interscript 0.1.2 → 0.1.7

data/maps/alalc-ukr-Cyrl-Latn-2011.yaml

@@ -0,0 +1,16 @@
+---
+authority_id: alalc
+id: 2011
+language: iso-639-2:ukr
+source_script: Cyrl
+destination_script: Latn
+name: ALA-LC Romanization System 1997
+url: https://www.loc.gov/catdir/cpso/romanization/ukrainia.pdf
+creation_date: 2011
+
+notes:
+  - The z͡h ligature is necessary to distinguish ж from the combination зг.
+  - The t͡s ligature is necessary to distinguish ц from the combination тс.
+
+map:
+  inherit: alalc-ukr-Cyrl-Latn-1997

data/maps/apcbg-bul-Cyrl-Latn-1995.yaml

@@ -0,0 +1,283 @@
+---
+authority_id: apcbg
+id: 1995
+language: iso-639-2:bul
+source_script: Cyrl
+destination_script: Latn
+name: Toponymic Guidelines for Antarctica, Lyubomir Ivanov. Antarctic Place-names Commission of Bulgaria, Sofia, 1995.
+url: http://www.apcbg.org/Toponymic-Guidelines.htm
+creation_date: 1995
+description: |
+  In Antarctica geographical names are important elements of identification, orientation, localization and navigation,
+  providing an essential reference system for logistic operations, including search and rescue measures, and for
+  international scientific research.  They facilitate information exchange in the field, in scientific publications and in
+  administrative measures under the Antarctic Treaty System. Geographical names also reflect the history of Antarctic
+  exploration.
+
+tests:
+  - source: нунатак Абрит
+    expected: nunatak Abrit
+  - source: връх Академия
+    expected: vrah Akademiya
+  - source: връх Ами Буе
+    expected: vrah Ami Bue
+  - source: нос Айтос
+    expected: nos Aytos
+  - source: залив Баба Тонка
+    expected: zaliv Baba Tonka
+  - source: Балабански камък
+    expected: Balabanski kamak
+  - source: Бедечки поток
+    expected: Bedechki potok
+  - source: нос Бяга
+    expected: nos Byaga
+  - source: остров Качо
+    expected: ostrov Kacho # wiki ostrov Cacho
+  - source: Чакъров остров
+    expected: Chakarov ostrov
+  - source: връх Дъбник
+    expected: vrah Dabnik
+  - source: залив Десислава
+    expected: zaliv Desislava
+  - source: ледник Джераси
+    expected: lednik Dzherasi
+  - source: Джегова скала
+    expected: Dzhegova skala
+  - source: Нунатак Едуард
+    expected: Nunatak Eduard
+  - source: Елховска седловина
+    expected: Elhovska sedlovina
+  - source: ледник Етър
+    expected: lednik Etar
+  - source: нунатак Филип Тотю
+    expected: nunatak Filip Totyu
+  - source: ледник Габаре
+    expected: lednik Gabare
+  - source: риф Гергини
+    expected: rif Gergini
+  - source: Гяуров връх
+    expected: Gyaurov vrah
+  - source: Гуцалски рид
+    expected: Gutsalski rid # not found
+  - source: Хараламбиев остров
+    expected: Haralambiev ostrov
+  - source: връх Ичера
+    expected: vrah Ichera
+  - source: полуостров Йоан Павел II
+    expected: poluostrov Yoan Pavel II
+  - source: нос Иван Александър
