280 lines
8.7 KiB
XML
280 lines
8.7 KiB
XML
<?xml version="1.0" encoding="UTF-8" ?>
|
||
<!DOCTYPE supplementalData SYSTEM "../../common/dtd/ldmlSupplemental.dtd">
|
||
<!--
|
||
Copyright © 1991-2013 Unicode, Inc.
|
||
CLDR data files are interpreted according to the LDML specification (http://unicode.org/reports/tr35/)
|
||
For terms of use, see http://www.unicode.org/copyright.html
|
||
-->
|
||
<supplementalData>
|
||
<version number="$Revision$"/>
|
||
<transforms>
|
||
<transform source="bg" target="bg_Latn" variant="BGN" direction="forward" draft="contributed" alias="Bulgarian-Latin/BGN bg-Latn-t-bg-m0-bgn">
|
||
<tRule><![CDATA[
|
||
#
|
||
########################################################################
|
||
# BGN/PCGN 1952 System
|
||
#
|
||
# This system was adopted by the BGN in 1949 and by the PCGN in 1952.
|
||
# It reflects the much simplified Bulgarian orthography as officially
|
||
# revised in February 1945. The Bulgarian alphabet contains all of
|
||
# the characters present in the Russian alphabet with the exception
|
||
# of Ёё, Ыы, and Ээ. Two obsolete letters Ѫѫ and Ѣѣ are also given.
|
||
#
|
||
# The Bulgarian Alphabet as defined by the BGN (Page 15):
|
||
#
|
||
# АБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЬЮЯѪѢ
|
||
# абвгдежзийклмнопрстуфхцчшщъьюяѫѣ
|
||
#
|
||
# Originally prepared by Michael Everson <everson@evertype.com>
|
||
########################################################################
|
||
#
|
||
# MINIMAL FILTER: Bulgarian-Latin
|
||
#
|
||
|
||
:: [АБВГДЕЖЗИЙКЛМНОПРСТУФХЦЧШЩЪЬЮЯѪѢабвгдежзийклмнопрстуфхцчшщъьюяѫѣ] ;
|
||
:: NFD (NFC) ;
|
||
#
|
||
#
|
||
########################################################################
|
||
|
||
|
||
#
|
||
########################################################################
|
||
#
|
||
# Define All Transformation Variables
|
||
#
|
||
########################################################################
|
||
#
|
||
|
||
$upperConsonants = [БВГДЖЗЙКЛМНПРСТФХЦЧШЩЬ] ;
|
||
$lowerConsonants = [бвгджзйклмнпрстфхцчшщь] ;
|
||
$consonants = [$upperConsonants $lowerConsonants] ;
|
||
|
||
$upperVowels = [АЕИОУЪЮЯѪѢ] ;
|
||
$lowerVowels = [аеиоуъюяѫѣ] ;
|
||
$vowels = [$upperVowels $lowerVowels] ;
|
||
|
||
$lower = [$lowerConsonants $lowerVowels] ;
|
||
|
||
$bulgarian = [ $lower $upperConsonants $upperVowels ] ;
|
||
#
|
||
#
|
||
# 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 → 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
|
||
Е → E ; # CYRILLIC CAPITAL LETTER DE
|
||
е → e ; # CYRILLIC SMALL LETTER DE
|
||
Ж} $lower → Zh ; # CYRILLIC CAPITAL LETTER ZHE
|
||
Ж → ZH ; # CYRILLIC CAPITAL LETTER ZHE
|
||
ж → zh ; # 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
|
||
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 16 Note 4
|
||
#
|
||
# 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 Note 4
|
||
#
|
||
########################################################################
|
||
|
||
|
||
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
|
||
|
||
#
|
||
########################################################################
|
||
#
|
||
# Implied rule from BGN Russian-Latin transliteration (Page 94 Note 3.6).
|
||
#
|
||
# шт becomes sh·t
|
||
#
|
||
########################################################################
|
||
#
|
||
|
||
ШТ → SH·T ; # CYRILLIC CAPITAL LETTER SHA
|
||
Шт → Sh·t ; # CYRILLIC CAPITAL LETTER SHA
|
||
шт → sh·t ; # 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 → Sht ; # CYRILLIC CAPITAL LETTER SHCHA
|
||
Щ → SHT ; # CYRILLIC CAPITAL LETTER SHCHA
|
||
щ → sht ; # CYRILLIC SMALL LETTER SHCHA
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End Implied rule
|
||
#
|
||
########################################################################
|
||
|
||
|
||
Ъ → Ŭ ; # CYRILLIC CAPITAL LETTER HARD SIGN
|
||
ъ → ŭ ; # CYRILLIC SMALL LETTER HARD SIGN
|
||
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 16 Note 1
|
||
#
|
||
# In modern Bulgarian orthography, the character ъ does not occur in
|
||
# word-final position. It should be omitted in romanization when found
|
||
# on older sources.
|
||
#
|
||
# The following rule removes all Ъъ at the end of a word. It is assumed
|
||
# that when the condition is met, the text must be from an older source.
|
||
# Comment out with a '#' at the start of a line to disable.
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
|
||
$bulgarian { [Ъъ] } $wordBoundary > ;
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End BGN Page 16 Note 1
|
||
#
|
||
########################################################################
|
||
|
||
|
||
Ь → ’ ; # CYRILLIC CAPITAL LETTER SOFT SIGN
|
||
ь → ’ ; # CYRILLIC SMALL LETTER SOFT SIGN
|
||
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
|
||
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 16 Note 2
|
||
#
|
||
# The obsolete character Ѫ, which was replaced by Ъ in 1945, should be
|
||
# romanized Ŭ.
|
||
#
|
||
########################################################################
|
||
#
|
||
|
||
Ѫ → Ŭ ; # CYRILLIC CAPITAL LETTER BIG YUS
|
||
ѫ → ŭ ; # CYRILLIC SMALL LETTER BIG YUS
|
||
#
|
||
#
|
||
########################################################################
|
||
#
|
||
# End BGN Page 16 Note 2
|
||
#
|
||
########################################################################
|
||
|
||
|
||
#
|
||
########################################################################
|
||
#
|
||
# BGN Page 16 Note 3
|
||
#
|
||
# The obsolete character Ѣ, replaced in 1945 by Я or Е according to local
|
||
# pronunciation, should be romanized as e or ya, accordingly, if the
|
||
# pronunciation is known; otherwise as ye.
|
||
#
|
||
########################################################################
|
||
#
|
||
|
||
Ѣ} $lower → Ye ; # CYRILLIC CAPITAL LETTER YAT
|
||
Ѣ → YE ; # CYRILLIC CAPITAL LETTER YAT
|
||
ѣ → ye ; # CYRILLIC SMALL LETTER YAT
|
||
#
|
||
#
|
||
# Alternative rule where appropriate for local pronunciation. To apply
|
||
# uncomment the following by removing the '#' mark at the start of the
|
||
# line and insert before the three rule lines above.
|
||
#
|
||
# Ѣ} $lower → e ; # CYRILLIC CAPITAL LETTER YAT
|
||
# Ѣ → E ; # CYRILLIC CAPITAL LETTER YAT
|
||
# ѣ → e ; # CYRILLIC SMALL LETTER YAT
|
||
#
|
||
########################################################################
|
||
#
|
||
# End BGN Page 16 Note 3
|
||
#
|
||
########################################################################
|
||
|
||
]]></tRule>
|
||
</transform>
|
||
</transforms>
|
||
</supplementalData>
|