244 lines
11 KiB
Plaintext
244 lines
11 KiB
Plaintext
# © 2016 and later: Unicode, Inc. and others.
|
||
# License & terms of use: http://www.unicode.org/copyright.html#License
|
||
#
|
||
# File: ru_ru_Latn_BGN.txt
|
||
# Generated from CLDR
|
||
#
|
||
|
||
# BGN/PCGN 1947 System
|
||
#
|
||
# The BGN/PCGN system for Russian was adopted by the BGN in 1944 and
|
||
# by the PCGN in 1947 for use in romanizing names written in the
|
||
# Russian Cyrillic alphabet.
|
||
#
|
||
# The Russian Alphabet as defined by the BGN (Page 93):
|
||
# АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯ
|
||
# абвгдеёжзийклмнопрстуфхцчшщъыьэюя
|
||
#
|
||
# Originally prepared by Michael Everson everson@evertype.com
|
||
# Fixed by Frank Yung-Fong Tang ftang@google.com
|
||
#
|
||
# Test Data from http://en.wikipedia.org/wiki/BGN/PCGN_romanization_of_Russian
|
||
########################################################################
|
||
# MINIMAL FILTER: Russian-Latin
|
||
::[АБВГДЕЁЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЫЬЭЮЯабвгдеёжзийклмнопрстуфхцчшщъыьэюя];
|
||
# BUG(ftang) remove the following line. Otherwise the rule for
|
||
# Й й Ё ё will break since the rule is written in NFC but
|
||
# the line decomposes the text.
|
||
# :: NFD (NFC) ;
|
||
########################################################################
|
||
# Define All Transformation Variables
|
||
########################################################################
|
||
$prime = ʹ ;
|
||
$doublePrime = ʺ ;
|
||
$wordBoundary = [^[:L:][:M:][:N:]] ;
|
||
$upperVowels = [АЕЁЭИОУЫЮЯ] ;
|
||
$lowerVowels = [аеёэиоуыюя] ;
|
||
$vowels = [$upperVowels $lowerVowels] ;
|
||
$upperConsonants = [[:Uppercase:]-$vowels] ;
|
||
$lowerConsonants = [[:Lowercase:]-$vowels] ;
|
||
$consonants = [$upperConsonants $lowerConsonants] ;
|
||
$upper = [:Uppercase:];
|
||
$lower = [:Lowercase:];
|
||
########################################################################
|
||
# Rules moved to front to avoid masking
|
||
########################################################################
|
||
$lowerVowels { ы → ·y ;
|
||
$upperVowels { [Ыы] → ·Y ;
|
||
[$consonants - [Йй]]{Э → ·E ;
|
||
[$consonants - [Йй]]{э → ·e ;
