tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[transliteration] Regenerated script data

Browse Source
This commit is contained in:
Al
2015-06-17 23:46:29 -04:00
parent 2408cfa6f0
commit be8353ad9b
1 changed files with 39 additions and 35 deletions
Download Patch File Download Diff File Expand all files Collapse all files

74
src/transliteration_scripts.h
View File

@@ -14,44 +14,47 @@ typedef struct script_transliteration_rule {
} script_transliteration_rule_t; } script_transliteration_rule_t;
script_transliteration_rule_t script_transliteration_rules[] = { script_transliteration_rule_t script_transliteration_rules[] = {
{SCRIPT_GURMUKHI, NULL, 0, 1}, {SCRIPT_THAANA, NULL, 0, 2},
{SCRIPT_TELUGU, NULL, 1, 1}, {SCRIPT_GURMUKHI, NULL, 2, 1},
{SCRIPT_CYRILLIC, "be", 2, 1}, {SCRIPT_TELUGU, NULL, 3, 1},
{SCRIPT_CYRILLIC, "bg", 3, 1}, {SCRIPT_CYRILLIC, "be", 4, 1},
{SCRIPT_CYRILLIC, NULL, 4, 1}, {SCRIPT_CYRILLIC, "bg", 5, 1},
{SCRIPT_CYRILLIC, "ru", 5, 1}, {SCRIPT_CYRILLIC, NULL, 6, 1},
{SCRIPT_CYRILLIC, "uz", 6, 1}, {SCRIPT_CYRILLIC, "ru", 7, 1},
{SCRIPT_CYRILLIC, "kk", 7, 1}, {SCRIPT_CYRILLIC, "uz", 8, 1},
{SCRIPT_CYRILLIC, "sr", 8, 1}, {SCRIPT_CYRILLIC, "kk", 9, 1},
{SCRIPT_CYRILLIC, "mn", 9, 1}, {SCRIPT_CYRILLIC, "sr", 10, 1},
{SCRIPT_CYRILLIC, "mk", 10, 1}, {SCRIPT_CYRILLIC, "mn", 11, 1},
{SCRIPT_CYRILLIC, "uk", 11, 1}, {SCRIPT_CYRILLIC, "mk", 12, 1},
{SCRIPT_CYRILLIC, "ky", 12, 1}, {SCRIPT_CYRILLIC, "uk", 13, 1},
{SCRIPT_ORIYA, NULL, 13, 1}, {SCRIPT_CYRILLIC, "ky", 14, 1},
{SCRIPT_HANGUL, NULL, 14, 1}, {SCRIPT_ORIYA, NULL, 15, 1},
{SCRIPT_GUJARATI, NULL, 15, 1}, {SCRIPT_HANGUL, NULL, 16, 2},
{SCRIPT_HAN, NULL, 16, 1}, {SCRIPT_GUJARATI, NULL, 18, 1},
{SCRIPT_ARMENIAN, NULL, 17, 1}, {SCRIPT_HAN, NULL, 19, 1},
{SCRIPT_TAMIL, NULL, 18, 1}, {SCRIPT_ARMENIAN, NULL, 20, 1},
{SCRIPT_BENGALI, NULL, 19, 1}, {SCRIPT_TAMIL, NULL, 21, 1},
{SCRIPT_MALAYALAM, NULL, 20, 1}, {SCRIPT_BENGALI, NULL, 22, 1},
{SCRIPT_HIRAGANA, NULL, 21, 1}, {SCRIPT_MALAYALAM, NULL, 23, 1},
{SCRIPT_KANNADA, NULL, 22, 1}, {SCRIPT_HIRAGANA, NULL, 24, 1},
{SCRIPT_LATIN, NULL, 23, 1}, {SCRIPT_KANNADA, NULL, 25, 1},
{SCRIPT_GEORGIAN, NULL, 24, 2}, {SCRIPT_LATIN, NULL, 26, 1},
{SCRIPT_DEVANAGARI, NULL, 26, 1}, {SCRIPT_GEORGIAN, NULL, 27, 2},
{SCRIPT_THAI, NULL, 27, 1}, {SCRIPT_DEVANAGARI, NULL, 29, 1},
{SCRIPT_GREEK, NULL, 28, 3}, {SCRIPT_THAI, NULL, 30, 1},
{SCRIPT_CANADIAN_ABORIGINAL, NULL, 31, 1}, {SCRIPT_GREEK, NULL, 31, 3},
{SCRIPT_ARABIC, "fa", 32, 1}, {SCRIPT_CANADIAN_ABORIGINAL, NULL, 34, 1},
{SCRIPT_ARABIC, NULL, 33, 2}, {SCRIPT_ARABIC, "fa", 35, 1},
{SCRIPT_ARABIC, "ps", 35, 1}, {SCRIPT_ARABIC, NULL, 36, 2},
{SCRIPT_HEBREW, NULL, 36, 2}, {SCRIPT_ARABIC, "ps", 38, 1},
{SCRIPT_KATAKANA, NULL, 38, 2}, {SCRIPT_HEBREW, NULL, 39, 2},
{SCRIPT_COMMON, NULL, 40, 1} {SCRIPT_KATAKANA, NULL, 41, 2},
{SCRIPT_COMMON, NULL, 43, 1}
}; };
char *script_transliterators[] = { char *script_transliterators[] = {
"thaana-latin",
"maldivian-latin-bgn",
"gurmukhi-latin", "gurmukhi-latin",
"telugu-latin", "telugu-latin",
"belarusian-latin-bgn", "belarusian-latin-bgn",
@@ -66,6 +69,7 @@ char *script_transliterators[] = {
"ukrainian-latin-bgn", "ukrainian-latin-bgn",
"kirghiz-latin-bgn", "kirghiz-latin-bgn",
"oriya-latin", "oriya-latin",
"hangul-latin",
"korean-latin-bgn", "korean-latin-bgn",
"gujarati-latin", "gujarati-latin",
"han-latin", "han-latin",