ICU-59117.0.1.tar.gz

[apple/icu.git] / icuSources / data / translit / Latin_NumericPinyin.txt

diff --git a/icuSources/data/translit/Latin_NumericPinyin.txt b/icuSources/data/translit/Latin_NumericPinyin.txt
index cc8ce2eca6718ff4d736225265ef852775d56bab..316e0758ab1ea58194bcc3e2c16123e435da2ea4 100644 (file)
--- a/icuSources/data/translit/Latin_NumericPinyin.txt
+++ b/icuSources/data/translit/Latin_NumericPinyin.txt
@@ -1,18 +1,30 @@
-# ***************************************************************************
-# *
-# *  Copyright (C) 2004-2010, International Business Machines
-# *  Corporation; Unicode, Inc.; and others.  All Rights Reserved.
-# *
-# ***************************************************************************
+# © 2016 and later: Unicode, Inc. and others.
+# License & terms of use: http://www.unicode.org/copyright.html#License
+#
 # File: Latin_NumericPinyin.txt
-# Generated from CLDR 
+# Generated from CLDR
 #
+
+# According to the pinyin definitions I've been able to find:
+# 'a', 'e' are the preferred bases
+# otherwise 'o'
+# otherwise last vowel
+# The trailing form of syllables are the following:
+#         "a", "ai", "ao", "an", "ang",
+#         "o", "ou", "ong",
+#         "e", "ei", "er", "en", "eng",
+#         "i", "ia", "iao", "ie", "iu", "ian", "in", "iang", "ing", "iong",
+#         "u", "ua", "uo", "uai", "ui", "uan", "un", "uang", "ueng",
+#         "ü", "üe", "üan", "ün"
+# so the letters the tone will 'hop' are:
 ::NFD (NFC);
 $tone = [\u0304\u0301\u030C\u0300\u0306] ;
+# Move the tone to the end of a syllable, and convert to number
 e {($tone) r} → r &Pinyin-NumericPinyin($1);
 ($tone) ( [i o n u {o n} {n g}]) → $2 &Pinyin-NumericPinyin($1);
 ($tone) → &Pinyin-NumericPinyin($1);
-$vowel = [aAeEiIoOuUüÜ {u\u0308} {U\u0308} ];
+# The following backs up until it finds the right vowel, then deposits the tone
+$vowel = [aAeEiIoOuU {u\u0308} {U\u0308} vV];
 $consonant = [[a-z A-Z] - [$vowel]];
 $digit = [1-5];
 $1 &NumericPinyin-Pinyin($3) $2 ← ([aAeE]) ($vowel* $consonant*) ($digit);
@@ -20,3 +32,4 @@ $1 &NumericPinyin-Pinyin($3) $2 ← ([oO]) ([$vowel-[aeAE]]* $consonant*) ($digi
 $1 &NumericPinyin-Pinyin($3) $2 ← ($vowel) ($consonant*) ($digit);
 &NumericPinyin-Pinyin($1) ← [:letter:] {($digit)};
 ::NFC (NFD);
+