+    expected: nos Ivan Aleksandar
+  - source: нос Иречек
+    expected: nos Irechek
+  - source: нос Кърджали
+    expected: nos Kardzhali
+  - source: седловина Кърнаре
+    expected: sedlovina Karnare
+  - source: нунатак Керсеблепт
+    expected: nunatak Kerseblept
+  - source: Кондофрейски възвишения
+    expected: Kondofreyski vazvisheniya
+  - source: Костинбродски проход
+    expected: Kostinbrodski prohod
+  - source: връх Кожух
+    expected: vrah Kozhuh
+  - source: Кукерски нунатаци
+    expected: Kukerski nunatatsi
+  - source: залив Лазурен бряг
+    expected: zaliv Lazuren bryag
+  - source: връх Лудогорие
+    expected: vrah Ludogorie
+  - source: Лютибродски скали
+    expected: Lyutibrodski skali
+  - source: Масларов нунатак
+    expected: Maslarov nunatak
+  - source: Михневски връх
+    expected: Mihnevski vrah
+  - source: рид Митино
+    expected: rid Mitino
+  - source: езеро Наяда
+    expected: ezero Nayada
+  - source: нос Никюп
+    expected: nos Nikyup
+  - source: рид Оборище
+    expected: rid Oborishte
+  - source: залив Олуша
+    expected: zaliv Olusha
+  - source: Оряховски възвишения
+    expected: Oryahovski vazvisheniya
+  - source: нунатак Памидово
+    expected: nunatak Pamidovo
+  - source: връх Парангалица
+    expected: vrah Parangalitsa
+  - source: Първомайски провлак
+    expected: Parvomayski provlak
+  - source: ледник Патлейна
+    expected: lednik Patleyna
+  - source: полуостров Перник
+    expected: poluostrov Pernik
+  - source: връх Петко Войвода
+    expected: vrah Petko Voyvoda
+  - source: остров Фанагория
+    expected: ostrov Fanagoriya
+  - source: нос Плас
+    expected: nos Plas
+  - source: Пресиянов рид
+    expected: Presiyanov rid
+  - source: връх Принсипе де Астуриас
+    expected: vrah Prinsipe de Asturias # wiki: vrah Príncipe de Asturias
+  - source: нунатак Ръченица
+    expected: nunatak Rachenitsa
+  - source: връх Райна Княгиня
+    expected: vrah Rayna Knyaginya
+  - source: Рид Ръжана
+    expected: Rid Razhana
+  - source: връх Ригс
+    expected: vrah Rigs
+  - source: остров Рогулят
+    expected: ostrov Rogulyat
+  - source: ледник Сабазий
+    expected: lednik Sabaziy
+  - source: ледник Съединение
+    expected: lednik Saedinenie
+  - source: нунатак Сенокос
+    expected: nunatak Senokos
+  - source: Сейдолски камък
+    expected: Seydolski kamak
+  - source: ледник Щерна
+    expected: lednik Shterna
+  - source: връх Шишман
+    expected: vrah Shishman
+  - source: ледник Сигмен
+    expected: lednik Sigmen
+  - source: Седловина Синитово
+    expected: Sedlovina Sinitovo
+  - source: Ледник Скаплизо
+    expected: Lednik Skaplizo
+  - source: залив Слънчев бряг
+    expected: zaliv Slanchev bryag
+  - source: остров Соатрис
+    expected: ostrov Soatris
+  - source: планина Софийски Университет
+    expected: planina Sofiyski Universitet
+  - source: ледник Сребърна
