interscript 0.1.6 → 0.1.7

data/maps/bgnpcgn-ukr-Cyrl-Latn-1965.yaml

@@ -1,10 +1,14 @@
 ---
 authority_id: bgnpcgn
 id: 1965
-language: ukr
+language: iso-639-2:ukr
 source_script: Cyrl
 destination_script: Latn
 name: BGN/PCGN 1965 System
+alias:
+  ogc11122:
+    code: ukr_Cyrl2Latn_BGN_1965
+    description: Ukrainian Board on Geographic Names/Permanent Committee on Geographical Names for British Official Use(PCGN) 1965 System
 url: https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/816788/ROMANIZATION_OF_UKRAINIAN.pdf
 creation_date: 1947
 confirmation_date: 2019-06

data/maps/bgnpcgn-ukr-Cyrl-Latn-2019.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bgnpcgn
 id: 2019
-language: ukr
+language: iso-639-2:ukr
 source_script: Cyrl
 destination_script: Latn
 name: BGN/PCGN 2019 Agreement
@@ -113,96 +113,7 @@ tests:
     expected: Znamianka
 
 map:
-  rules:
-    - pattern: (?<=З|з)(Г|г)
-      result: gh
-    - pattern: (?<!\b\u2019)\b\u0404 # Є in initial position -> Ye
-      result: Ye
-    - pattern: (?<!\b\u2019)\b\u0454 # є in initial position -> ye
-      result: ye
-    - pattern: (?<!\b\u2019)\b\u0407 # Ї in initial position -> Yi
-      result: Yi
-    - pattern: (?<!\b\u2019)\b\u0457 # ї in initial position -> yi
-      result: yi
-    - pattern: (?<!\b\u2019)\b\u0419 # Й in initial position -> Y
-      result: "Y"
-    - pattern: (?<!\b\u2019)\b\u0419 # й in initial position -> y
-      result: "y"
-    - pattern: (?<!\b\u2019)\b\u042e # Ю in initial position -> Yu
-      result: Yu
-    - pattern: (?<!\b\u2019)\b\u044e # ю in initial position -> yu
-      result: yu
-    - pattern: (?<!\b\u2019)\b\u042f # Я in initial position -> Ya
-      result: Ya
-    - pattern: (?<!\b\u2019)\b\u044f # я in initial position -> ya
-      result: ya
-    - pattern: \b\u2019\b # remove ’
-      result: ""
+  inherit: un-ukr-Cyrl-Latn-2012
 
   characters:
-    "\u0410": "A"        # А
-    "\u0411": "B"        # Б
-    "\u0412": "V"        # В
-    "\u0413": "H"        # Г
-    "\u0490": "G"        # Ґ
-    "\u0414": "D"        # Д
-    "\u0415": "E"        # Е
-    "\u0404": "Ie"       # Є
-    "\u0416": "Zh"       # Ж
-    "\u0417": "Z"        # З
-    "\u0418": "Y"        # И
-    "\u0406": "I"        # І
-    "\u0407": "I"        # Ї
-    "\u0419": "I"        # Й
-    "\u041a": "K"        # К
-    "\u041b": "L"        # Л
-    "\u041c": "M"        # М
-    "\u041d": "N"        # Н
-    "\u041e": "O"        # О
-    "\u041f": "P"        # П
-    "\u0420": "R"        # Р
-    "\u0421": "S"        # С
-    "\u0422": "T"        # Т
-    "\u0423": "U"        # У
-    "\u0424": "F"        # Ф
-    "\u0425": "Kh"       # Х
-    "\u0426": "Ts"       # Ц
-    "\u0427": "Ch"       # Ч
-    "\u0428": "Sh"       # Ш
-    "\u0429": "Shch"     # Щ
-    "\u042e": "Iu"       # Ю
-    "\u042f": "Ia"       # Я
-    "\u042c": ""         # Ь
-    "\u0430": "a"        # а
-    "\u0431": "b"        # б
-    "\u0432": "v"        # в
-    "\u0433": "h"        # г
-    "\u0491": "g"        # ґ
-    "\u0434": "d"        # д
-    "\u0435": "e"        # е
-    "\u0454": "ie"       # є
-    "\u0436": "zh"       # ж
-    "\u0437": "z"        # з
-    "\u0438": "y"        # и
-    "\u0456": "i"        # і
-    "\u0457": "i"        # ї
-    "\u0439": "i"        # й
-    "\u043a": "k"        # к
-    "\u043b": "l"        # л
-    "\u043c": "m"        # м
-    "\u043d": "n"        # н
-    "\u043e": "o"        # о
-    "\u043f": "p"        # п
-    "\u0440": "r"        # р
-    "\u0441": "s"        # с
-    "\u0442": "t"        # т
-    "\u0443": "u"        # у
-    "\u0444": "f"        # ф
-    "\u0445": "kh"       # х
-    "\u0446": "ts"       # ц
-    "\u0447": "ch"       # ч
-    "\u0448": "sh"       # ш
-    "\u0449": "shch"     # щ
-    "\u044e": "iu"       # ю
-    "\u044f": "ia"       # я
-    "\u044c": ""         # Ь
+    "\u0027": ''     # ' ->

