273 lines
9.0 KiB
Plaintext
273 lines
9.0 KiB
Plaintext
# ***************************************************************************
|
||
# Copyright (C) 2016 and later: Unicode, Inc. and others.
|
||
# License & terms of use: http://www.unicode.org/copyright.html
|
||
# *
|
||
# * Copyright (C) 2004-2016, International Business Machines
|
||
# * Corporation; Unicode, Inc.; and others. All Rights Reserved.
|
||
# *
|
||
# ***************************************************************************
|
||
# File: uz_Cyrl_uz_BGN.txt
|
||
# Generated from CLDR
|
||
#
|
||
|
||
#
|
||
########################################################################
|
||
# BGN/PCGN 1979 System
|
||
#
|
||
# The BGN/PCGN system for Uzbek was designed for use in
|
||
# romanizing names written in the Uzbek alphabet.
|
||
# The Uzbek alphabet contains four letters not present
|
||
# in the Russian alphabet: Ўў, Ққ, Ғғ, and Ҳҳ.
|
||
#
|
||
# The Uzbek Alphabet as defined by the BGN (Page 107):
|
||
#
|
||
# АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЪЬЭЮЯЎҚҒҲ
|
||
# абвгдеёжзийклмнопрстуфхцчшъьэюяўқғҳ
|
||
#
|
||
# Originally prepared by Michael Everson <everson@evertype.com>
|
||
########################################################################
|
||
#
|
||
# MINIMAL FILTER: Uzbek-Latin
|
||
#
|
||
:: [АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЪЬЭЮЯЎҚҒҲабвгдеёжзийклмнопрстуфхцчшъьэюяўқғҳ] ;
|
||
:: NFD (NFC) ;
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
########################################################################
|
||
#
|
||
# Define All Transformation Variables
|
||
#
|
||
########################################################################
|
||
#
|
||
$prime = ʹ ;
|
||
$doublePrime = ʺ ;
|
||
$upperConsonants = [БВГДЖЗЙКЛМНПРСТФХЦЧШЪЬҚҒҲ] ;
|
||
$lowerConsonants = [бвгджзйклмнпрстфхцчшъьқғҳ] ;
|
||
$consonants = [$upperConsonants $lowerConsonants] ;
|
||
$upperVowels = [АЕЁИОУЭЮЯЎ] ;
|
||
$lowerVowels = [аеёиоуэюяў] ;
|
||
$vowels = [$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:]] ;
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
########################################################################
|
||
#
|
||
# Rules moved to front to avoid masking
|
||
#
|
||
########################################################################
|
||
#
|
||
$lowerVowels ы → y ;
|
||
$upperVowels[Ыы] → Y ;
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# Start of Alphabetic Transformations
|
||
#
|
||
########################################################################
|
||
#
|
||
А → A ; # CYRILLIC CAPITAL LETTER A
|
||
а → a ; # CYRILLIC SMALL LETTER A
|
||
Б → B ; # CYRILLIC CAPITAL LETTER BE
|
||
б → b ; # CYRILLIC SMALL LETTER BE
|
||
В → W ; # CYRILLIC CAPITAL LETTER VE
|
||
в → w ; # CYRILLIC SMALL LETTER VE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 108 Rule 2
|
||
#
|
||
# The character sequences гҳ, кҳ, сҳ, and цҳ may be romanized g·h,
|
||
# k·h, s·h, and ts·h in order to differentiate those romanizations from
|
||
# the digraphs gh, kh, sh, and the letter sequence tsh, which are used
|
||
# to render the chаracters г, х, ш, and the character sequence тш.
|
||
#
|
||
########################################################################
|
||
#
|
||
ГҲ → G·H ; # CYRILLIC CAPITAL LETTER GHE
|
||
Гҳ → G·h ; # CYRILLIC CAPITAL LETTER GHE
|
||
гҳ → g·h ; # CYRILLIC SMALL LETTER GHE
|
||
Г → G ; # CYRILLIC CAPITAL LETTER GHE
|
||
г → g ; # CYRILLIC SMALL LETTER GHE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Rule 2
|
||
#
|
||
########################################################################
|
||
#
|
||
Д → D ; # CYRILLIC CAPITAL LETTER DE
|
||
д → d ; # CYRILLIC SMALL LETTER DE
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 108 Rule 1:
