Module:bn-translit

Ang dokumentasyon ng ng modyul na ito ay maaaring likhain sa Module:bn-translit/doc

-- Transliteration for Bengali

local export = {}
local gsub = mw.ustring.gsub
local match = mw.ustring.match
 
local conv = {
	-- consonants
	['ক্ষ'] = 'kkh', ['জ্ঞ'] = 'gg',
	['ক'] = 'k',	['খ'] = 'kh',	['গ'] = 'g',	['ঘ'] = 'gh',	['ঙ'] = 'ṅ',
	['চ'] = 'c',	['ছ'] = 'ch',	['জ'] = 'j',	['ঝ'] = 'jh',	['ঞ'] = 'ñ',
	['ট'] = 'ṭ',	['ঠ'] = 'ṭh',	['ড'] = 'ḍ',	['ঢ'] = 'ḍh',	['ণ'] = 'ṇ',
	['ত'] = 't',	['থ'] = 'th',	['দ'] = 'd',	['ধ'] = 'dh',	['ন'] = 'n',
	['প'] = 'p',	['ফ'] = 'ph',	['ব'] = 'b',	['ভ'] = 'bh',	['ম'] = 'm',
	['য'] = 'j',	['র'] = 'r',	['ল'] = 'l',	['ৱ'] = 'w', 
	['শ'] = 'ś',	['ষ'] = 'ś',	['স'] = 'ś',	['হ'] = 'h',	
	['য়'] = 'ẏ',	['ড়'] = 'ṛ',	['ঢ়'] = 'ṛh',

	-- visarga
	['ঃ'] = 'ḥ',

	-- vowel diacritics
	['ি'] = 'i',	['ু'] = 'u',	['ৃ'] = 'ri',	['ে'] = 'e',	['ো'] = 'o',
	['া'] = 'a',	['ী'] = 'ī',	['ূ'] = 'ū',	['ৈ'] = 'oi',	['ৌ'] = 'ou',

	-- vowel signs
	['অ'] = 'ô',	['ই'] = 'i',	['উ'] = 'u',	['ঋ'] = 'ri',	['এ'] = 'e',	['ও'] = 'o',
	['আ'] = 'a',	['ঈ'] = 'ī',	['ঊ'] = 'ū',	['ঐ'] = 'oi',	['ঔ'] = 'ou',

	--hôshôntô
	['্'] = '',

	-- chôndrôbindu
	["ঁ"] = "̃",
		
	-- ônussar
	['ং'] = 'ṅ',
	
	-- ôbôgrôhô
	['ঽ']='’',

	-- khôndô tô
	['ৎ'] = 't',

	-- numerals
	['০'] = '0', ['১'] = '1', ['২'] = '2', ['৩'] = '3', ['৪'] = '4', ['৫'] = '5', ['৬'] = '6', ['৭'] = '7', ['৮'] = '8', ['৯'] = '9',
 
	-- punctuation
	['।'] = '.', -- dari
}

local deaspirate = {
	['খ'] = 'ক',	['ঘ'] = 'গ',
	['ছ'] = 'চ',	['ঝ'] = 'জ',
	['ঠ'] = 'ট',	['ঢ'] = 'ড',
	['থ'] = 'ত',	['ধ'] = 'দ',
	['ফ'] = 'প',	['ভ'] = 'ব', 	
	['ঢ়'] = 'ড়',
}

function export.tr(text, lang, sc)
	local c = '([কষজঞকখগঘঙচছজঝঞটঠডঢণতথদধনপফবভমযরলৱশষসহ]়?)'
	local y = 'য'
	local r = 'র'
	local v = '([ô্িুৃেোাীূৈৌঅইউঋএওআঈঊঐঔ])'
	local virama = '্'
	local n = '(ং?)'
	
	local no_virama = gsub(v,virama,"")
	
	text = text .. " "
	
	text = gsub(text, c .. '্য', function(f)
		if deaspirate[f] then
			return deaspirate[f] .. '্' .. f
		end
	end)
	
	text = gsub(text, c, "%1ô")
	text = gsub(text, "ô".. "([ô্িুৃেোাীূৈৌ])", "%1")
	
	text = gsub(text, v .. n .. c .. "ô ", function(j, k, l) -- ending
		return l == y and j .. k .. l .. "ô " or j .. k .. l .. " "
	end)
	
	local pattern = v .. n .. c .. "ô" .. c .. no_virama
	local continue = true
	while continue do
		continue = false
		text = gsub(text,"(.*)" .. pattern,
			function(d, e, f, g, h, i)
				if g ~= y and g ~= r then
					continue = true
				end
				return (g == y or g == r) and d .. e .. f .. g .. "ô" .. h .. i or d .. e .. f .. g .. h .. i
			end)
	end
	
	text = gsub(text, "([যডঢ]়)", conv)
	text = gsub(text, "ক্ষ", "kkh")
	text = gsub(text, "জ্ঞ", "gg")
	text = gsub(text, ".", conv)
	
	text = gsub(text, "ː(.)", "%1%1")
	
	text = gsub(text, " ?।", ".")
	
	-- vowel fix
	text = gsub(text, "([āēeo]y)ô ", "%1 ")
	
	-- nasalization fix
	text = gsub(text, "ôN ", "ông ")
	text = gsub(text, "N", "ng")
	
	-- tb / sb -> t / s
	text = gsub(text, "([tsś])b", "%1")
	
	text = gsub(text, " $", "")
	
	return mw.ustring.toNFC(text)
end
 
return export