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

@@ -0,0 +1,127 @@
+---
+authority_id: bgnpcgn
+id: 1979
+language: iso-639-2:uzb
+source_script: Cyrl
+destination_script: Latn
+name: BGN/PCGN Romanization System -- Uzbek Cyrillic (1979)
+url: http://transliteration.eki.ee/pdf/Uzbek.pdf
+creation_date: 1979
+
+notes:
+  - At the beginning of a syllable, after a vowel, ъ or ь.
+
+tests:
+  # https://ru.wikipedia.org/wiki/Узбекский_язык
+  - source: Ўзбек ёзуви
+    expected: Ŭzbek yozuwi
+  - source: Ўзбек тили
+    expected: Ŭzbek tili
+  - source: катта
+    expected: katta
+  - source: куп
+    expected: kup
+  - source: кальта
+    expected: kalʼta
+  - source: Бори элға яхшилик қилғилки, мундин яхши йўқ Ким, дегайлар даҳр аро қолди фалондин яхшилик
+    expected: Bori elgha yakhshilik qilghilki, mundin yakhshi yŭq Kim, degaylar dahr aro qoldi falondin yakhshilik
+  - source: Бахр ул-худо
+    expected: Bakhr ul-khudo
+  - source: Рисале-йи маариф-и Шейбани
+    expected: Risale-yi maarif-i Sheybani
+  - source: Карами Хакка нихоят йукдур
+    expected: Karami Khakka nikhoyat yukdur
+  - source: Йахши
+    expected: Yakhshi
+  - source: Тутук белгись
+    expected: Tutuk belgisʼ
+  - source: |
+      Барча одамлар эркин, қадр-қиммат ва ҳуқуқларда тенг бўлиб туғиладилар.
+      Улар ақл ва виждон соҳибидирлар ва бир-бирлари ила биродарларча муомала қилишлари зарур.
+    expected: |
+      Barcha odamlar erkin, qadr-qimmat wa huquqlarda teng bŭlib tughiladilar.
+      Ular aql wa wizhdon sohibidirlar wa bir-birlari ila birodarlarcha muomala qilishlari zarur.
+  - source: ПАПАПАЧУКА Респект!
+    expected: PAPAPACHUKA Respekt!
+
+map:
+  rules:
+    # note[1]
+    - pattern: (?<=[АаЕеЁёИиОоУуЭэЮюЯяЪъЬь])\u0415
+      result: Ye
+    - pattern: (?<=[АаЕеЁёИиОоУуЭэЮюЯяЪъЬь])\u0435
+      result: ye
+
+  characters:
+    '\u0410': 'A'        # А
+    '\u0411': 'B'        # Б
+    '\u0412': 'W'        # В
+    '\u0413': 'G'        # Г
+    '\u0492': 'Gh'       # Ғ
+    '\u0414': 'D'        # Д
+    '\u0415': 'E'        # Е
+    '\u0401': 'Yo'       # Ё
+    '\u0416': 'Zh'       # Ж
+    '\u0417': 'Z'        # З
+    '\u0418': 'I'        # И
+    '\u0419': 'Y'        # Й
+    '\u041A': 'K'        # К
+    '\u049A': 'Q'        # Қ
+    '\u041B': 'L'        # Л
+    '\u041C': 'M'        # М
+    '\u041D': 'N'        # Н
+    '\u041E': 'O'        # О
+    '\u041F': 'P'        # П
+    '\u0420': 'R'        # Р
+    '\u0421': 'S'        # С
+    '\u0422': 'T'        # Т
+    '\u0423': 'U'        # У
+    '\u040E': 'Ŭ'        # Ў
+    '\u0424': 'F'        # Ф
+    '\u0425': 'Kh'       # Х
+    '\u04B2': 'H'        # Ҳ
+    '\u0426': 'Ts'       # Ц
+    '\u0427': 'Ch'       # Ч
+    '\u0428': 'Sh'       # Ш
+    '\u042a': "\u02BC"   # Ъ
+    '\u042c': "\u02BC"   # Ь
+    '\u042D': 'E'        # Э
+    '\u042E': 'Yu'       # Ю
+    '\u042F': 'Ya'       # Я
+
+    '\u0430': 'a'        # а
+    '\u0431': 'b'        # б
+    '\u0432': 'w'        # в
+    '\u0433': 'g'        # г
+    '\u0493': 'gh'       # ғ
+    '\u0434': 'd'        # д
+    '\u0435': 'e'        # e
+    '\u0451': 'yo'       # ё
+    '\u0436': 'zh'       # ж
+    '\u0437': 'z'        # з
+    '\u0438': 'i'        # и
+    '\u0439': 'y'        # й
+    '\u043A': 'k'        # к
+    '\u049B': 'q'        # қ
+    '\u043B': 'l'        # л
+    '\u043C': 'm'        # м
+    '\u043D': 'n'        # н
+    '\u043E': 'o'        # о
+    '\u043F': 'p'        # п
+    '\u0440': 'r'        # р
+    '\u0441': 's'        # с
+    '\u0442': 't'        # т
+    '\u0443': 'u'        # у
+    '\u045E': 'ŭ'        # ў
+    '\u0444': 'f'        # ф
+    '\u0445': 'kh'       # х
+    '\u04B3': 'h'        # ҳ
+    '\u0446': 'ts'       # ц
+    '\u0447': 'ch'       # ч
+    '\u0448': 'sh'       # ш
+    '\u044a': "\u02BC"   # ъ
+    '\u044c': "\u02BC"   # ь
+    '\u044D': 'e'        # э
+    '\u044F': 'ya'       # я
+    '\u044E': 'yu'       # ю
+