+    expected: lednik Srebarna
+  - source: Средногорски възвишения
+    expected: Srednogorski vazvisheniya
+  - source: Св. Евтимиев камък
+    expected: Sv. Evtimiev kamak
+  - source: база Св. Климент Охридски
+    expected: baza Sv. Kliment Ohridski
+  - source: връх Стъргел
+    expected: vrah Stargel
+  - source: нунатак Сторгозия
+    expected: nunatak Storgoziya # nunatak Storgozia according to wiki
+  - source: нунатак Сурвакари
+    expected: nunatak Survakari
+  - source: ледник Световрачене
+    expected: lednik Svetovrachene
+  - source: остров Теменуга
+    expected: ostrov Temenuga
+  - source: Тракийски възвишения
+    expected: Trakiyski vazvisheniya
+  - source: хълм Цамблак
+    expected: halm Tsamblak
+  - source: ледник Урдовиза
+    expected: lednik Urdoviza
+  - source: остров Вълчедръм
+    expected: ostrov Valchedram
+  - source: острови Вардим
+    expected: ostrovi Vardim
+  - source: Владигеров проток
+    expected: Vladigerov protok
+  - source: ледник Ябланица
+    expected: lednik Yablanitsa
+  - source: залив Ямфорина
+    expected: zaliv Yamforina
+  - source: Йовков нос
+    expected: Yovkov nos
+  - source: рид Заберново
+    expected: rid Zabernovo
+  - source: ледник Збелсурд
+    expected: lednik Zbelsurd
+  - source: Жефарович камък
+    expected: Zhefarovich kamak
+  - source: връх Зиези
+    expected: vrah Ziezi
+  - source: залив Златни пясъци
+    expected: zaliv Zlatni pyasatsi
+  - source: ледник Злокучене
+    expected: lednik Zlokuchene
+  - source: проток Злогош
+    expected: protok Zlogosh
+
+map:
+  characters:
+    '\u0410': 'A'
+    '\u0411': 'B'
+    '\u0412': 'V'
+    '\u0413': 'G'
+    '\u0414': 'D'
+    '\u0415': 'E'
+    '\u0416': 'Zh'
+    '\u0417': 'Z'
+    '\u0418': 'I'
+    '\u0419': 'Y'
+    '\u041a': 'K'
+    '\u041b': 'L'
+    '\u041c': 'M'
+    '\u041d': 'N'
+    '\u041e': 'O'
+    '\u041f': 'P'
+    '\u0420': 'R'
+    '\u0421': 'S'
+    '\u0422': 'T'
+    '\u0423': 'U'
+    '\u0424': 'F'
+    '\u0425': 'H'
+    '\u0426': 'Ts'
+    '\u0427': 'Ch'
+    '\u0428': 'Sh'
+    '\u0429': 'Sht'
+    '\u042a': 'A'
+    '\u042c': 'Y'
+    '\u042e': 'Yu'
+    '\u042f': 'Ya'
+
+    '\u0430': 'a'
+    '\u0431': 'b'
+    '\u0432': 'v'
+    '\u0433': 'g'
+    '\u0434': 'd'
+    '\u0435': 'e'
+    '\u0436': 'zh'
+    '\u0437': 'z'
+    '\u0438': 'i'
+    '\u0439': 'y'
+    '\u043a': 'k'
+    '\u043b': 'l'
+    '\u043c': 'm'
+    '\u043d': 'n'
+    '\u043e': 'o'
+    '\u043f': 'p'
+    '\u0440': 'r'
+    '\u0441': 's'
+    '\u0442': 't'
+    '\u0443': 'u'
+    '\u0444': 'f'
+    '\u0445': 'h'
+    '\u0446': 'ts'
+    '\u0447': 'ch'
+    '\u0448': 'sh'
+    '\u0449': 'sht'
+    '\u044a': 'a'
+    '\u044c': 'y'
+    '\u044e': 'yu'
+    '\u044f': 'ya'
+
+    # note 2
+    '\u046A': "U\u0306" # Ѫ
+    '\u046B': "u\u0306" # ѫ
+
+    # note[3]
+    '\u0462': "YE" # Ѣ
+    '\u0463': "ye" # ѣ