|
||
[$upperVowels [ЙЪЬ]] { Е } $upper → YE ; # CYRILLIC CAPITAL LETTER IE
|
||
[$upperVowels [ЙЪЬ]] { Е → Ye ; # CYRILLIC CAPITAL LETTER IE
|
||
[$upperVowels $lowerVowels [ЙйЪъЬь]] { е → ye ; # CYRILLIC SMALL LETTER IE
|
||
[$upperVowels [ЙЪЬ]] { Ё } $upper → YË ; # CYRILLIC CAPITAL LETTER IO
|
||
[$upperVowels [ЙЪЬ]] { Ё → Yë ; # CYRILLIC CAPITAL LETTER IO
|
||
[$upperVowels $lowerVowels [ЙйЪъЬь]] { ё → yë ; # CYRILLIC SMALL LETTER IO
|
||
# Since in the above rule we look use the context before the characters,
|
||
# we have to perform them in a separate pass before we change the vowels
|
||
# the ::Null forces a separate pass.
|
||
::Null;
|
||
########################################################################
|
||
# Start of Alphabetic Transformations
|
||
########################################################################
|
||
А → A ; # CYRILLIC CAPITAL LETTER A
|
||
а → a ; # CYRILLIC SMALL LETTER A
|
||
Б → B ; # CYRILLIC CAPITAL LETTER BE
|
||
б → b ; # CYRILLIC SMALL LETTER BE
|
||
В → V ; # CYRILLIC CAPITAL LETTER VE
|
||
в → v ; # CYRILLIC SMALL LETTER VE
|
||
Г → G ; # CYRILLIC CAPITAL LETTER GHE
|
||
г → g ; # CYRILLIC SMALL LETTER GHE
|
||
Д → D ; # CYRILLIC CAPITAL LETTER DE
|
||
д → d ; # CYRILLIC SMALL LETTER DE
|
||
########################################################################
|
||
# BGN Page 94 Rule 1:
|
||
# # The character e should be romanized ye
|
||
# initially, after the vowel # characters a, e, ё, и, о, у, ы, э, ю,
|
||
# and я, and after й, ъ, and ь.
|
||
# In all other instances, it should
|
||
# be romanized e.
|
||
########################################################################
|
||
# BUG(ftang)- the following two lines said BEFORE the vowels, instead of AFTER
|
||
# Е}[$upperVowels [ЙЪЬ]] → YE ; # CYRILLIC CAPITAL LETTER IE
|
||
# Е}[$lowerVowels [йъь]] → Ye ; # CYRILLIC CAPITAL LETTER IE
|
||
$wordBoundary{Е} $upper → YE ; # CYRILLIC CAPITAL LETTER IE
|
||
$wordBoundary{Е → Ye ; # CYRILLIC CAPITAL LETTER IE
|
||
Е → E ; # CYRILLIC CAPITAL LETTER IE
|
||
#
|
||
# BUG(ftang)- the following line said BEFORE the vowels, instead of AFTER
|
||
# е}[$upperVowels $lowerVowels [ЙйЪъЬь]] → ye ; # CYRILLIC SMALL LETTER IE
|
||
$wordBoundary{е → ye ; # CYRILLIC SMALL LETTER IE
|
||
е → e ; # CYRILLIC SMALL LETTER IE
|
||
########################################################################
|
||
# End of Rule 1
|
||
########################################################################
|
||
########################################################################
|
||
# BGN Page 94 Rule 2:
|
||
#
|
||
# The character ё is not considered a separate character of the
|
||
# Russian alphabet and the dieresis is generally not shown. When the
|
||
# dieresis is shown, the character should be romanized yë initially,
|
||
# after the vowel characters a, e, ё, и, о, у, ы, э, ю, and я, and
|
||
# after й, ъ, and ь, In all other instances, it should be romanized
|
||
# ё. When the dieresis is not shown, the character may still be
|
||
# romanized in the preceding manner or, alternatively, in accordance
|
||
# with note 1.
|
||
########################################################################
|
||
# BUG(ftang)- the following two lines said BEFORE the vowels, instead of AFTER
|
||
# Ё}[$upperVowels [ЙЪЬ]] → YË ; # CYRILLIC CAPITAL LETTER IO
|
||
# Ё}[$lowerVowels [йъь]] → Yë ; # CYRILLIC CAPITAL LETTER IO
|
||
$wordBoundary {Ё} $upper → YË ; # CYRILLIC CAPITAL LETTER IO
|
||
$wordBoundary {Ё} $lower → Yë ; # CYRILLIC CAPITAL LETTER IO
|
||
Ё → Ë ; # CYRILLIC CAPITAL LETTER IO
|
||
# BUG(ftang)- the following line said BEFORE the vowels, instead of AFTER
|
||
# ё}[$upperVowels $lowerVowels [ЙйЪъЬь]] → yë ; # CYRILLIC SMALL LETTER IO
|
||
$wordBoundary{ё → yë ; # CYRILLIC SMALL LETTER IO
|
||
ё → ë ; # CYRILLIC SMALL LETTER IO
|
||
########################################################################
|
||
# End of Rule 2
|
||
########################################################################
|
||
Ж} $lower → Zh ; # CYRILLIC CAPITAL LETTER ZHE
|
||
Ж → ZH ; # CYRILLIC CAPITAL LETTER ZHE
