interscript 0.1.3 → 0.1.9

data/maps/bgnpcgn-isl-Latn-Latn-1964.yaml

@@ -0,0 +1,37 @@
+---
+authority_id: bgnpcgn
+id: 1964
+language: iso-639-2:isl
+source_script: Latn
+destination_script: Latn
+name: TRANSLITERATION OF ICELANDIC BGN/PCGN 1947 System
+url: https://github.com/interscript/interscript/files/5180785/BGN_Romanization_Guide_1964_icelandic_1947.pdf
+creation_date: 1964
+description: |
+  The BGN and the PCGn in 1947 jointly agreed to the transliteration of two letters of the Icelandic alphabet which,
+  although used in writing Old English, have disappeared from the modern English alphabet.
+
+  The transliterated letters are the edh (Ð ð) and the thorn (Þ þ), pronounved as th in "thus" and th in "think," respectively.
+
+  It was felt that it was better to transliterate these letters into familiar symbols than preserve such unfamiliar letters in the nomenclature.
+
+notes:
+  - More about "edh" letter - https://en.wikipedia.org/wiki/Eth
+  - More about "thorn" letter - https://en.wikipedia.org/wiki/Thorn_(letter)
+
+tests:
+  - source: Fyrirgefðu
+    expected: Fyrirgefdhu
+  - source: þu ert velkominn
+    expected: thu ert velkominn
+  - source: GOÐAN DAGINN
+    expected: GODHAN DAGINN
+  - source: Þakka
+    expected: Thakka
+
+map:
+  characters:
+    "\u00D0": "Dh" # Ð
+    "\u00F0": "dh" # ð
+    "\u00DE": "Th" # Þ
+    "\u00FE": "th" # þ

data/maps/bgnpcgn-jpn-Hrkt-Latn-1976.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bgnpcgn
 id: 1930
-language: jpn
+language: iso-639-2:jpn
 source_script: Hrkt
 destination_script: Latn
 name: Japanese Kana Modified Hepburn 1930 System

data/maps/bgnpcgn-kat-Geor-Latn-1981.yaml

@@ -1,10 +1,14 @@
 ---
 authority_id: bgnpcgn
 id: 1981
-language: kat
+language: iso-639-2:kat
 source_script: Geor
 destination_script: Latn
 name: ROMANIZATION OF GEORGIAN; BGN/PCGN 1981 System
+alias:
+  ogc11122:
+    code: kat_Geor2Latn_BGN_1981
+    description: Georgian US Board on Geographic Names(BGN)/Permanent Committee on Geographical Names for British Official Use(PCGN) 1981 System
 url: https://transliteration.eki.ee/pdf/Georgian.pdf
 creation_date: 1981
 confirmation_date: 1981

data/maps/bgnpcgn-kat-Geor-Latn-2009.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bgnpcgn
 id: 2009
-language: kat
+language: iso-639-2:kat
 source_script: Geor
 destination_script: Latn
 name: ROMANIZATION OF GEORGIAN; Georgia 2011 national system; BGN/PCGN 2009 Agreement
@@ -40,4 +40,3 @@ tests:
 