data/maps/{bas-rus-Cyrl-Latn-bss.yaml → bas-rus-Cyrl-Latn-2017-bss.yaml}

@@ -1,7 +1,7 @@
 ---
 authority_id: bas
-id: bss
-language: rus
+id: 2017-bss
+language: iso-639-2:rus
 source_script: Cyrl
 destination_script: Latn
 name: Streamlined Romanization of Russian Cyrillic -- Basic Streamlined System
@@ -41,23 +41,48 @@ tests:
 
       Н.В. Гоголь
     expected: |
-      Eh, troyka! ptitsa troyka, kto tebya vidumal? znat, u boykogo
-      naroda ti mogla tolko roditsya, v toy zemle, chto ne lyubit shutit, a
-      rovnem-gladnem razmetnulas na polsveta, da i stupay schitay versti,
-      poka ne zaryabit tebe v ochi. I ne hitriy, kazhis, dorozhniy snaryad,
-      ne zheleznim shvachen vintom, a naskoro zhivyem s odnim toporom da
-      dolotom sobral tebya yaroslavskiy rastoropniy muzhik. Ne v nemetskih
-      botfortah yamshchik: boroda da rukavitsi, i sidit chert znaet na chem;
-      a privstal, da zamahnulsya, da zatyanul pesnyu — koni vihrem, spitsi v
-      kolesah smeshalis v odin gladkiy krug, tolko drognula doroga, da
-      vskriknul v ispuge ostanovivshiysya peshehod — i von ona poneslas,
-      poneslas, poneslas!
+      Eh, troyka! ptitsa troyka, kto tebya vidumal? znat, u boykogo naroda
+      ti mogla tolko roditsya, v toy zemle, chto ne lyubit shutit, a
+      rovnem-gladnem razmetnulas na polsveta, da i stupay schitat versti, poka
+      ne zaryabit tebe v ochi. I ne hitriy, kazhis, dorozhniy snaryad, ne
+      zheleznim shvachen vintom, a naskoro zhivyem s odnim toporom da dolotom
+      snaryadil i sobral tebya yaroslavskiy rastoropniy muzhik. Ne v nemetskih
+      botfortah yamshchik: boroda da rukavitsi, i sidit chert znaet na chem; a
+      privstal, da zamahnulsya, da zatyanul pesnyu — koni vihrem, spitsi v
+      kolesah smeshalis v odin gladkiy krug, tolko drognula doroga, da vskriknul
+      v ispuge ostanovivshiysya peshehod — i von ona poneslas, poneslas,
+      poneslas!
 
       N.V. Gogol
 
+  - source: ЁЖ Ёж ёж
+    expected: EZH Ezh ezh
+  - source: Цветущий сад
+    expected: Tsvetushchiy sad
+  - source: Чувство юмора
+    expected: Chuvstvo yumora
+  - source: Широкий выбор
+    expected: Shirokiy vibor
+  - source: Все подъезды заблокированны
+    expected: Vse podezdi zablokirovanni
+  - source: Ожерелье
+    expected: Ozherelye
+  - source: Ручьи
+    expected: Ruchyi
+  - source: Каньон
+    expected: Kanyon
+  - source: Бельэтаж
+    expected: Belyetazh
+
 map:
+  rules:
+    - pattern: \u042c(?=[ЕеЁёИиОоЭэ]) # Ь (before Е, Ё, И, O, Э)
+      result: Y
+    - pattern: \u044c(?=[ЕеЁёИиОоЭэ]) # ь (before Е, Ё, И, O, Э)
+      result: y
+
   characters:
-    "\u0027": ""    # '
+    # "\u0027": ""    # '
     "\u0410": "A"    # А
     "\u0411": "B"    # Б
     "\u0412": "V"    # В
@@ -65,7 +90,7 @@ map:
     "\u0414": "D"    # Д
     "\u0401": "E"    # Ё
     "\u0415": "E"    # Е
-    "\u0416": "ZH"   # Ж
+    "\u0416": "Zh"   # Ж
     "\u0417": "Z"    # З
     "\u042D": "E"    # Э
     "\u0418": "I"    # И
@@ -82,23 +107,24 @@ map:
     "\u0423": "U"    # У
     "\u0424": "F"    # Ф
     "\u0425": "H"    # Х
-    "\u0426": "TS"   # Ц
-    "\u0427": "CH"   # Ч
-    "\u0428": "SH"   # Ш
-    "\u0429": "SHCH" # Щ
+    "\u0426": "Ts"   # Ц
+    "\u0427": "Ch"   # Ч
+    "\u0428": "Sh"   # Ш
+    "\u0429": "Shch" # Щ
     "\u042B": "I"    # Ы
-    "\u042F": "YA"   # Я
-    "\u042E": "YU"   # Ю
+    "\u042F": "Ya"   # Я
+    "\u042E": "Yu"   # Ю
 
     # Ь (before Е, Ё, И, O, Э)
-    "\u042c\u0401": "YE"    # Ё
-    "\u042c\u0415": "YE"    # Е
-    "\u042c\u0418": "YI"    # И
-    "\u042c\u041E": "YO"    # O
-    "\u042c\u0417": "YE"    # Э
+    # "\u042c\u0401": "YE"    # Ё
+    # "\u042c\u0415": "YE"    # Е
+    # "\u042c\u0418": "YI"    # И
+    # "\u042c\u041E": "YO"    # O
+    # "\u042c\u0417": "YE"    # Э
 
