Module:tg-Latn-Cyrl-translit

From Wiktionary, the free dictionary
Jump to navigation Jump to search

This module will transliterate Tajik language text per WT:TG TR. 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:tg-Latn-Cyrl-translit/testcases.

Functions[edit]

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 = {text}

local tt = {
	["t"] = "т",
	["r"] = "р",
	["f"] = "ф",
	["š"] = "ш",
	["h"] = "ҳ",
	["ʾ"] = "ъ",
	["n"] = "н",
	["p"] = "п",
	["y"] = "й",
	["l"] = "л",
	["z"] = "з",
	["e"] = "е",
	["g"] = "г",
	["b"] = "б",
	["u"] = "у",
	["ü"] = "ӯ",
	["s"] = "с",
	["x"] = "х",
	["č"] = "ч",
	["m"] = "м",
	["o"] = "о",
	["i"] = "и",
	["ī"] = "ӣ",
	["ž"] = "ж",
	["k"] = "к",
	["d"] = "д",
	["v"] = "в",
	["a"] = "а",
	["j"] = "ҷ",
	["q"] = "қ",
	["ġ"] = "ғ",
	--capital letters
	-- fa-IPA doesnt support letter case.
	-- these are here as a precaution
	["T"] = "Т",
	["R"] = "Р",
	["F"] = "Ф",
	["Š"] = "Ш",
	["H"] = "Ҳ",
	["N"] = "Н",
	["P"] = "П",
	["Y"] = "Й",
	["L"] = "Л",
	["Z"] = "З",
	["E"] = "Е",
	["G"] = "Г",
	["B"] = "Б",
	["U"] = "У",
	["Ü"] = "Ӯ",
	["S"] = "С",
	["X"] = "Х",
	["Č"] = "Ч",
	["M"] = "М",
	["O"] = "О",
	["I"] = "И",
	["Ī"] = "Ӣ",
	["Ž"] = "Ж",
	["K"] = "К",
	["D"] = "Д",
	["V"] = "В",
	["A"] = "А",
	["J"] = "Ҷ",
	["Q"] = "Қ",
	["Ġ"] = "Ғ",
	-- ["ʾ"] = "Ъ" - this cant be uppercase in latin
	["ye"] = "e",
	["yi"] = "i",
	["yī"] = "ī",
	["yo"] = "ё",
	["yu"] = "ю",
	["ya"] = "я",
	-- uppercase (fa-IPA doesnt support letter case, these are
	-- precautionary)
	["Ye"] = "Е",
	["Yi"] = "I",
	["Yī"] = "Ī",
	["Yo"] = "Ё",
	["Yu"] = "Ю",
	["Ya"] = "Я",
}

local allchar = "AaBbVvGgĠġDdEeŽžZzIiĪīYyKkQqLlMmNnOoPpCcTtUuÜüFfXxHhČčJjŠšʾ"

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("tg"):findBestScript(text):getCode()
	end

	if sc ~= "Latn" then
		return nil
	end

	--initial "e" forms
	text = mw.ustring.gsub(text, "([^" .. allchar .. "])" .. "e", "%1э")
	text = mw.ustring.gsub(text, "([^" .. allchar .. "])" .. "E", "%1Э")
	--tajik orthography uses dashes but NOT the way fa-IPA does
	--so they need to be removed
	text = mw.ustring.gsub(text, "%-", "")
	text = mw.ustring.gsub(text, "[Yy][eiīoua]", tt)
	text = mw.ustring.gsub(text, ".", tt)

	return text
end

return export