 map:
   inherit: "ggg-kat-Geor-Latn-2002"
-

data/maps/bgnpcgn-kaz-Cyrl-Latn-1979.yaml

@@ -0,0 +1,247 @@
+
+---
+authority_id: bgnpcgn
+id: 1979
+language: iso-639-2:kaz
+source_script: Cyrl
+destination_script: Latn
+name: Romanization of Kazakh
+url: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/811511/ROMANIZATION_OF_KAZAKH.pdf
+creation_date: 1979
+confirmation_date: 2019
+description: |
+  The BGN/PCGN system for Kazakh was designed for use in romanizing names written in the Kazakh Cyrillic alphabet.
+  The Kazakh Cyrillic alphabet contains nine characters not present in the Russian alphabet: ә, ғ, қ, ң, ө, ұ, ү, һ and і.
+
+notes:
+  - The character sequences гһ, зһ, кһ, нг, сһ and цһ may be romanized g·h, z·h, k·h, n·g, s·h
+    and ts·h in order to differentiate those romanizations from from the digraphs gh, zh, kh, ng, sh,
+    and the letter sequence tsh, which are used to render the characters ғ, ж, х, ң, ш, and the character sequence тш.
+  - The character ы may be romanized i̵ (Unicode encoding 0069+0335) instead of у, if so desired.
+  - Unicode values are shown with the uppercase Cyrillic character first, followed by the lowercase character.
+  - The Kazakh government has adopted a programme to move to using the Roman-script as the principal writing system for Kazakh.
+  - 'An inventory of letter-diacritic combinations, with their Unicode encoding, in addition to the
+    unmodified letters of the basic Roman script is:
+    Ä (U+00C4)    ӓ (U+00E4)
+    Ī (U+012A)    ī (U+012B)
+    Ö (U+00D6)    ö (U+00F6)
+    Ū (U+016A)    ū (U+016B)
+    Ü (U+00DC)    ü (U+00FC)
+    ” (U+201D)    ’ (U+2019)
+    Ė (U+0116)    ė (U+0117)'
+  - The Romanization column shows only lowercase forms but, when romanizing, uppercase and
+    lowercase Roman letters as appropriate should be used.
+
+tests:
+  - source: Өңірек
+    expected: Öngirek
+  - source: Өтебас Артезиан Құдығы
+    expected: Ötebas Artezīan Qudyghy
+  - source: Өскенбай
+    expected: Öskenbay
+  - source: Өсек Көлі
+    expected: Ösek Köli
+  - source: Өрмексу
+    expected: Örmeksū
+  - source: Өмірзақ
+    expected: Ömirzaq
+  - source: Өлеңті
+    expected: Ölengti
+  - source: Өл-Фараби Даңғылы
+    expected: Öl-Farabī Dangghyly
+  - source: Өкпекті Тауы
+    expected: Ökpekti Taūy
+  - source: Өкенсоркен Қыстауы
+    expected: Ökensorken Qystaūy
+  - source: Өзен Ойысы
+    expected: Özen Oyysy
+  - source: Өзен
+    expected: Özen
+  - source: Өгізтөбе Тауы
+    expected: Ögiztöbe Taūy
+  - source: Өгізтау Қыстауы
+    expected: Ögiztaū Qystaūy
+  - source: Өгізмүйіз Тауы
+    expected: Ögizmüyiz Taūy
+  - source: Өгізбұлақ
+    expected: Ögizbulaq
+  - source: Өгіз Үреулі
+    expected: Ögiz Üreūli
+  - source: Өгем Жотасы
+    expected: Ögem Zhotasy
+  - source: Өгем
+    expected: Ögem
+  - source: Әшім
+    expected: Äshim
+  - source: Әулиетөбе Тауы
+    expected: Äūlīetöbe Taūy
+  - source: Әулиекөл
+    expected: Äūlīeköl
+  - source: Әндіжан Құдығы
+    expected: Ändizhan Qudyghy
+  - source: Ұясай
+    expected: Uyasay
+  - source: Ұялы Метеорологиялық Станциясы
+    expected: Uyaly Meteorologīyalyq Stantsīyasy
+  - source: Ұшқын Қыстауы
+    expected: Ushqyn Qystaūy
+  - source: Үңгіртас
+    expected: Ünggirtas
+  - source: Үшқұлын
+    expected: Üshqulyn
+  - source: Құтырғы Асуы
+    expected: Qutyrghy Asūy
+  - source: Ярмы Стансасы
+    expected: Yarmy Stansasy
+  - source: Юпитер Қыстауы
+    expected: Yupīter Qystaūy
+  - source: Энгельс Көшесi
+    expected: Ėngel’s Köshesi
+  - source: Ырғызбай Жайлауы
+    expected: Yrghyzbay Zhaylaūy
+  - source: Щебнюха Тауы
+    expected: Shchebnyukha Taūy
+  - source: Шөміштікөл Соры
+    expected: Shömishtiköl Sory
+  - source: Чалов Барак Қыстауы
+    expected: Chalov Barak Qystaūy
+  - source: Чайкино
+    expected: Chaykīno
+  - source: Цуриковка
+    expected: Tsūrīkovka
+  - source: Хамитқора Қыстауы
+    expected: Khamītqora Qystaūy
