Module:xchc-translit: Difference between revisions

From Linguifex
Jump to navigation Jump to search
No edit summary
No edit summary
Line 4: Line 4:
local D = u(0x0308) -- COMBINING DIAERESIS
local D = u(0x0308) -- COMBINING DIAERESIS
local A = u(0x0301) -- COMBINING ACUTE
local A = u(0x0301) -- COMBINING ACUTE
local DA = U(0x030B) -- COMBINING DOUBLE ACUTE
local DA = u(0x030B) -- COMBINING DOUBLE ACUTE
 


local export = {}
local export = {}

Revision as of 17:15, 2 October 2023



local u = mw.ustring.char
local M = u(0x0304) -- COMBINING MACRON
local B = u(0x0306) -- COMBINING BREVE
local D = u(0x0308) -- COMBINING DIAERESIS
local A = u(0x0301) -- COMBINING ACUTE
local DA = u(0x030B) -- COMBINING DOUBLE ACUTE

local export = {}

local tt = { -- Unofficial transliteration loosely based on ISO 9
	["І"]='J', ["і"]='j',
	["Э"]='E', ["э"]='e',
	["О"]='O', ["о"]='o',
	["Е" .. M .. D]='Jō', ["е" .. M .. D]='jō', ["Ё"]='Jo', ["ё"]='jo',
	["А" .. M]='Ā', ["а" .. M]='ā', ["А"]='A', ["а"]='a',
	["Б"]='B', ["б"]='b',
	["В"]='V', ["в"]='v',
	["Г"]='G', ["г"]='g',
	["Ғ"]='Ǧ', ["ғ"]='ǧ',
	["Д"]='D', ["д"]='d',
	["Е" .. M]='Jē', ["е" .. M]='jē', ["Е"]='Je', ["е"]='je',
	["Ж"]='Ž', ["ж"]='ž',
	["З"]='Z', ["з"]='z', 
	["Ӣ"]='Jī', ["ӣ"]='jī', ["И"]='Ji', ["и"]='ji',
	["К"]='K', ["к"]='k',
	["Л"]='L', ["л"]='l',
	["М"]='M', ["м"]='m',
	["Н"]='N', ["н"]='n',
	["Ӈ"]='Ŋ', ["ӈ"]='ŋ',
	["О" .. M]='Ō', ["о" .. M]='ō', ["О"]='O', ["о"]='o',
	["Ө" .. M]='Ȫ', ["ө" .. M]='ȫ', ["Ө"]='Ö', ["ө"]='ö',
	["П"]='P', ["п"]='p',
	["Р"]='R', ["р"]='r',
	["С"]='S', ["с"]='s',
	["Т"]='T', ["т"]='t',
	["Ӯ"]='Ū', ["ӯ"]='ū', ["У"]='U', ["у"]='u',
	["Ү" .. M]='Ȳ', ["ү" .. M]='ȳ', ["Ү"]='Y', ["ү"]='y',
	["Ф"]='F', ["ф"]='f',
	["ХӀ"]='H', ["хӀ"]='h', ["Х"]='X', ["х"]='x',
	["Ц"]='Ts', ["ц"]='ts',
	["Ч"]='Č', ["ч"]='č',
	["Ш"]='Š', ["ш"]='š',
	["Ы" .. M]='Ī', ["ы" .. M]='ī', ["Ы"]='I', ["ы"]='i',
	["Ю" .. M]='Jū', ["ю" .. M]='jū', ["Ю"]='Ju', ["ю"]='ju',
	["Я" .. M]='Jā', ["я" .. M]='jā', ["Я"]='Ja', ["я"]='ja',
	
	["Lь"] = "Lj", ["lь"] = "lj",
	["Nь"]='Nj', ["n[ьj]"]='nj', 
	["[ьЬ]"]="'", ["[Ъъ]"]='"',
	
	["Й"]='J', ["й"]='j',
};

function export.tr(text, lang, sc)
	for pat, repl in pairs(tt) do
		text = mw.ustring.gsub(text, pat, repl)
	end
	return text
end

local tt_lepik= { 
	[B .. M] = DA, [M] = A,
	["Ь"]="Y", ["Ь"]="y",
	["І"]='J', ["і"]='j',
	["Э"]='E', ["э"]='e',
	["О"]='O', ["о"]='o',
	["Е" .. A .. D]='Jō', ["е" .. A .. D]='jō', ["Е" .. D]='Jo', ["е" .. D]='jo',
	["А"]='A', ["а"]='a',
	["Б"]='B', ["б"]='b',
	["В"]='V', ["в"]='v',
	["Г"]='G', ["г"]='g',
	["Ғ"]='Gh', ["ғ"]='gh',
	["Дз"]='Dc', ["дз"]='dc', ["Дж"]='Dz', ["дж"]='dz',
	["Д"]='D', ["д"]='d',
	["Е"]='Je', ["е"]='je',
	["Ж"]='Ž', ["ж"]='ž',
	["З"]='Z', ["з"]='z', 
	["И" .. B]='J', ["и" .. B]='j', ["И"]='Ji', ["и"]='ji',
	["К"]='K', ["к"]='k',
	["Л"]='L', ["л"]='l',
	["М"]='M', ["м"]='m',
	["Н"]='N', ["н"]='n',
	["Ӈ"]='Ng', ["ӈ"]='ng',
	["О"]='O', ["о"]='o',
	["Ө" .. A]='O' .. DA, ["ө" .. A]='o' .. DA, ["Ө"]='O' .. D, ["ө"]='o' .. D,
	["П"]='P', ["п"]='p',
	["Р"]='R', ["р"]='r',
	["С"]='S', ["с"]='s',
	["Т"]='T', ["т"]='t',
	["У"]='U', ["у"]='u',
	["Ү" .. A]='U' .. DA, ["ү" .. A]='u' .. DA, ["Ү"]='U' .. D, ["ү"]='u' .. D,
	["Ф"]='F', ["ф"]='f',
	["ХӀ"]='H', ["хӀ"]='h', ["Х"]='Ch', ["х"]='ch',
	["Ц"]='C', ["ц"]='c',
	["Ч"]='Cs', ["ч"]='cs',
	["Ш"]='Š', ["ш"]='š',
	["Ы"]='I', ["ы"]='i',
	["Ю"]='Ju', ["ю"]='ju',
	["Я"]='Ja', ["я"]='ja',
	["[Ъъ]"]="'",
};

function export.tr_lepik(text, lang, sc)
	
	text = mw.ustring.toNFD(text)
	
	for pat, repl in pairs(tt) do
		text = mw.ustring.gsub(text, pat, repl)
	end
	return mw.ustring.toNFC(text)
end

return export