data/maps/bgnpcgn-uzb-Cyrl-Latn-2000.yaml

@@ -0,0 +1,82 @@
+---
+authority_id: bgnpcgn
+id: 2000
+language: iso-639-2:uzb
+source_script: Cyrl
+destination_script: Latn
+name: TABLE OF CORRESPONDENCES CYRILLIC - ROMAN BGN/PCGN 2000 Agreement
+description: |
+  In 1995, the Uzbek government adopted the Roman alphabet to replace the existing Cyrillic alphabet.
+  The presentation below provides a table of correspondences between the former Cyrillic alphabet and the
+  current Roman alphabet. When Uzbek Roman-alphabet spellings are not available, this table can be used to
+  convert Uzbek Cyrillic spellings. This table of correspondences supersedes the BGN/PCGN 1979 romanization
+  system for Uzbek.
+url: http://transliteration.eki.ee/pdf/Uzbek.pdf
+creation_date: 2000
+confirmation_date: 2017-11
+
+notes:
+  - The letter sequence ye is used initially, after the vowel characters 1, 6, 7, 10, 16, 21, 29, 30, 31, and 32, and after characters 11 and 28.
+  - The Unicode encoding of the apostrophe appearing in rows 27 and 28 is U+2019. The inverted apostrophe appearing in rows 32 (o‘) and 34 (g‘) is U+2018.
+  - The Roman-script columns show only lowercase forms but, when applying the table, uppercase and lowercase Roman letters as appropriate should be used.
+
+tests:
+  # https://ru.wikipedia.org/wiki/Узбекский_язык
+  - source: Ўзбек ёзуви
+    expected: O‘zbek yozuwi
+  - source: Ўзбек тили
+    expected: O‘zbek tili
+  - source: катта
+    expected: katta
+  - source: куп
+    expected: kup
+  - source: кальта
+    expected: kal’ta
+  - source: Бори элға яхшилик қилғилки, мундин яхши йўқ Ким, дегайлар даҳр аро қолди фалондин яхшилик
+    expected: Bori elg‘a yaxshilik qilg‘ilki, mundin yaxshi yo‘q Kim, degaylar dahr aro qoldi falondin yaxshilik
+  - source: Бахр ул-худо
+    expected: Baxr ul-xudo
+  - source: Рисале-йи маариф-и Шейбани
+    expected: Risale-yi maarif-i Sheybani
+  - source: Карами Хакка нихоят йукдур
+    expected: Karami Xakka nixoyat yukdur
+  - source: Йахши
+    expected: Yaxshi
+  - source: Тутук белгись
+    expected: Tutuk belgis’
+  - source: |
+      Барча одамлар эркин, қадр-қиммат ва ҳуқуқларда тенг бўлиб туғиладилар.
+      Улар ақл ва виждон соҳибидирлар ва бир-бирлари ила биродарларча муомала қилишлари зарур.
+    expected: |
+      Barcha odamlar erkin, qadr-qimmat wa huquqlarda teng bo‘lib tug‘iladilar.
+      Ular aql wa wijdon sohibidirlar wa bir-birlari ila birodarlarcha muomala qilishlari zarur.
+  - source: ПАПАПАЧУКА Респект!
+    expected: PAPAPACHUKA Respekt!
+
+map:
+  inherit: bgnpcgn-uzb-Cyrl-Latn-1979
+
+  rules:
+    # note[1]
+    - pattern: (?<=[АаЕеЁёИиОоУуЭэЮюЯяЙйЬь])\u0415
+      result: Ye
+    - pattern: (?<=[АаЕеЁёИиОоУуЭэЮюЯяЙйЬь])\u0435
+      result: ye
+
+  characters:
+    '\u0412': 'V'        # В
+    '\u0492': "G\u2018"  # Ғ
+    '\u0416': 'J'        # Ж
+    '\u040E': "O\u2018"  # Ў
+    '\u0425': 'X'        # Х
+    '\u042a': "\u2019"   # Ъ note[2]
+    '\u042c': "\u2019"   # Ь note[2]
+
+    '\u0432': 'w'        # в
+    '\u0493': "g\u2018"  # ғ
+    '\u0436': 'j'        # ж
+    '\u045E': "o\u2018"  # ў
+    '\u0445': 'x'        # х
+    '\u044a': "\u2019"   # ъ note[2]
+    '\u044c': "\u2019"   # ь note[2]
+