+  - source: Фыкалка
+    expected: Fykalka
+  - source: Уақбай Қыстауы
+    expected: Ūaqbay Qystaūy
+  - source: Төңірекшың Тоғайы
+    expected: Töngirekshyng Toghayy
+  - source: Сабағали Қыстауы
+    expected: Sabaghalī Qystaūy
+  - source: Рысқұлов Даңғылы
+    expected: Rysqulov Dangghyly
+  - source: Пірназар Құдығы
+    expected: Pirnazar Qudyghy
+  - source: Оңтүстік Қазақстан Облысы
+    expected: Ongtüstik Qazaqstan Oblysy
+  - source: Нөмір Үшінші Суторабының Бөгені
+    expected: Nömir Üshinshi Sūtorabynyng Bögeni
+  - source: Мәмбетқазған Құдығы
+    expected: Mämbetqazghan Qudyghy
+  - source: Мемлекеттік Аудандық Электр Стансасы - Бір
+    expected: Memlekettik Aūdandyq Ėlektr Stansasy - Bir
+  - source: Линейский Белок Тауы
+    expected: Līneyskīy Belok Taūy
+  - source: Көшердік Бөгені
+    expected: Kösherdik Bögeni
+  - source: Көлфонтан Артезиан Құдығы
+    expected: Kölfontan Artezīan Qudyghy
+  - source: Изендіарал Мүйісі
+    expected: Īzendiaral Müyisi
+  - source: Злиха Метеорологиялық Станциасы
+    expected: Zlīkha Meteorologīyalyq Stantsīasy
+  - source: Жұлжұрған Көлі
+    expected: Zhulzhurghan Köli
+  - source: Ескі Үшал Қыстауы
+    expected: Eski Üshal Qystaūy
+  - source: Дөңгелексор Қыстауы
+    expected: Dönggeleksor Qystaūy
+  - source: Горько-Солёное Көлі
+    expected: Gor’ko-Solyonoe Köli
+  - source: Вагулино
+    expected: Vagūlīno
+  - source: Бөстай Учаскесі
+    expected: Böstay Ūchaskesi
+  - source: Аққолқы Тоғайы
+    expected: Aqqolqy Toghayy
+  - source: Іңқардария
+    expected: Ingqardarīya
+
+map:
+  characters:
+    '\u0410': 'A'      # А
+    '\u04D8': 'Ä'      # Ә
+    '\u0411': 'B'      # Б
+    '\u0412': 'V'      # В
+    '\u0413': 'G'      # Г
+    '\u0492': 'Gh'     # Ғ
+    '\u0414': 'D'      # Д
+    '\u0415': 'E'      # Е
+    '\u0401': 'Yo'     # Ё
+    '\u0416': 'Zh'     # Ж
+    '\u0417': 'Z'      # З
+    '\u0418': 'Ī'      # И
+    '\u0419': 'Y'      # Й
+    '\u041A': 'K'      # К
+    '\u049A': 'Q'      # Қ
+    '\u041B': 'L'      # Л
+    '\u041C': 'M'      # М
+    '\u041D': 'N'      # Н
+    '\u04A2': 'Ng'     # Ң
+    '\u041E': 'O'      # О
+    '\u04E8': 'Ö'      # Ө
+    '\u041F': 'P'      # П
+    '\u0420': 'R'      # Р
+    '\u0421': 'S'      # С
+    '\u0422': 'T'      # Т
+    '\u0423': 'Ū'      # У
+    '\u04B0': 'U'      # Ұ
+    '\u04AE': 'Ü'      # Ү
+    '\u0424': 'F'      # Ф
+    '\u0425': 'Kh'     # Х
+    '\u04BA': 'H'      # Һ
+    '\u0426': 'Ts'     # Ц
+    '\u0427': 'Ch'     # Ч
+    '\u0428': 'Sh'     # Ш
+    '\u0429': 'Shch'   # Щ
+    '\u042A': '”'      # Ъ
+    '\u042B': 'Y'      # Ы
+    '\u0406': 'I'      # І
+    '\u042C': '’'      # Ь
+    '\u042D': 'Ė'      # Э
+    '\u042E': 'Yu'     # Ю
+    '\u042F': 'Ya'     # Я
+
+    '\u0430': 'a'      # а
+    '\u04D9': 'ä'      # ә
+    '\u0431': 'b'      # б
+    '\u0432': 'v'      # в
+    '\u0433': 'g'      # г
+    '\u0493': 'gh'     # ғ
+    '\u0434': 'd'      # д
+    '\u0435': 'e'      # e
+    '\u0451': 'yo'     # ё
+    '\u0436': 'zh'     # ж
+    '\u0437': 'z'      # з
+    '\u0438': 'ī'      # и
+    '\u0439': 'y'      # й
+    '\u043A': 'k'      # к
+    '\u049B': 'q'      # қ
+    '\u043B': 'l'      # л
+    '\u043C': 'm'      # м
+    '\u043D': 'n'      # н
+    '\u04A3': 'ng'     # ң
+    '\u043E': 'o'      # о
+    '\u04E9': 'ö'      # ө
+    '\u043F': 'p'      # п
+    '\u0440': 'r'      # р
+    '\u0441': 's'      # с
+    '\u0442': 't'      # т
+    '\u0443': 'ū'      # у
+    '\u04B1': 'u'      # ұ
+    '\u04AF': 'ü'      # ү
+    '\u0444': 'f'      # ф
+    '\u0445': 'kh'     # х
+    '\u04BB': 'h'      # һ
+    '\u0446': 'ts'     # ц
+    '\u0447': 'ch'     # ч
+    '\u0448': 'sh'     # ш
+    '\u0449': 'shch'   # щ
+    '\u044A': '”'      # ъ
+    '\u044B': 'y'      # ы
+    '\u0456': 'i'      # і
+    '\u044C': '’'      # ь
+    '\u044D': 'ė'      # э
+    '\u044E': 'yu'     # ю
+    '\u044F': 'ya'     # я

