维基词典

模組:Shah-translit

下列說明文檔位於Module:Shah-translit/doc。[編輯]
相關連結：子頁面列表 • 链入 • 嵌入包含 • 測試用例 • 沙盒

這個模組會將沙穆基文文字轉寫為拉丁字母。

最好不要直接從模板或其他模組調用此模組。要從模板中使用它，請以{{xlit}}做為替代；若要在模組中使用，則以Module:languages#Language:transliterate替代。

關於測試用例，請參閱Module:Shah-translit/testcases。

函數

编辑
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.
-- TODO: long í support
-- TODO: sort out short e vs palatalisation
-- TODO: add rule for CẹC = CyaC

local u = mw.ustring.char
local gsub = mw.ustring.gsub

local export = {}

local vav = u(0x0648)
local ye = u(0x06CC)
local alif = u(0x0627)
local he = 'ه'

local vw_s_cfu = u(0x0650) -- pesh (i)
local vw_s_ccu = u(0x0655) -- hamza below
local vw_s_cbr = u(0x064F) -- zer (u)
local vw_s_mcu = u(0x0654) -- hamza above
local vw_s_ocu = u(0x064E) -- zabar (a)

local vw_l_cbr = u(0x0657) -- inverted zer
local vw_l_cfu = u(0x0656) -- subscript alif

local hat = u(0x065A)
local inverted_hat = u(0x065B)
local hats = hat .. inverted_hat

local short_vowels_list = vw_s_cfu .. vw_s_ccu .. vw_s_cbr .. vw_s_mcu .. vw_s_ocu 

-- carrier + diacritic combos
local long_u = vav .. vw_l_cbr
local short_o = vav .. inverted_hat
local long_i = ye .. vw_l_cfu
local short_e = ye .. inverted_hat

local vocalised_carrier = short_a .. long_a .. short_i .. long_i .. short_e .. long_e .. short_o .. short_u .. 'ا' .. 'آ' .. ' ِ◌' .. 'ی' .. ' ُ◌' .. 'ے' .. 'و' 

local consonants = "بپتثجچحخدذرزژسشصضطظعغفقکگلمنڻوهىھٹڈڑھتھڈھجھدھٹھکھڑھ"

local conv = {
	--consonants
	['ب'] = 'b', ['پ'] = 'p', ['ت'] = 't', ['ث'] = 's', ['ج'] = 'j', 
	['چ'] = 'ch', ['ح'] = 'h', ['خ'] = 'kḥ', ['د'] = 'd', ['ذ'] = 'z',
	['ر'] = 'r', ['ز'] = 'z', ['ژ'] = 'z', ['س'] = 's', ['ش'] = 'sh',
	['ص'] = 's', ['ض'] = 'z', ['ط'] = 't', ['ظ'] = 'z', ['ع'] = 'a',
	['غ'] = 'ġ', ['ف'] = 'f', ['ق'] = 'q', ['ک'] = 'k', ['گ'] = 'g',
	['ل'] = 'l', ['م'] = 'm', ['ن'] = 'n', ['ڻ'] = 'ṅ', ['و'] = 'v',
	['ه'] = 'h', ['ى'] = 'y', ['ھ'] = 'h', ['ٹ'] = 'ṭ', ['ڈ'] = 'ḍ',
	['ڑ'] = 'ṛ', ['ھ'] = 'bh', ['تھ'] = 'th', ['ڈھ'] = 'ḍh', ['جھ'] = 'jh',
	['دھ'] = 'dh', ['ٹھ'] = 'ṭh', ['کھ'] = 'kh', ['ڑھ'] = 'ṛh', ['وا'] = 'wā',
	
	-- vowels
	[u(0x0627)] = 'ā', -- alif
	[' ِ◌'] = 'i', ['ی'] = 'ī' ,
	[' ُ◌'] = 'u',
	['ے'] = 'e', ['ے'] = 'ē',
	
	-- other diacritics
	['ں'] = 'N', --noon ghunna: nasalize
	['ء'] = 'ʔ', --hamza:glottal t
	[' ّ◌'] = 'ː', --shad: emphasise
	[u(0x0670)] = 'ā', -- khari zabar
	[u(0x064B)] = 'N', -- do zabar
	[' ٍ◌'] = 'in', -- do zer

	-- digits
	['۰'] = '0', ['۱'] = '1', ['۲'] = '2', ['۳'] = '3', ['۴'] = '4',
	['۵'] = '5', ['۶'] = '6', ['۷'] = '7', ['۸'] = '8', ['۹'] = '9', 
}

local nasal_assim = {
	['[kg]h?'] = 'ṅ',
	['[cj]h?'] = 'ñ',
	['[ṭḍ]h?'] = 'ṇ',
	['[td]h?'] = 'n',
	['[pb]h?'] = 'm',
	['n'] = 'n',
	['m'] = 'm',
    ['s'] = 'n',
}

local short_vowels = {
	-- high vowels
	[u(0x0627)] = 'ā',	[u(0x0650)] = 'i', [u(0x0655)] = 'ụ', [u(0x064F)] = 'u', [u(0x065F)] = 'ụ̄', [u(0x0648)] = 'ū',
	-- central vowels
	[u(0x0654)] = 'ạ',
	
