172 lines
2.8 KiB
XML
172 lines
2.8 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="zh_Latn_PINYIN" target="ru" direction="forward" alias="ru-t-zh-latn-pinyin">
|
||
<tRule><![CDATA[
|
||
# Cyrillization of Mandarin Chinese from Pinyin into Russian (Palladius system).
|
||
#
|
||
# References:
|
||
# http://ru.wikipedia.org/wiki/Транскрипционная_система_Палладия
|
||
# http://www.omniglot.com/writing/mandarin_pts.htm
|
||
# http://www.pinyin.info/romanization/russian/index.html
|
||
# These differ in the treatment of some syllables (e.g. <jue>) from the rules
|
||
# below.
|
||
#
|
||
# Further commentary:
|
||
# http://languagelog.ldc.upenn.edu/nll/?p=604
|
||
#
|
||
# Remove tone marks.
|
||
|
||
:: NFD (NFC);
|
||
[\u0304\u0301\u030C\u0300\u0306] → ;
|
||
:: NFC (NFD);
|
||
#
|
||
#
|
||
# Syllabify. Add apostrophes to disambiguate whether <n> and <g> belong to the
|
||
# coda or onset of a syllable.
|
||
|
||
:: Null ();
|
||
ng } [aeou] → n\'g;
|
||
ng → ng\';
|
||
n } [aeiouü] → \'n;
|
||
n → n\';
|
||
#
|
||
#
|
||
# Main pass.
|
||
|
||
:: Null ();
|
||
#
|
||
#
|
||
|
||
Ai → Ай;
|
||
A → А;
|
||
B → Б;
|
||
Ch → Ч;
|
||
Ci → Цы;
|
||
C → Ц;
|
||
D → Д;
|
||
E → Э;
|
||
F → Ф;
|
||
G → Г;
|
||
H → Х;
|
||
Ju → Цз | ü;
|
||
J → Цз;
|
||
K → К;
|
||
L → Л;
|
||
M → М;
|
||
N → Н;
|
||
O → О;
|
||
P → П;
|
||
Qu → Ц | ü;
|
||
Q → Ц;
|
||
R → Ж;
|
||
Sh → Ш;
|
||
Si → Сы;
|
||
S → С;
|
||
T → Т;
|
||
Wu → У;
|
||
W → В;
|
||
Xu → С | ü;
|
||
X → С;
|
||
Yai → Яй;
|
||
Ya → Я;
|
||
Ye → Е;
|
||
Yi → И;
|
||
You → Ю;
|
||
Yo } ng → Ю;
|
||
Yo → Ё;
|
||
Yu → Ю | v;
|
||
Zh → Чж;
|
||
Zi → Цзы;
|
||
Z → Цз;
|
||
#
|
||
#
|
||
|
||
[$] { n → н;
|
||
\'n → н;
|
||
\' → ;
|
||
#
|
||
#
|
||
|
||
ai → ай;
|
||
a → а;
|
||
b → б;
|
||
ch → ч;
|
||
ci → цы;
|
||
c → ц;
|
||
d → д;
|
||
ei → эй;
|
||
e → э;
|
||
f → ф;
|
||
g → г;
|
||
h → х;
|
||
ia → я;
|
||
ie → е;
|
||
io } ng → ю;
|
||
iu → ю;
|
||
i → и;
|
||
ju → цз | ü;
|
||
j → цз;
|
||
k → к;
|
||
l → л;
|
||
m → м;
|
||
ng → н;
|
||
n → нь;
|
||
o } ng → у;
|
||
o → о;
|
||
p → п;
|
||
qu → ц | ü;
|
||
q → ц;
|
||
r } [^aeiou] → р;
|
||
r → ж;
|
||
sh → ш;
|
||
si → сы;
|
||
s → с;
|
||
t → т;
|
||
[Хх] { ui → уэй; # special exception
|
||
ui → уй;
|
||
uo → о;
|
||
u → у;
|
||
ü → ю | v;
|
||
#
|
||
# We use the dummy symbol <v> to signal that an <ü> or equivalent character has
|
||
# just been processed.
|
||
|
||
ve → э;
|
||
#
|
||
#alternative: ve → е;
|
||
|
||
v } [an] → ;
|
||
v → й;
|
||
wu → у;
|
||
w → в;
|
||
xu → с | ü;
|
||
x → с;
|
||
yai → яй;
|
||
ya → я;
|
||
ye → е;
|
||
yi → и;
|
||
you → ю;
|
||
yo } ng → ю;
|
||
yo → ё;
|
||
yu → ю | v;
|
||
zh → чж;
|
||
zi → цзы;
|
||
z → цз;
|
||
#
|
||
#
|
||
# Final pass: Make sure that the output consists entirely of Cyrillic letters.
|
||
|
||
:: NFC ();
|
||
]]></tRule>
|
||
</transform>
|
||
</transforms>
|
||
</supplementalData>
|