data/maps/bgnpcgn-kir-Cyrl-Latn-1979.yaml

@@ -0,0 +1,218 @@
+---
+authority_id: bgnpcgn
+id: 1979
+language: iso-639-2:kir
+source_script: Cyrl
+destination_script: Latn
+name: Romanization of Kyrgyz
+url: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/816663/ROMANIZATION_OF_KYRGYZ.pdf
+creation_date: 1979
+confirmation_date: 2019
+description: |
+  The BGN/PCGN system for Kyrgyz Cyrillic was designed for use in romanizing names written
+  in the Kyrgyz Cyrillic alphabet. The Kyrgyz Cyrillic alphabet contains three characters not present in
+  the Russian alphabet: Ң , Ө, and Y.
+
+notes:
+  - Both Kyrgyz and Kirghiz may frequently be seen as the language name; both these spellings are used in the
+    ISO 639 Standard on the representation of names for languages.
+  - The character sequence н г may be romanized n·g in order to differentiate that romanization
+    from the digraph ng, which is used to render the character ң.
+  - The character ы may be romanized i (Unicode encoding 0069+0335) instead of y, if so desired.
+  - 'An inventory of letter-diacritic combinations, with their Unicode encoding, in addition to the
+    unmodified letters of the basic Roman script is:
+    All apostrophes appearing in romanization are U+2019
+    Ö (U+00D6)  ö (U+00F6)
+    Ü (U+00DC) ü (U+00FC)'
+  - The Romanization column shows only lowercase forms but, when romanizing, uppercase and
+    lowercase Roman letters as appropriate should be used.
+
+tests:
+  - source: Ысык-Көл Облусу
+    expected: Ysyk-Köl Oblusu
+  - source: Ысык-Көл
+    expected: Ysyk-Köl
+  - source: Шедвик-Сай
+    expected: Shedvik-Say
+  - source: Чүй Облусу
+    expected: Chüy Oblusu
+  - source: Чүй
+    expected: Chüy
+  - source: Чирик-Сай
+    expected: Chirik-Say
+  - source: Хребет Джети-Сандал
+    expected: Khrebet Djeti-Sandal
+  - source: Узук-Булак
+    expected: Uzuk-Bulak
+  - source: Торугарт Ашуу
+    expected: Torugart Ashuu
+  - source: Торетал
+    expected: Toretal
+  - source: Терек
+    expected: Terek
+  - source: Талды-Булак
+    expected: Taldy-Bulak
+  - source: Талас Облусу
+    expected: Talas Oblusu
+  - source: Талас
+    expected: Talas
+  - source: Сарык-Кёль
+    expected: Saryk-Kyol’
+  - source: Родник Кара-Суу
+    expected: Rodnik Kara-Suu
+  - source: Родник Бейрёк-Булак
+    expected: Rodnik Beyryok-Bulak
+  - source: Перевал Сары-Челек
+    expected: Pereval Sary-Chelek
+  - source: Перевал Макмал
+    expected: Pereval Makmal
+  - source: Перевал Кара-Токой
+    expected: Pereval Kara-Tokoy
+  - source: Перевал Ашуу-Тёр
+    expected: Pereval Ashuu-Tyor
+  - source: Перевал Ашуу
+    expected: Pereval Ashuu
+  - source: Ош Шаары
+    expected: Osh Shaary
+  - source: Ош Облусу
+    expected: Osh Oblusu
+  - source: Ош
+    expected: Osh
+  - source: Ош
+    expected: Osh
+  - source: Осоавиахим
+    expected: Osoaviakhim
+  - source: Озеро Афлатук
+    expected: Ozero Aflatuk
+  - source: Нарын Облусу
+    expected: Naryn Oblusu
+  - source: Нарын
+    expected: Naryn
+  - source: Метеорологическая Станция Чамкал
+    expected: Meteorologicheskaya Stantsiya Chamkal
+  - source: Марза-Булак
+    expected: Marza-Bulak
+  - source: Макмал
+    expected: Makmal
