i18n_data 0.12.0 → 0.15.0

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/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 = "0.12.0"
+  VERSION = Version = "0.15.0"
 end

data/lib/i18n_data.rb

@@ -78,7 +78,7 @@ module I18nData
         options.each do |code, name|
           # support "Dutch" and "Dutch; Flemish", checks for inclusion first to skip the splitting
           # then check for exact match
-          return code if name.include?(search) && (name == search || name.split(';').each(&:strip!).include?(search))
+          return code if name.include?(search) && (name == search || name.split('; ').include?(search))
         end
       end
 

metadata

@@ -1,15 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: i18n_data
 version: !ruby/object:Gem::Version
-  version: 0.12.0
+  version: 0.15.0
 platform: ruby
 authors:
 - Michael Grosser
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-10-03 00:00:00.000000000 Z
-dependencies: []
+date: 2021-12-20 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: []
@@ -82,7 +96,6 @@ files:
 - cache/file_data_provider/countries-KM.txt
 - cache/file_data_provider/countries-KN.txt
 - cache/file_data_provider/countries-KO.txt
-- cache/file_data_provider/countries-KU.txt
 - cache/file_data_provider/countries-KV.txt
 - cache/file_data_provider/countries-KW.txt
 - cache/file_data_provider/countries-KY.txt
@@ -251,7 +264,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.1.2
+rubygems_version: 3.2.16
 signing_key: 
 specification_version: 4
 summary: country/language names and 2-letter-code pairs, in 85 languages

