Module:kk-Cyrl-Arab-translit

---

The following documentation is generated by Module:documentation/functions/translit. ^[edit]

Useful links: subpage list • links • transclusions • testcases • sandbox

This module will transliterate Kazakh language text. The module should preferably not be called directly from templates or other modules. To use it from a template, use {{xlit}}. Within a module, use Module:languages#Language:transliterate.

For testcases, see Module:kk-Cyrl-Arab-translit/testcases.

Functions

tr(text, lang, sc)

Transliterates a given piece of text written in the script specified by the code sc, and language specified by the code lang.

When the transliteration fails, returns nil.

local export = {}
local rsplit = mw.text.split
local U = mw.ustring.char
local rsubn = mw.ustring.gsub
local ulen = mw.ustring.len

local frontvowels = "үҮәӘіІөӨ"
local backvowels = "ӯӮАаоОЫыҰұ"
local allvowels = "үҮиИеЕәӘөӨӯӮАаоОЫыІіЯяЮюЁё"
local allchar = "АаӘәБбВвГгҒғДдЕеЁёЖжЗзИиЙйКкҚқЛлМмНнҢңОоӨөПпРрСсТтУуҰұҮүФфХхҺһЦцЧчШшЩщЪъЫыІіЬьЭэЮюЯя"
local hmzh = U(0x674)

local mapping = {
 ["ү"] = "ۇ"
,["Ү"] = "ۇ"
,["т"] = "ت"
,["Т"] = "ت"
,["һ"] = "ھ"
,["Һ"] = "ھ"
,["р"] = "ر"
,["Р"] = "ر"
,["ф"] = "ف"
,["Ф"] = "ف"
,["ш"] = "ش"
,["Ш"] = "ش"
,["н"] = "ن"
,["Н"] = "ن"
,["п"] = "پ"
,["П"] = "پ"
,["й"] = "ي"
,["Й"] = "ي"
,["л"] = "ل"
,["Л"] = "ل"
,["з"] = "ز"
,["З"] = "ز"
,["е"] = "ە"
,["Е"] = "ە"
,["г"] = "گ"
,["Г"] = "گ"
,["б"] = "ب"
,["Б"] = "ب"
,["у"] = "ۋ"
,["У"] = "ۋ"
,["с"] = "س"
,["С"] = "س"
,["х"] = "ح"
,["Х"] = "ح"
,["ы"] = "ى"
,["Ы"] = "ى"
,["м"] = "م"
,["М"] = "م"
,["о"] = "و"
,["О"] = "و"
,["и"] = "ي"
,["И"] = "ي"
,["ж"] = "ج"
,["Ж"] = "ج"
,["к"] = "ك"
,["К"] = "ك"
,["д"] = "د"
,["Д"] = "د"
,["в"] = "ۆ"
,["В"] = "ۆ"
,["а"] = "ا"
,["А"] = "ا"
,["ң"] = "ڭ"
,["Ң"] = "ڭ"
,["ә"] = "ا"
,["Ә"] = "ا"
,["э"] = "ە"
,["Э"] = "ە"
,["ұ"] = "ۇ"
,["Ұ"] = "ۇ"
,["қ"] = "ق"
,["Қ"] = "ق"
,["ғ"] = "ع"
,["Ғ"] = "ع"
,["і"] = "ى"
,["І"] = "ى"--?
,["ө"] = "و"
,["Ө"] = "و"


--mostly in loanwords from Russian
,["ё"] = "يو"
,["Ё"] = "يو"
,["ц"] = "تس"
,["Ц"] = "تس"
,["ч"] = "چ"
,["Ч"] = "چ"
,["щ"] = "شش"
,["Щ"] = "شش"
,["ъ"] = ""
,["Ъ"] = ""
,["ь"] = ""
,["Ь"] = ""
,["э"] = "ە"
,["Э"] = "ە"
,["ю"] = "يۋ"
,["Ю"] = "يۋ"
,["я"] = "يا"
,["Я"] = "يا"
--хостинг - hostiñ
,["нг"] = "ڭ"
,["Нг"] = "ڭ"
--is it used?
,["ӯ"] = "ۇ"
,["Ӯ"] = "ۇ"

--punctuation 
,[","] = "،"
,["%"] = "٪"
,["?"] = "؟"
}

--[==[ disabled: hamza ligatures have been depricated by unicode
local function fvowels(text)
	-- front vowels
	text = rsubn(text, hmzh .. "[үҮ]", "ٷ")
	text = rsubn(text, hmzh .. "[Әә]", "ٵ")
	text = rsubn(text, hmzh .. "[Іі]", "ٸ")
	text = rsubn(text, hmzh .. "[Өө]", "ٶ")
	text = rsubn(text, hmzh .. "([" .. backvowels .. "])", "%1")
	return text
end ]==]

function export.tr(text, lang, sc)
	if type(text) == "table" then
		options = {}
		text, script = text.args[1], text.args[2]
	end
	
	if not sc then
		sc = require("Module:languages").getByCode("kk"):findBestScript(text):getCode()
	end
	
	if sc ~= "Cyrl" then
		return nil
	end
	text = rsubn(text, "^" .. "[-]", "-")
	text = rsubn(text, "([" .. allchar .. "])([^" .. allchar .. "])", "%1#_#%2")
	text = rsubn(text, "([^" .. allchar .. "])([" .. allchar .. "])", "%1#_#%2")
	text = rsubn(text, " | ", "# | #")
	text = "##" .. rsubn(text, " ", "# #") .. "##"
	-- treat every word as an individual string
	rsplit( "#" .."%s" .. "#", ",")
	rsplit( "#" .."_" .. "#", ",")
	text = rsubn(text, "[иИ][яЯ]", "йа")
	text = rsubn(text, "[иИ]([".. allvowels .."])", "й%1")
	text = rsubn(text, "([".. allvowels .."])[яЯ]", "%1йа")
	text = rsubn(text, "#[яЯ]", "#йа")
	text = rsubn(text, "[яЯ]", "ә")
	text = rsubn(text, "#([" .. allchar .. "]+)([" .. frontvowels .. "])", "#" .. hmzh .. "%1%2")
	text = rsubn(text, "#([" .. frontvowels .. "])", "#" .. hmzh .. "%1")
	text = rsubn(text, "#" .. hmzh .. "([" .. allchar .. allvowels .. "]+)([эЭЕеГгҒғКкҚқ])", "#%1%2")
	text = rsubn(text, "#" .. hmzh .. "([эЭЕеГгКк])", "#%1")
	-- text = fvowels(text) -- hamza ligatures have been depricated by unicode
	text = rsubn(text, '.', mapping)
	text = rsubn(text,"-", "ـ") -- Kashida
	text = rsubn(text, "#", "")
	text = rsubn(text, "_", "")

	return text
end
return export