+  - source: Кыргызстан
+    expected: Kyrgyzstan
+  - source: Кыргыз Республикасы
+    expected: Kyrgyz Respublikasy
+  - source: Куру-Сай
+    expected: Kuru-Say
+  - source: Куру-Сай
+    expected: Kuru-Say
+  - source: Кур-Пырылды
+    expected: Kur-Pyryldy
+  - source: Кок-Бель-Таш
+    expected: Kok-Bel’-Tash
+  - source: Кичи-Сандык
+    expected: Kichi-Sandyk
+  - source: Кель-Сай
+    expected: Kel’-Say
+  - source: Карагайлы
+    expected: Karagayly
+  - source: Кара-Суу
+    expected: Kara-Suu
+  - source: Жалал-Абад Облусу
+    expected: Jalal-Abad Oblusu
+  - source: Жалал-Абад
+    expected: Jalal-Abad
+  - source: Долина Беш-Башат
+    expected: Dolina Besh-Bashat
+  - source: Гора Арпа-Турча
+    expected: Gora Arpa-Turcha
+  - source: Бишкек Шаары
+    expected: Bishkek Shaary
+  - source: Бишкек
+    expected: Bishkek
+  - source: Бишкек
+    expected: Bishkek
+  - source: Баткен Облусу
+    expected: Batken Oblusu
+  - source: Баткен
+    expected: Batken
+  - source: Аяк-Терек
+    expected: Ayak-Terek
+  - source: Аюу-Чача
+    expected: Ayuu-Chacha
+  - source: Арпа
+    expected: Arpa
+  - source: Ак-Суу
+    expected: Ak-Suu
+
+map:
+  characters:
+    '\u0410': 'A'      # А
+    '\u0411': 'B'      # Б
+    '\u0412': 'V'      # В
+    '\u0413': 'G'      # Г
+    '\u0414': 'D'      # Д
+    '\u0415': 'E'      # Е
+    '\u0401': 'Yo'     # Ё
+    '\u0416': 'J'      # Ж
+    '\u0417': 'Z'      # З
+    '\u0418': 'I'      # И
+    '\u0419': 'Y'      # Й
+    '\u041A': 'K'      # К
+    '\u041B': 'L'      # Л
+    '\u041C': 'M'      # М
+    '\u041D': 'N'      # Н
+    '\u04A2': 'Ng'     # Ң
+    '\u041E': 'O'      # О
+    '\u04E8': 'Ö'      # Ө
+    '\u041F': 'P'      # П
+    '\u0420': 'R'      # Р
+    '\u0421': 'S'      # С
+    '\u0422': 'T'      # Т
+    '\u0423': 'U'      # У
+    '\u04AE': 'Ü'      # Ү
+    '\u0424': 'F'      # Ф
+    '\u0425': 'Kh'     # Х
+    '\u0426': 'Ts'     # Ц
+    '\u0427': 'Ch'     # Ч
+    '\u0428': 'Sh'     # Ш
+    '\u0429': 'Shch'   # Щ
+    '\u042A': '”'      # Ъ
+    '\u042B': 'Y'      # Ы
+    '\u042C': '’'      # Ь
+    '\u042D': 'E'      # Э
+    '\u042E': 'Yu'     # Ю
+    '\u042F': 'Ya'     # Я
+
+    '\u0430': 'a'      # а
+    '\u0431': 'b'      # б
+    '\u0432': 'v'      # в
+    '\u0433': 'g'      # г
+    '\u0434': 'd'      # д
+    '\u0435': 'e'      # e
+    '\u0451': 'yo'     # ё
+    '\u0436': 'j'      # ж
+    '\u0437': 'z'      # з
+    '\u0438': 'i'      # и
+    '\u0439': 'y'      # й
+    '\u043A': 'k'      # к
+    '\u043B': 'l'      # л
+    '\u043C': 'm'      # м
+    '\u043D': 'n'      # н
+    '\u04A3': 'ng'     # ң
+    '\u043E': 'o'      # о
+    '\u04E9': 'ö'      # ө
+    '\u043F': 'p'      # п
+    '\u0440': 'r'      # р
+    '\u0441': 's'      # с
+    '\u0442': 't'      # т
+    '\u0443': 'u'      # у
+    '\u04AF': 'ü'      # ү
+    '\u0444': 'f'      # ф
+    '\u0445': 'kh'     # х
+    '\u0446': 'ts'     # ц
+    '\u0447': 'ch'     # ч
+    '\u0448': 'sh'     # ш
+    '\u0449': 'shch'   # щ
+    '\u044A': '”'      # ъ
+    '\u044B': 'y'      # ы
+    '\u044C': '’'      # ь
+    '\u044D': 'e'      # э
+    '\u044E': 'yu'     # ю
+    '\u044F': 'ya'     # я