|
||
ж → zh ; # CYRILLIC SMALL LETTER ZHE
|
||
########################################################################
|
||
# BGN Page 94 Rule 3.4
|
||
# э after any consonant character except
|
||
# й becomes ·е
|
||
########################################################################
|
||
З → Z ; # CYRILLIC CAPITAL LETTER ZE
|
||
з → z ; # CYRILLIC SMALL LETTER ZE
|
||
# BUG(ftang) The following two lines said those consonant becomes ·е
|
||
# [$consonants - [Йй]]}Э → ·Е ;
|
||
# [$consonants - [Йй]]}э → ·е ;
|
||
########################################################################
|
||
# End of Rule 3.4
|
||
########################################################################
|
||
И → I ; # CYRILLIC CAPITAL LETTER I
|
||
и → i ; # CYRILLIC SMALL LETTER I
|
||
########################################################################
|
||
# BGN Page 94 Rule 3:
|
||
#
|
||
# Unusual Russian character sequences occurring primarily in
|
||
# non-Russian-language names may be romanized as shown below in order
|
||
# to provide differentiation from regularly-occurring digraphs and
|
||
# character sequences.
|
||
#
|
||
# BGN Page 94 Rule 3.1
|
||
# й before а, у, ы, or э becomes у·
|
||
########################################################################
|
||
Й}[АаУуЫыЭэ] → Y· ; # CYRILLIC CAPITAL LETTER I
|
||
й}[АаУуЫыЭэ] → y· ; # CYRILLIC SMALL LETTER I
|
||
Й → Y ; # CYRILLIC CAPITAL LETTER I
|
||
й → y ; # CYRILLIC SMALL LETTER I
|
||
########################################################################
|
||
# End Rule 3.1
|
||
########################################################################
|
||
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
|
||
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
|
||
С → S ; # CYRILLIC CAPITAL LETTER ES
|
||
с → s ; # CYRILLIC SMALL LETTER ES
|
||
########################################################################
|
||
# BGN Page 94 Rule 3.5
|
||
# тс becomes t·s
|
||
########################################################################
|
||
ТС → T·S ; # CYRILLIC CAPITAL LETTER TE
|
||
Тс → T·s ; # CYRILLIC CAPITAL LETTER TE
|
||
тс → t·s ; # CYRILLIC SMALL LETTER TE
|
||
Т → T ; # CYRILLIC CAPITAL LETTER TE
|
||
т → t ; # CYRILLIC SMALL LETTER TE
|
||
########################################################################
|
||
# End Rule 3.5
|
||
########################################################################
|
||
У → U ; # CYRILLIC CAPITAL LETTER U
|
||
у → u ; # 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
|
||
Ц} $lower → Ts ; # CYRILLIC CAPITAL LETTER TSE
|
||
Ц → TS ; # CYRILLIC CAPITAL LETTER TSE
|
||
ц → ts ; # CYRILLIC SMALL LETTER TSE
|
||
Ч} $lower → Ch ; # CYRILLIC CAPITAL LETTER CHE
|
||
Ч → CH ; # CYRILLIC CAPITAL LETTER CHE
|
||
ч → ch ; # CYRILLIC SMALL LETTER CHE
|
||
########################################################################
|
||
# BGN Page 94 Rule 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 Rule 3.6
|
||
########################################################################
|
||
Ъ → $doublePrime ; # CYRILLIC CAPITAL LETTER HARD SIGN
|
||
ъ → $doublePrime ; # CYRILLIC SMALL LETTER HARD SIGN
|
||
########################################################################
|
||
# BGN Page 94 Rule 3.2
|
||
# ы before а, у, ы, or э becomes у·
|
||
#
|
||
# BGN Page 94 Rule 3.3
|
||
# ы after any vowel character becomes ·у
|
||
########################################################################
|
||
#
|
||
# BUG(ftang) the following line said the vowels will change
|
||
# $vowels}Ы → ·Y ; # CYRILLIC CAPITAL LETTER I
|
||
# $vowels}ы → ·y ; # CYRILLIC CAPITAL LETTER I
|
||
Ы}[АаУуЫыЭэ] → Y· ; # CYRILLIC CAPITAL LETTER YERU
|
||
ы}[ауыэ] → y· ; # CYRILLIC SMALL LETTER YERU
|
||
Ы → Y ; # CYRILLIC CAPITAL LETTER YERU
|
||
ы → y ; # CYRILLIC SMALL LETTER YERU
|
||
########################################################################
|
||
# End Rule 3.2 and 3.3
|
||
########################################################################
|
||
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
|
||
|