310 lines
10 KiB
Plaintext
310 lines
10 KiB
Plaintext
# © 2016 and later: Unicode, Inc. and others.
|
||
# License & terms of use: http://www.unicode.org/copyright.html
|
||
# Generated using tools/cldr/cldr-to-icu/build-icu-data.xml
|
||
#
|
||
# File: tk_Cyrl_tk_BGN.txt
|
||
# Generated from CLDR
|
||
#
|
||
|
||
#
|
||
########################################################################
|
||
# BGN/PCGN 1979 System
|
||
#
|
||
# The BGN/PCGN system for Turkmen was designed for use in
|
||
# romanizing names written in the Turkmen alphabet.
|
||
# The Turkmen alphabet contains five letters not present
|
||
# in the Russian alphabet: Җҗ, Ңң, Өө, Үү, and Әә.
|
||
#
|
||
# The Turkmen Cyrillic Alphabet as defined by the BGN (Page 103):
|
||
#
|
||
# АБВГДЕЁЖҖЗИЙКЛМНҢОӨПРСТУҮФХЦЧШЩЪЫЬЭӘЮЯ
|
||
# абвгдеёжзҗийклмнңоөпрстуүфхцчшщъыьэәюя
|
||
#
|
||
# Originally prepared by Michael Everson <everson@evertype.com>
|
||
########################################################################
|
||
#
|
||
# MINIMAL FILTER: Turkmen-Latin
|
||
#
|
||
# :: [АБВГДЕЁЖҖЗИЙКЛМНҢОӨПРСТУҮФХЦЧШЩЪЫЬЭӘЮЯабвгдеёжзҗийклмнңоөпрстуүфхцчшщъыьэәюя] ;
|
||
:: NFD (NFC) ;
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
########################################################################
|
||
#
|
||
# Define All Transformation Variables
|
||
#
|
||
########################################################################
|
||
#
|
||
$prime = ʹ ;
|
||
$doublePrime = ʺ ;
|
||
$upperConsonants = [БВГДЖҖЗЙКЛМНҢПРСТФХЦЧШЩЪЬ] ;
|
||
$lowerConsonants = [бвгджҗзйклмнңпрстфхцчшщъь] ;
|
||
$consonants = [$upperConsonants $lowerConsonants] ;
|
||
$upperVowels = [АЕЁИОӨУҮЫЭӘЮЯ] ;
|
||
$lowerVowels = [аеёиоөуүыэәюя] ;
|
||
$vowels = [$upperVowels $lowerVowels] ;
|
||
$lower = [$lowerConsonants $lowerVowels] ;
|
||
#
|
||
#
|
||
# Use this $wordBoundary until bug 2034 is fixed in ICU:
|
||
# http://bugs.icu-project.org/cgi-bin/icu-bugs/transliterate?id=2034;expression=boundary;user=guest
|
||
#
|
||
$wordBoundary = [^[:L:][:M:][:N:]] ;
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
########################################################################
|
||
#
|
||
# Start of Alphabetic Transformations
|
||
#
|
||
########################################################################
|
||
#
|
||
А → A ; # CYRILLIC CAPITAL LETTER A
|
||
а → a ; # CYRILLIC SMALL LETTER A
|
||
Б → B ; # CYRILLIC CAPITAL LETTER BE
|
||
б → b ; # CYRILLIC SMALL LETTER BE
|
||
В → W ; # CYRILLIC CAPITAL LETTER VE
|
||
в → w ; # CYRILLIC SMALL LETTER VE
|
||
Г → G ; # CYRILLIC CAPITAL LETTER GHE
|
||
г → g ; # CYRILLIC SMALL LETTER GHE
|
||
Д → D ; # CYRILLIC CAPITAL LETTER DE
|
||
д → d ; # CYRILLIC SMALL LETTER DE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 104 Rule 1:
|
||
#
|
||
# The character e should be romanized ye initially, after the vowel
|
||
# characters a, e, ё, и, о, ө, у, ү, ы, э, ю, and я, and after й, ъ, and ь.
|
||
# In all other instances, it should be romanized e.
|
||
#
|
||
########################################################################
|
||
#
|
||
Е}[[$upperVowels - [Ә]] [ЙЪЬ]] → YE ; # CYRILLIC CAPITAL LETTER IE
|
||
Е}[[$lowerVowels - [ә]] [йъь]] → Ye ; # CYRILLIC CAPITAL LETTER IE
|
||
$wordBoundary{Е → Ye ; # CYRILLIC CAPITAL LETTER IE
|
||
Е → E ; # CYRILLIC CAPITAL LETTER IE
|
||
е}[[$upperVowels - [Ә]] [$lowerVowels - [ә]] [ЙйЪъЬь]] → ye ; # CYRILLIC SMALL LETTER IE
|
||
$wordBoundary{е → ye ; # CYRILLIC SMALL LETTER IE
|
||
е → e ; # CYRILLIC SMALL LETTER IE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End of Rule 1
|
||
#
|
||
########################################################################
|
||
#
|
||
Ё} $lower → Yo ; # CYRILLIC CAPITAL LETTER IO
|
||
Ё → YO ; # CYRILLIC CAPITAL LETTER IO
|
||
ё → yo ; # CYRILLIC SMALL LETTER IO
|
||
Ж} $lower → Zh ; # CYRILLIC CAPITAL LETTER ZHE
|
||
Ж → ZH ; # CYRILLIC CAPITAL LETTER ZHE
|
||
ж → zh ; # CYRILLIC SMALL LETTER ZHE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 104 Rule 2
|
||
#
|
||
# The character sequences зх, нг, сх, and цх may be romanized z·h,
|
||
# n·g, s·h, and ts·h in order to differentiate those romanizations form
|
||
# the digraphs zh, ng, sh, and the letter sequence tsh, which are used
|
||
# to render the characters ж, ң, ш, and the character sequence тш.
|
||
#
|
||
########################################################################
|
||
#
|
||
ЗХ → Z·H ; # CYRILLIC CAPITAL LETTER ZE
|
||
Зх → Z·h ; # CYRILLIC CAPITAL LETTER ZE
|
||
зх → z·h ; # CYRILLIC SMALL LETTER ZE
|
||
З → Z ; # CYRILLIC CAPITAL LETTER ZE
|
||
з → z ; # CYRILLIC SMALL LETTER ZE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Rule 2
|
||
#
|
||
########################################################################
|
||
#
|
||
И → И ; # CYRILLIC CAPITAL LETTER I
|
||
и → и ; # CYRILLIC SMALL LETTER I
|
||
Й → Y ; # CYRILLIC CAPITAL LETTER I
|
||
й → y ; # CYRILLIC SMALL LETTER I
|
||
К → K ; # CYRILLIC CAPITAL LETTER KA
|
||
к → k ; # CYRILLIC SMALL LETTER KA
|
||
Л → L ; # CYRILLIC CAPITAL LETTER EL
|
||
л → l ; # CYRILLIC SMALL LETTER EL
|
||
М → M ; # CYRILLIC CAPITAL LETTER EM
|
||
м → m ; # CYRILLIC SMALL LETTER EM
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 104 Rule 2
|
||
#
|
||
# нг becomes n·g
|
||
#
|
||
########################################################################
|
||
#
|
||
НГ → N·G ; # CYRILLIC CAPITAL LETTER EN
|
||
Нг → N·g ; # CYRILLIC CAPITAL LETTER EN
|
||
нг → n·g ; # CYRILLIC SMALL LETTER EN
|
||
Н → N ; # CYRILLIC CAPITAL LETTER EN
|
||
н → n ; # CYRILLIC SMALL LETTER EN
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Rule 2
|
||
#
|
||
########################################################################
|
||
#
|
||
Ң} $lower → Ng ; # CYRILLIC CAPITAL LETTER EN WITH DESCENDER
|
||
Ң → NG ; # CYRILLIC CAPITAL LETTER EN WITH DESCENDER
|
||
ң → ng ; # CYRILLIC SMALL LETTER EN WITH DESCENDER
|
||
О → O ; # CYRILLIC CAPITAL LETTER O
|
||
о → o ; # CYRILLIC SMALL LETTER O
|
||
Ө → Ö ; # CYRILLIC CAPITAL LETTER BARRED O
|
||
ө → ö ; # CYRILLIC SMALL LETTER BARRED O
|
||
П → P ; # CYRILLIC CAPITAL LETTER PE
|
||
п → p ; # CYRILLIC SMALL LETTER PE
|
||
Р → R ; # CYRILLIC CAPITAL LETTER ER
|
||
р → r ; # CYRILLIC SMALL LETTER ER
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 104 Rule 2
|
||
#
|
||
# сх becomes s·h
|
||
#
|
||
########################################################################
|
||
#
|
||
СХ → S·H ; # CYRILLIC CAPITAL LETTER ES
|
||
Сх → S·h ; # CYRILLIC CAPITAL LETTER ES
|
||
сх → s·h ; # CYRILLIC SMALL LETTER ES
|
||
С → S ; # CYRILLIC CAPITAL LETTER ES