-    # Ъ (otherwise) -> (none)
+    # Ь (otherwise) -> (none)
     "\u042c": ""
+
     # Ъ  -> (none)
     "\u042a": ""
 
@@ -135,15 +161,14 @@ map:
     "\u044E": "yu"   # ю
 
     # ь (before е, ё, и, o, э)
-    "\u044c\u0435": "ye"    # ё
-    "\u044c\u0451": "ye"    # е
-    "\u044c\u0438": "yi"    # и
-    "\u044c\u006f": "yo"    # o
-    "\u044c\u044d": "ye"    # э
+    # "\u044c\u0435": "ye"    # ё
+    # "\u044c\u0451": "ye"    # е
+    # "\u044c\u0438": "yi"    # и
+    # "\u044c\u006f": "yo"    # o
+    # "\u044c\u044d": "ye"    # э
 
     # ь (otherwise) -> (none)
     "\u044c": ""
 
     # ъ -> (none)
     "\u044a": ""
-

data/maps/{bas-rus-Cyrl-Latn-oss.yaml → bas-rus-Cyrl-Latn-2017-oss.yaml}

@@ -1,7 +1,7 @@
 ---
 authority_id: bas
-id: oss
-language: rus
+id: 2017-oss
+language: iso-639-2:rus
 source_script: Cyrl
 destination_script: Latn
 name: Streamlined Romanization of Russian Cyrillic -- Optimized Streamlined System
@@ -26,8 +26,7 @@ notes:
     for Cyrillic х.
 
 tests:
-  - source: |
-      Эх, тройка! птица тройка, кто тебя выдумал? знать, у бойкого народа
+  - source: "Эх, тройка! птица тройка, кто тебя выдумал? знать, у бойкого народа
       ты могла только родиться, в той земле, что не любит шутить, а
       ровнем-гладнем разметнулась на полсвета, да и ступай считать версты, пока
       не зарябит тебе в очи. И не хитрый, кажись, дорожный снаряд, не
@@ -37,27 +36,48 @@ tests:
       привстал, да замахнулся, да затянул песню — кони вихрем, спицы в
       колесах смешались в один гладкий круг, только дрогнула дорога, да вскрикнул
       в испуге остановившийся пешеход — и вон она понеслась, понеслась,
-      понеслась!
+      понеслась!\nН.В. Гоголь"
 
-      Н.В. Гоголь
-    expected: |
-      `Eh, troyka! ptitsa troyka, kto tebya v`idumal? znat', u boykogo
+    expected: "`Eh, troyka! ptitsa troyka, kto tebya v`idumal? znat', u boykogo
       naroda t`i mogla tol'ko rodit'sya, v toy zemle, chto ne lyubit shutit',
-      a rovnem-gladnem razmetnulas' na polsveta, da i stupay schitay verst`i,
+      a rovnem-gladnem razmetnulas' na polsveta, da i stupay schitat' verst`i,
       poka ne zaryabit tebe v ochi. I ne hitr`iy, kazhis', dorozhn`iy
       snaryad, ne zhelezn`im shvachen vintom, a naskoro zhivy``em s odnim
-      toporom da dolotom sobral tebya yaroslavskiy rastoropn`iy muzhik. Ne v
+      toporom da dolotom snaryadil i sobral tebya yaroslavskiy rastoropn`iy muzhik. Ne v
       nemetskih botfortah yamshchik: boroda da rukavits`i, i sidit ch``ert
       znaet na ch``em; a privstal, da zamahnulsya, da zatyanul pesnyu — koni
       vihrem, spits`i v kolesah smeshalis' v odin gladkiy krug, tol'ko
       drognula doroga, da vskriknul v ispuge ostanovivshiysya peshehod — i
-      von ona poneslas', poneslas', poneslas'!
+      von ona poneslas', poneslas', poneslas'!\nN.V. Gogol'"
 
