Модул:fa-translit

Test cases

(refresh)

test_translit_persian:

	Text	Expected	Actual	Differs at
	سَرانجام	saranjâm	саканҷам	1
	سَرانْجام	saranjâm	саканҷам	1
	سَرَانْجَام	saranjâm	сакаанҷаам	1
	کُروز	koruz	кокӯз	1
	کُرُوز	korouz	кокuз	1
	طَنین	tanin	еанин	1
	طَنِین	taneyn	еанeyн	1
	عَصاً	’asan	’асan	2
	خانه	xâne	ханe	1
	خانِه	xâne	ханиҳ	1
	کُرِۀ شُمالی	kore-ye šomâli	коки-ye шомали	1
	ضَمّه	zamme	заммe	1
	ضَمِّه	zamme	замииҳ	1
	وُدکا	vodkâ	ӯодка	1
	اَرمَنِستان	armanestân	аакманистан	1
	باکو	bâku	бакӯ	1

---

local export = {}

local U = mw.ustring.char

local fatHatan = U(0x64B) -- What is the Persian term for this?
local fathe = U(0x64E) -- also zabar
local zamme = U(0x64F) -- also piš
local kasre = U(0x650) -- also zir
local tashdid = U(0x651) -- also called shadda
local jazm = U(0x652)

local waw = U(0x0648)
local ye = U(0x06CC)

local group = "بپتثجچحخدذرزژسشصضطظغفقکگلمنوهی"

 
local mapping = {
	["ا"] = 'а', ["ب"] = 'б', ["پ"] = 'п', ["ت"] = 'т', ["ث"] = 'с', ["ج"] = 'ҷ', ["چ"] = 'ч', ["ح"] = 'ҳ', ["خ"] = 'х', 
	["د"] = 'д', ["ذ"] = 'з', ["ر"] = 'к', ["ز"] = 'з', ["ژ"] = 'ж', ["س"] = 'с', ["ش"] = 'ш', ["ص"] = 'с', ["ض"] = 'з', 
	["ط"] = 'е', ["ظ"] = 'з', ["غ"] = 'ғ', ["ف"] = 'ф', ["ق"] = 'қ', ["ک"] = 'к', ["گ"] = 'г', ["ل"] = 'л', 
	["م"] = 'м', ["ن"] = 'н', ["و"] = 'ӯ', ["ه"] = 'ҳ', ["ی"] = 'и', ["آ"] = 'о',

	-- displaying on separate lines as the viewing becomes distorted on these combinations
	["ع"] = "’",
	["ء"] = "’",
	["ئ"] = "’", 
	["ؤ"] = "’",
	["أ"] = "’",
	
	-- diacritics
	[fathe] = "а",
	[kasre] = "и",
	[zamme] = "о",
	[jazm] = "", -- also sokun - no vowel
	[U(0x200C)] = "-", -- ZWNJ (zero-width non-joiner)
	[fatHatan] = "н",
	-- ligatures
    ["ﻻ"] = "ло",
	["ﷲ"] = "ллоҳ",
	-- kashida
	["ـ"] = "", -- kashida, no sound
	-- numerals
	["۱"] = "1", ["۲"] = "2", ["۳"] = "3", ["۴"] = "4", ["۵"] = "5",
	["۶"] = "6", ["۷"] = "7", ["۸"] = "8", ["۹"] = "9", ["۰"] = "0",
	-- normal arabic variants to numerals
	["١"] = "1", ["٢"] = "2", ["٣"] = "3", ["٤"] = "4", ["٥"] = "5",
	["٦"] = "6", ["٧"] = "7", ["٨"] = "8", ["٩"] = "9", ["٠"] = "0",
	-- punctuation (leave on separate lines)
	["؟"] = "?", -- question mark
	["،"] = ",", -- comma
	["؛"] = ";", -- semicolon
	["«"] = '“', -- quotation mark
	["»"] = '”', -- quotation mark
	["٪"] = "%", -- percent
	["؉"] = "‰", -- per mille
	["٫"] = ".", -- decimals
	["٬"] = ",", -- thousand
	["ۀ"] = "-ye" -- he ye (in ezâfe)
}
 
function export.tr(text, lang, sc)

    text = mw.ustring.gsub(text, '([' .. group .. ']' .. tashdid .. '?)ه$', '%1e') 

    -- ou
    -- NOT WORKING
    text = mw.ustring.gsub(text, zamme .. waw .. '([' .. group .. '])', "ou%1")

    -- ey
    -- WORKING
    text = mw.ustring.gsub(text, kasre .. ye .. '([' .. group .. '])', "ey%1")

	text = mw.ustring.gsub(text, 'ىٰ', "â")
	text = mw.ustring.gsub(text, 'ا' .. fatHatan, "an")
	-- text = mw.ustring.gsub(text, 'الله', "ﷲ")
	-- text = mw.ustring.gsub(text, 'لا', "ﻻ")
	text = mw.ustring.gsub(text, '.', mapping)
	text = mw.ustring.gsub(text, 'ou', "u")
	text = mw.ustring.gsub(text, 'aâ', "â")
	text = mw.ustring.gsub(text, 'âa', "a")
	text = mw.ustring.gsub(text, 'ei', "i")
	text = mw.ustring.gsub(text, 'ai', "ay")
	text = mw.ustring.gsub(text, 'au', "aw")
	text = mw.ustring.gsub(text, 'u([aâeiou])', "v%1")
	text = mw.ustring.gsub(text, 'i([aâeiou])', "y%1")
	text = mw.ustring.gsub(text, "([aâeiou])(" .. tashdid .. ")", "%2%1") -- swapping tašdid with vowels
	text = mw.ustring.gsub(text, "(.)" .. tashdid, "%1%1") -- implementing tašdid
	text = mw.ustring.gsub(text, 'eh$', "e")
	text = mw.ustring.gsub(text, 'eh([^aâeiouy’bdfghjklmnpqrstvyxzčğšž])', "e%1")

	return text
end
 
return export