data/cache/file_data_provider/countries-KU.txt

@@ -1,249 +0,0 @@
-AW;;Aruba
-AF;;Efxanistan
-AO;;Angola
-AI;;Angîla
-AX;;Giravên Åland
-AL;;Arnavût
-AD;;Andora
-AE;;Mîrtiyên Ereb ên Yekbuyî
-AR;;Arjantîn
-AM;;Ermenistan
-AS;;Samoya Emerîkî
-AQ;;Antarktîka
-TF;;Herêma Basurî ya Fransî
-AG;;Antîguya û Barbuda
-AU;;Awustralya
-AT;;Awusturya
-AZ;;Azerbaycan
-BI;;Bûrûndî
-BE;;Belçîka
-BJ;;Benîn
-BQ;;Bonaire, Sint Eustatius û Saba
-BF;;Burkîna Faso
-BD;;Banglades
-BG;;Bulgaristan
-BH;;Bahrayîn
-BS;;Giravên Bahama
-BA;;Bosna Hersek
-BL;;Saint Barthélemy
-BY;;Belarus
-BZ;;Belîze
-BM;;Bermûda
-BO;;Bolîvya
-BR;;Brezîlya
-BB;;Barbados
-BN;;Brûney Darûssalam
-BT;;Bûtan
-BV;;Girava Boyuvet
-BW;;Botswana
-CF;;Komara Afrikaya Navîn
-CA;;Kanada
-CC;;Grava Kokos (Kîlîng)
-CH;;Swîsre
-CL;;Sîlî
-CN;;Çîn
-CI;;Sahila Diranên Fîlan
-CM;;Kamerûn
-CD;;Komara Demokratîk a Kongo
-CG;;Kongo
-CK;;Giravên Cook
-CO;;Kolombiya
-KM;;Komoros
-CV;;Kap Verde
-CR;;Kosta Rîka
-CU;;Kûba
-CW;;Curaçao
-CX;;Girava Noel
-KY;;Giravên Kayman
-CY;;Kibris
-CZ;;Çekistan
-DE;;Almanya
-DJ;;Cîbutî
-DM;;Domînîka
-DK;;Danîmarka
-DO;;Komara Domînîk
-DZ;;Cezayîr
-EC;;Ekvator
-EG;;Misir
-ER;;Erître
-EH;;Sehraya Rojava
-ES;;Spanya
-EE;;Estonya
-ET;;Etîyopya
-FI;;Fînlandiya
-FJ;;Fîjî
-FK;;Giravên Falkland (Malvin)
-FR;;Fransa
-FO;;Giravên Faroye
-FM;;Mîkronesya, Dewletên Federal yên
-GA;;Gabon
-GB;;Şahitiya Yekbuyî
-GE;;Gurcistan
-GG;;Guernsey
-GH;;Gana
-GI;;Cebelîtariq
-GN;;Gîne
-GP;;Guadelûp
-GM;;Gambiya
-GW;;Gîneya Bisoî
-GQ;;Gîneya Ekvatorî
-GR;;Yewnanistan
-GD;;Girava Grenada
-GL;;Greenland
-GT;;Guatemala
-GF;;Gîneya Fransî
-GU;;Guam
-GY;;Guyana
-HK;;Hong Kong
-HM;;Giravên Heard û McDonald
-HN;;Honduras
-HR;;Hirvatistan
-HT;;Haîtî
-HU;;Macaristan
-ID;;Endonezya
-IM;;Girava Man
-IN;;Hindistan
-IO;;Herêma Îngîlîz a Deryaya Hindî
-IE;;Îrlanda
-IR;;Îran, Komara Îslamî ya
-IQ;;Iraq
-IS;;Îzlanda
-IL;;Îsraîl
-IT;;Îtalya
-JM;;Jamayîka
-JE;;Jersey
-JO;;Urdin
-JP;;Japonya
-KZ;;Kazaxstan
-KE;;Kenya
-KG;;Kirgîsistan
-KH;;Kamboca
-KI;;Kirîbatî
-KN;;Saint Kitts û Nevis
-KR;;Kore, Komara
-KW;;Kûveyt
-LA;;Komara Demokratîka Gelê Lao
-LB;;Lubnan
-LR;;Lîberya
-LY;;Lîbya
-LC;;Saint Lucia
-LI;;Lihtenstayn
-LK;;Srî Lanka
-LS;;Lesoto
-LT;;Lîtvanya
-LU;;Luksemburg
-LV;;Latviya
-MO;;Makao
-MF;;Saint Martin (beşa firansî)
-MA;;Fas
-MC;;Monako
-MD;;Moldova
-MG;;Madagaskar
-MV;;Giravên Maldiv
-MX;;Mexîko
-MH;;Giravên Marsal
-MK;;Kaledonyaya Nû
-ML;;Malî
-MT;;Malta
-MM;;Miyanmar
-ME;;Montenegro
-MN;;Mongolya
-MP;;Giravên Marîyana Bakûr
-MZ;;Mozambîk
-MR;;Morîtanya
-MS;;Montserat
-MQ;;Martinîk
-MU;;Morîtiyus
-MW;;Malawî
-MY;;Malezya
-YT;;Mayote
-NA;;Namîbiya
-NC;;Kaledonyaya Nû
-NE;;Nîjer
-NF;;Girava Norfolk
-NG;;Nîjerya
-NI;;Nîkaragûa
-NU;;Nîûe
-NL;;Holanda
-NO;;Norweç
-NP;;Nepal
-NR;;Naûrû
-NZ;;Zelandaya Nû
-OM;;Uman
-PK;;Pakistan
-PA;;Panama
-PN;;Pitkayirn
-PE;;Perû
-PH;;Fîlîpîn
-PW;;Palawu
-PG;;Papûa Gîneya Nû
-PL;;Polonya
-PR;;Porto Rîko
-KP;;Kore, Komara Demokratîk a Gelê
-PT;;Portekîz
-PY;;Paraguay
-PS;;Palestine, State of
-PF;;Polînesya Fransî
-QA;;Qatar
-RE;;Reûnyon
-RO;;Romanya
-RU;;Federasyona Rûsî
-RW;;Rûanda
-SA;;Si'ûdî Erebistan
-SD;;Sûdan
-SN;;Senegal
-SG;;Sîngapur
-GS;;Georgiaya Basûr û Giravên Sandwicha Basûr
-SH;;Saint Helena, Ascension and Tristan da Cunha
-SJ;;Svalbard û Jan Mayen
-SB;;Giravên Solomon
-SL;;Sîera Leone
-SV;;El Salvador
-SM;;San Marîno
-SO;;Somalî
-PM;;Saint Pierre û Miquelon
-RS;;Sirbistan
-SS;;Komara Sûdan
-ST;;Sao Tome û Prinsipe
-SR;;Surînam
-SK;;Slovakya
-SI;;Slovenya
-SE;;Swêd
-SZ;;Eswatini
-SX;;Saint Martin (beşa firansî)
-SC;;Seyselan
-SY;;Sûriye
-TC;;Giravên Turks û Caicos
-TD;;Çad
-TG;;Togo
-TH;;Tayland
-TJ;;Tacîkistan
-TK;;Tokelau
-TM;;Tirkmenîstan
-TL;;Tîmora Rojhilat
-TO;;Tonga
-TT;;Trînîdad û Tobago
-TN;;Tûnis
-TR;;Tirkiye
-TV;;Tûvalû
-TW;;Taywan
-TZ;;Tanzania
-UG;;Ûganda
-UA;;Ukrayna
-UM;;Giravên biçûk derve Emerîkayê
-UY;;Uruguay
-US;;Dewletên Yekbûyî
-UZ;;Ozbekîstan
-VA;;Texta Pîroz (Dewleta Bajarî ya Vatîkan)
-VC;;Saint Vincent û Grenadines
-VE;;Venezûela
-VG;;Giravên Virjin, Brîtanî
-VI;;Giravên Virjin, D.Y.
-VN;;Viyetnam
-VU;;Vanûatû
-WF;;Walis û Futuna
-WS;;Samoa
-YE;;Yemen
-ZA;;Efrîka Basur
-ZM;;Zambiya
-ZW;;Zimbabwe