-      N.V. Gogol'
+  - source: ЁЖ Ёж ёж
+    expected: "``EZH ``Ezh ``ezh"
+  - source: Цветущий сад
+    expected: Tsvetushchiy sad
+  - source: Чувство юмора
+    expected: Chuvstvo yumora
+  - source: Широкий выбор
+    expected: Shirokiy v`ibor
+  - source: Все подъезды заблокированны
+    expected: Vse pod"ezd`i zablokirovann`i
+  - source: Ожерелье
+    expected: Ozherelye
+  - source: Ручьи
+    expected: Ruchyi
+  - source: Каньон
+    expected: Kanyon
+  - source: Бельэтаж
+    expected: Bely`etazh
 
 map:
+  rules:
+    - pattern: \u042c(?=[ЕеЁёИиОоЭэ]) # Ь (before Е, Ё, И, O, Э)
+      result: Y
+    - pattern: \u044c(?=[ЕеЁёИиОоЭэ]) # ь (before Е, Ё, И, O, Э)
+      result: y
+
   characters:
-    "\u0027": ""     # '
+    # "\u0027": ""     # '
     "\u0410": "A"    # А
     "\u0411": "B"    # Б
     "\u0412": "V"    # В
@@ -65,7 +85,7 @@ map:
     "\u0414": "D"    # Д
     "\u0401": "``E"  # Ё
     "\u0415": "E"    # Е
-    "\u0416": "ZH"   # Ж
+    "\u0416": "Zh"   # Ж
     "\u0417": "Z"    # З
     "\u042D": "`E"   # Э
     "\u0418": "I"    # И
@@ -82,25 +102,25 @@ map:
     "\u0423": "U"    # У
     "\u0424": "F"    # Ф
     "\u0425": "H"    # Х
-    "\u0426": "TS"   # Ц
-    "\u0427": "CH"   # Ч
-    "\u0428": "SH"   # Ш
-    "\u0429": "SHCH" # Щ
+    "\u0426": "Ts"   # Ц
+    "\u0427": "Ch"   # Ч
+    "\u0428": "Sh"   # Ш
+    "\u0429": "Shch" # Щ
     "\u042B": "`I"   # Ы
-    "\u042F": "YA"   # Я
-    "\u042E": "YU"   # Ю
+    "\u042F": "Ya"   # Я
+    "\u042E": "Yu"   # Ю
 
     # Ь (before Е, Ё, И, O, Э)
-    "\u042c\u0401": "Y``E"  # Ё
-    "\u042c\u0415": "YE"    # Е
-    "\u042c\u0418": "YI"    # И
-    "\u042c\u041E": "YO"    # O
-    "\u042c\u0417": "Y`E"    # Э
-
+    # "\u042c\u0401": "Y``e"  # Ё
+    # "\u042c\u0415": "Ye"    # Е
+    # "\u042c\u0418": "Yi"    # И
+    # "\u042c\u041E": "Yo"    # O
+    # "\u042c\u0417": "Y`e"    # Э
 
-    # Ъ (otherwise) -> " (or none)
+    # Ь (otherwise) -> ' (or none)
     "\u042c": "'"
-    # Ъ  -> ' (or none)
+
+    # Ъ  -> " (or none)
     "\u042a": '"'
 
     "\u0430": "a"    # а
@@ -136,14 +156,14 @@ map:
     "\u044E": "yu"   # ю
 
     # ь (before е, ё, и, o, э)
-    "\u044c\u0435": "ye"    # ё
-    "\u044c\u0451": "y``e"  # e
-    "\u044c\u0438": "yi"    # и
-    "\u044c\u006f": "yo"    # o
-    "\u044c\u044d": "y`e"    # э
+    # "\u044c\u0435": "ye"    # ё
+    # "\u044c\u0451": "y``e"  # e
+    # "\u044c\u0438": "yi"    # и
+    # "\u044c\u006f": "yo"    # o
+    # "\u044c\u044d": "y`e"    # э
 
-    # Ъ (otherwise) -> " (or none)
+    # ь (otherwise) -> ' (or none)
     "\u044c": "'"
 
-    # Ъ  -> ' (or none)
+    # ъ  -> " (or none)
     "\u044a": '"'