	-- low vowels
	[u(0x064E)] = 'a',
	
}

function export.tr(text, lang, sc)
local waw = 'و'
local ye = 'ی'
local noon = '[ًں]?'
local shadda = ' ّ'

	text = text .. " "
	
	--two consonants
	
	text = mw.ustring.gsub(text,
		consonants .. consonants,
		"%1a%2") --example ਕਰ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. consonants,
		"%1%2%3") --example ਸ਼ੇਰ
	
	text = mw.ustring.gsub(text,
		consonants .. consonants ..  short_vowel ,
		"%1a%2%3") --example ਦਰੀ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. consonants ..  short_vowel ,
		"%1%2%3%4") --example ਖੀਰਾ 

		-- shadda
	text = mw.ustring.gsub(text,
		consonants .. shadda .. consonants,
		"%1a%2%3") --example ਰੱਬ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. shadda .. consonants,
		"%1%2%3%4") --example ਸਿੱਖ 
	
	text = mw.ustring.gsub(text,
		consonants .. shadda .. consonants ..  short_vowel ,
		"%1a%2%3%4") --example ਗੱਡੀ 
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. shadda .. consonants ..  short_vowel ,
		"%1%2%3%4%5") --example ਤਿੱਖਾ 
	
				-- noon
	text = mw.ustring.gsub(text,
		consonants .. noon .. consonants,
		"%1a%2%3") --example ਬੰਦ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. noon .. consonants,
		"%1%2%3%4") --example ਜਿੰਦ 
	
	text = mw.ustring.gsub(text,
		consonants .. noon .. consonants ..  short_vowel,
		"%1a%2%3%4") --example ਬੰਦੀ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. noon .. consonants ..  short_vowel,
		"%1%2%3%4%5") --example ਜ਼ਿੰਦਾ
	
	
	--three consonants
	
	text = mw.ustring.gsub(text,
		consonants .."a" .. consonants .. consonants,
		"%1a%2a%3") --example ਕਦਰ 
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. consonants .. consonants,
		"%1%2%3a%4") --example ਜਿਸਮ
	
	text = mw.ustring.gsub(text,
		consonants .. consonants ..  short_vowel .. consonants,
		"%1a%2%3%4") --example ਤਸੀਰ
	
	text = mw.ustring.gsub(text,
		consonants .. consonants .. consonants ..  short_vowel,
		"%1a%2%3%4") --example ਤਸਬੀ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. consonants  .. "a" .. consonants ..  short_vowel,
		"%1%2%3%4%5") --example ਖ਼ਾਲਸਾ 
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. consonants ..  short_vowel .. consonants ..  short_vowel,
		"%1%2%3%4%5%6") --example ਖ਼ੁਲਾਸਾ
	
		-- shadda
		
	text = mw.ustring.gsub(text,
		consonants .. shadda .. consonants .. "a" .. consonants ..  short_vowel,
		"%1%2%3%4%5%6") --example ਨੱਠਣਾ
		
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. shadda .. consonants .. "a" .. consonants ..  short_vowel,
		"%1%2%3%4%5%6") --example ਸਿੱਖਣਾ
	
		-- noon
		
	text = mw.ustring.gsub(text,
		consonants .. noon .. consonants .. consonants,
		"%1a%2%3a%4") --example ਜੰਗਲ
	
	text = mw.ustring.gsub(text,
		consonants .. noon .. consonants ..  short_vowel,
		"%1a%2%3%4") --example ਰੰਗਲਾ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. noon .. consonants .. consonants ..  short_vowel,
		"%1%2%3%4a%5%6") --example ਜ਼ਿੰਦਗੀ
	
	--four consonants
		
	text = mw.ustring.gsub(text,
		consonants .. consonants .. consonants .. consonants,
		"%1a%2%3a%4") --example ਸਰਬਤ
	
	text = mw.ustring.gsub(text,
		consonants ..  short_vowel .. consonants .. consonants .. consonants,
		"%1%2%3%4a%5") --example ਕੁਦਰਤ
	
	text = mw.ustring.gsub(text,
		consonants .. "a" .. consonants ..  short_vowel .. consonants .. consonants,
		"%1a%2%3%4a%5") --example ਕਬੂਤਰ
	
	text = mw.ustring.gsub(text,
		consonants .. consonants .. consonants ..  short_vowel .. consonants,
		"%1a%2%3%4%5") --example 
	
	text = mw.ustring.gsub(text,
		consonants .. consonants .. consonants .. consonants ..  short_vowel,
		"%1a%2%3%4%5") --example 
	
	text = mw.ustring.gsub(text, 
		consonants .. consonants .. consonants ..  short_vowel .. consonants ..  short_vowel,
		"%1a%2%3%4%5%6") --example ਦਰਵਾਜ਼ਾ
	