|
||
#
|
||
# The character e should be romanized ye initially, after the vowel
|
||
# characters a, e, ё, и, о, у, э, ю, я, and ў, and after й and ь.
|
||
# In all other instances, it should be romanized e.
|
||
#
|
||
########################################################################
|
||
#
|
||
Е}[$upperVowels [ЙЬ]] → YE ; # CYRILLIC CAPITAL LETTER IE
|
||
Е}[$lowerVowels [йь]] → Ye ; # CYRILLIC CAPITAL LETTER IE
|
||
$wordBoundary{Е → Ye ; # CYRILLIC CAPITAL LETTER IE
|
||
Е → E ; # CYRILLIC CAPITAL LETTER IE
|
||
е}[$upperVowels $lowerVowels [ЙйЬь]] → 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 → J ; # CYRILLIC CAPITAL LETTER ZHE
|
||
ж → j ; # CYRILLIC SMALL LETTER ZHE
|
||
З → Z ; # CYRILLIC CAPITAL LETTER ZE
|
||
з → z ; # CYRILLIC SMALL LETTER ZE
|
||
И → I ; # CYRILLIC CAPITAL LETTER I
|
||
и → i ; # CYRILLIC SMALL LETTER I
|
||
Й → Y ; # CYRILLIC CAPITAL LETTER I
|
||
й → y ; # CYRILLIC SMALL LETTER I
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 108 Rule 2
|
||
#
|
||
# кҳ becomes k·h
|
||
#
|
||
########################################################################
|
||
#
|
||
КҲ → K·H ; # CYRILLIC CAPITAL LETTER KA
|
||
Кҳ → K·h ; # CYRILLIC CAPITAL LETTER KA
|
||
кҳ → k·h ; # CYRILLIC SMALL LETTER KA
|
||
К → K ; # CYRILLIC CAPITAL LETTER KA
|
||
к → k ; # CYRILLIC SMALL LETTER KA
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Rule 2
|
||
#
|
||
########################################################################
|
||
#
|
||
Л → L ; # CYRILLIC CAPITAL LETTER EL
|
||
л → l ; # CYRILLIC SMALL LETTER EL
|
||
М → M ; # CYRILLIC CAPITAL LETTER EM
|
||
м → m ; # CYRILLIC SMALL LETTER EM
|
||
Н → N ; # CYRILLIC CAPITAL LETTER EN
|
||
н → n ; # CYRILLIC SMALL LETTER EN
|
||
О → O ; # CYRILLIC CAPITAL LETTER O
|
||
о → o ; # CYRILLIC SMALL LETTER 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 108 Rule 2
|
||
#
|
||
# сҳ becomes s·h
|
||
#
|
||
########################################################################
|
||
#
|
||
СҲ → S·H ; # CYRILLIC CAPITAL LETTER ES
|
||
Сҳ → S·h ; # CYRILLIC CAPITAL LETTER ES
|
||
сҳ → 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 → Ū ; # CYRILLIC CAPITAL LETTER U
|
||
у → ū ; # CYRILLIC SMALL LETTER U
|
||
Ф → F ; # CYRILLIC CAPITAL LETTER EF
|
||
ф → f ; # CYRILLIC SMALL LETTER EF
|
||
Х} $lower → Kh ; # CYRILLIC CAPITAL LETTER HA
|
||
Х → KH ; # CYRILLIC CAPITAL LETTER HA
|
||
х → kh ; # CYRILLIC SMALL LETTER HA
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 108 Rule 2
|
||
#
|
||
# цҳ becomes ts·h
|
||
#
|
||
########################################################################
|
||
#
|
||
ЦҲ → TS·H ; # CYRILLIC CAPITAL LETTER GHE
|
||
Цҳ → Ts·h ; # CYRILLIC CAPITAL LETTER GHE
|
||
цҳ → 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
|
||
Ш} $lower → Sh ; # CYRILLIC CAPITAL LETTER SHA
|
||
Ш → SH ; # CYRILLIC CAPITAL LETTER SHA
|
||
ш → sh ; # CYRILLIC SMALL LETTER SHA
|
||
Ъ → $prime ; # CYRILLIC CAPITAL LETTER HARD SIGN
|
||
ъ → $prime ; # CYRILLIC SMALL LETTER HARD SIGN
|
||
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
|
||
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
|
||
Ў → Ŭ ; # CYRILLIC CAPITAL LETTER SHORT U
|
||
ў → ŭ ; # CYRILLIC SMALL LETTER SHORT U
|
||
Қ → Q ; # CYRILLIC CAPITAL LETTER KA WITH DESCENDER
|
||
қ → q ; # CYRILLIC SMALL LETTER KA WITH DESCENDER
|
||
Ғ} $lower → Gh ; # CYRILLIC CAPITAL LETTER GHE WITH STROKE
|
||
Ғ → GH ; # CYRILLIC CAPITAL LETTER GHE WITH STROKE
|
||
ғ → gh ; # CYRILLIC SMALL LETTER GHE WITH STROKE
|
||
Ҳ → H ; # CYRILLIC CAPITAL LETTER HA WITH DESCENDER
|
||
ҳ → h ; # CYRILLIC SMALL LETTER HA WITH DESCENDER
|
||
#
|
||
#
|
||
########################################################################
|
||
|