +267,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.2.16
+rubygems_version: 3.1.6
 signing_key: 
 specification_version: 4
 summary: country/language names and 2-letter-code pairs, in 85 languages