data/maps/bgnpcgn-zho-Hans-Latn-1979.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bgnpcgn
 id: pinyin
-language: zho
+language: iso-639-2:zho
 source_script: Hans
 destination_script: Latn
 name: ROMANIZATION OF CHINESE -- BGN/PCGN 1979 AGREEMENT

data/maps/bis-asm-Beng-Latn-13194-1991.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: asm
+language: iso-639-2:asm
 source_script: Beng
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Assamese Romanization

data/maps/bis-ben-Beng-Latn-13194-1991.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: ben
+language: iso-639-2:ben
 source_script: Beng
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Bengali Romanization

data/maps/bis-dev-Deva-Latn-13194-1991.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: dev
+language: iso-639-2:dev
 source_script: Deva
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Devanagri Romanization

data/maps/{bis-gjr-Gujr-Latn-13194-1991.yaml → bis-guj-Gujr-Latn-13194-1991.yaml}

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: gjr
+language: iso-639-2:guj
 source_script: Gujr
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Gujrati Romanization
@@ -43,6 +43,8 @@ tests:
     expected: "srdār pṭēlē nkkī kryuṃ htuṃ kē kāśmīr bhārtnŏ hissŏ bnśē; 91 vrṣ phēlāṃ lāhŏr jēlmāṃ bhūkhhḍtāḷ drmiyān śhīd thyā htā jtīn dās"
   - source: "કોરોના પ્રોટોકોલ વચ્ચે આજે મેડિકલ પ્રવેશ પરીક્ષા લેવાશેઃ એન્ટ્રી ટચ ફ્રી રહેશે, એડમિટ કાર્ડ બાર કોડથી ચેક થશે"
     expected: "kŏrŏnā prŏṭŏkŏl vccē ājē mēḍikl prvēś prīkṣā lēvāśēḥ ēnṭrī ṭc phrī rhēśē, ēḍmiṭ kārḍ bār kŏḍthī cēk thśē"
+  - source: "૮૪૬૬૫૪૧૬૪૬૫૧"
+    expected: "846654164651"
 
 
 
@@ -163,4 +165,17 @@ map:
     '્': ''
     '઼': ''
     '।': '.'
-    "‍": ''# Used for joining
+    "‍": ''# Used for joining
+
+    # digits
+
+    '૦': '0'
+    '૧': '1'
+    '૨': '2'
+    '૩': '3'
+    '૪': '4'
+    '૫': '5'
+    '૬': '6'
+    '૭': '7'
+    '૮': '8'
+    '૯': '9'

data/maps/bis-knd-Knda-Latn-13194-1991.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: knd
+language: iso-639-2:knd
 source_script: Knda
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Kannada Romanization

data/maps/bis-mlm-Mlym-Latn-13194-1991.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: mlm
+language: iso-639-2:mlm
 source_script: Mlym
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Malayalam Romanization

data/maps/bis-ori-Orya-Latn-13194-1991.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: ori
+language: iso-639-2:ori
 source_script: Orya
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Oriya Romanization

data/maps/bis-pnj-Guru-Latn-13194-1991.yaml

@@ -1,7 +1,7 @@
 ---
 authority_id: bis
 id: 1991
-language: pnj
+language: iso-639-2:pnj
 source_script: Guru
 destination_script: Latn
 name: Indian script code for information interchange - ISCII - Panjabi(Gurmukhi) Romanization