			-- noon

	text = mw.ustring.gsub(text, 
		consonants .. "a" .. noon .. consonants .. "a" .. consonants ..  short_vowel .. consonants,
		"%1a%2%3%4%5%6") --example ਧੰਨਵਾਦ 
	
	text = mw.ustring.gsub(text, 
		consonants .. noon .. consonants .. consonants ..  short_vowel .. consonants,
		"%1a%2%3%4%5") --example ਫ਼ਰਜ਼ੰਦ
	
		--shadda
		
	text = mw.ustring.gsub(text, 
		consonants .. "a" .. shadda .. consonants .. "a" .. consonants .. consonants,
		"%1a%2%3%4a%5") --example ਯੱਕਦਮ
	
	-- five consonants

	text = mw.ustring.gsub(text, 
		consonants ..  short_vowel .. consonants .. consonants .. "a" .. consonants ..  short_vowel .. consonants,
		"%1%2%3a%4%5%6%7") --example ਵੇਪਰਵਾਹ
	
	text = mw.ustring.gsub(text, 
		consonants ..  short_vowel .. consonants .. "a" .. consonants .. consonants .. "a" .. consonants ..  short_vowel,
		"%1%2%3%4a%5%6%7") --example ਪਾਰਦਰਸ਼ੀ
	
	--shadda
	
	text = mw.ustring.gsub(text, 
		consonants  ..  short_vowel .. shadda  ..  consonants ..  consonants .. consonants ..  short_vowel .. consonants,
		"%1%2%3%4a%5%6%7%8") --example ਸ਼ੁੱਕਰਵਾਰ
	
	text = mw.ustring.gsub(text, 
		consonants .. "a" .. consonants .. consonants .. "a" .. consonants .. "a" .. shadda .. consonants,
		"%1a%2a%3%4a%5%6") --example ਮਗਰਮੱਛ
	
	text = mw.ustring.gsub(text, 
		consonants ..  short_vowel .. consonants ..  short_vowel .. consonants .. consonants .. "a" .. consonants .. consonants ..  short_vowel,
		"%1%2%3%4%5a%6a%7%8%9") --example ਯੂਨੀਵਰਸਟੀ
	
	-- vocalised_carrier
	
	text = mw.ustring.gsub(text, 
		vocalised_carrier .. consonants .. consonants ,
		"%1%2a%3") --example ਅਦਬ
	
	text = mw.ustring.gsub(text, 
		vocalised_carrier .. consonants .. "a".. consonants ..  short_vowel ,
		"%1%2%3%4") --example ਆਖਣਾ
	
	text = mw.ustring.gsub(text, 
		vocalised_carrier .. noon .. consonants .. "a".. consonants ..  short_vowel ,
		"%1%2%3%4%5") --example ਅੰਤੜੀ
	
	text = mw.ustring.gsub(text, 
		vocalised_carrier .. consonants .. "a".. consonants .. consonants ,
		"%1%2%3a%4") --example ਅਫ਼ਸਰ

	text = mw.ustring.gsub(text, 
		vocalised_carrier .. consonants .. "a" .. consonants ..  short_vowel .. consonants ..  short_vowel ,
		"%1%2%3%4%5%6") --example ਅਲਮਾਰੀ
	
	text = mw.ustring.gsub(text, 
		vocalised_carrier .. consonants .. "a" .. consonants ..  short_vowel .. consonants ,
		"%1%2%3%4%5") --example ਇਕਰਾਰ
	
	text = mw.ustring.gsub(text, 
		vocalised_carrier .. " ص" .. " ت" .. "a" .. consonants ..  short_vowel ,
		"%1st%2%3") --example ਉਸਤਰਾ
	
	text = mw.ustring.gsub(text, 
		consonants ..  short_vowel .. vocalised_carrier .. consonants .. consonants ,
		"%1%2%3%4a%5") --example ਸਾਈਕਲ
	
	text = mw.ustring.gsub(text, 
		consonants ..  short_vowel .. consonants ..  short_vowel .. vocalised_carrier .. consonants .. "a" .. consonants ..  short_vowel,
		"%1%2%3%4%5%6%7%8") --example ਵਿਦਿਆਰਥੀ
	
	for key,val in pairs(nasal_assim) do
		text = mw.ustring.gsub(text,"N("..key..")",val.."%1")
	end
	text = mw.ustring.gsub(text,"([aiueēoāīū])N ", "%1̃ ")
	text = mw.ustring.gsub(text,"(.?)N", "%1̃")
	
	text = mw.ustring.gsub(text,"ː(.)","%1%1")
	
	text = mw.ustring.gsub(text," ?।",".")
	
	text = mw.ustring.gsub(text," $","")

	return text
end

return export
来自“https://zh.wiktionary.org/w/index.php?title=Module:Shah-translit&oldid=6068193”