|
||
с → s ; # CYRILLIC SMALL LETTER ES
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Rule 2
|
||
#
|
||
########################################################################
|
||
#
|
||
Т → T ; # CYRILLIC CAPITAL LETTER TE
|
||
т → t ; # CYRILLIC SMALL LETTER TE
|
||
У → U ; # CYRILLIC CAPITAL LETTER U
|
||
у → u ; # CYRILLIC SMALL LETTER U
|
||
Ү → Ü ; # CYRILLIC CAPITAL LETTER STRAIGHT U
|
||
ү → ü ; # CYRILLIC SMALL LETTER STRAIGHT U
|
||
Ф → F ; # CYRILLIC CAPITAL LETTER EF
|
||
ф → f ; # CYRILLIC SMALL LETTER EF
|
||
Х → H ; # CYRILLIC CAPITAL LETTER HA
|
||
х → h ; # CYRILLIC SMALL LETTER HA
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 104 Rule 2
|
||
#
|
||
# цх becomes ts·h
|
||
#
|
||
########################################################################
|
||
#
|
||
ЦХ → TS·H ; # CYRILLIC CAPITAL LETTER GHE
|
||
Цх → Ts·h ; # CYRILLIC CAPITAL LETTER GHE
|
||
цх → ts·h ; # CYRILLIC SMALL LETTER GHE
|
||
Ц} $lower → Ts ; # CYRILLIC CAPITAL LETTER TSE
|
||
Ц → TS ; # CYRILLIC CAPITAL LETTER TSE
|
||
ц → ts ; # CYRILLIC SMALL LETTER TSE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Rule 2
|
||
#
|
||
########################################################################
|
||
#
|
||
Ч} $lower → Ch ; # CYRILLIC CAPITAL LETTER CHE
|
||
Ч → CH ; # CYRILLIC CAPITAL LETTER CHE
|
||
ч → ch ; # CYRILLIC SMALL LETTER CHE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# Implied rule from BGN Russian-Latin transliteration (Page 94 Note 3.6).
|
||
#
|
||
# шч becomes sh·ch
|
||
#
|
||
########################################################################
|
||
#
|
||
ШЧ → SH·CH ; # CYRILLIC CAPITAL LETTER SHA
|
||
Шч → Sh·ch ; # CYRILLIC CAPITAL LETTER SHA
|
||
шч → sh·ch ; # CYRILLIC SMALL LETTER SHA
|
||
Ш} $lower → Sh ; # CYRILLIC CAPITAL LETTER SHA
|
||
Ш → SH ; # CYRILLIC CAPITAL LETTER SHA
|
||
ш → sh ; # CYRILLIC SMALL LETTER SHA
|
||
Щ} $lower → Shch ; # CYRILLIC CAPITAL LETTER SHCHA
|
||
Щ → SHCH ; # CYRILLIC CAPITAL LETTER SHCHA
|
||
щ → shch ; # CYRILLIC SMALL LETTER SHCHA
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Implied rule
|
||
#
|
||
########################################################################
|
||
#
|
||
Ъ → $doublePrime ; # CYRILLIC CAPITAL LETTER HARD SIGN
|
||
ъ → $doublePrime ; # CYRILLIC SMALL LETTER HARD SIGN
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 104 Note 3
|
||
#
|
||
# The character Ыы may be romanized Ɨɨ instead of Yy, if so desired.
|
||
#
|
||
########################################################################
|
||
#
|
||
Ы → Y ; # CYRILLIC CAPITAL LETTER YERU
|
||
ы → y ; # CYRILLIC SMALL LETTER YERU
|
||
#
|
||
#
|
||
# Alternative rule to implement the option described here. To apply
|
||
# uncomment the following by removing the '#' mark at the start of the
|
||
# line and insert before the two rule lines above.
|
||
#
|
||
#Ы → Ɨ ; # CYRILLIC CAPITAL LETTER YERU
|
||
#ы → ɨ ; # CYRILLIC SMALL LETTER YERU
|
||
#
|
||
########################################################################
|
||
#
|
||
# End BGN Page 104 Note 2
|
||
#
|
||
########################################################################
|
||
#
|
||
Ь → $prime ; # CYRILLIC CAPITAL LETTER SOFT SIGN
|
||
ь → $prime ; # CYRILLIC SMALL LETTER SOFT SIGN
|
||
Э → E ; # CYRILLIC CAPITAL LETTER E
|
||
э → e ; # CYRILLIC SMALL LETTER E
|
||
Ә → Ä ; # CYRILLIC CAPITAL LETTER SCHWA
|
||
ә → ä ; # CYRILLIC SMALL LETTER SCHWA
|
||
Ю} $lower → Yu ; # CYRILLIC CAPITAL LETTER YU
|
||
Ю → YU ; # CYRILLIC CAPITAL LETTER YU
|
||
ю → yu ; # CYRILLIC SMALL LETTER YU
|
||
Я} $lower → Ya ; # CYRILLIC CAPITAL LETTER YA
|
||
Я → YA ; # CYRILLIC CAPITAL LETTER YA
|
||
я → ya ; # CYRILLIC SMALL LETTER YA
|
||
#
|
||
#
|
||
########################################################################
|
||
|