@@ Line 1: / Line 1: @@
 local p = {}
-local U = mw.ustring
+local arguments = require('Module:Arguments')
-local gsub = U.gsub
+local tt = require('Module:Tt')
-local match = U.match
-local sub = U.sub
-local len = U.len
-local INITIAL = {
+local AUTO_TOOLTIP = 'This romanization was generated automatically. Please add a user-curated romanization as soon as possible.'
+local THAI = '[ก-๛]'
+local consonants = {
 	['ก'] = 'k',
-	['ข'] = 'kh', ['ฃ'] = 'kh', ['ค'] = 'kh', ['ฅ'] = 'kh', ['ฆ'] = 'kh',
+	['ข'] = 'kh',
+	['ฃ'] = 'kh',
+	['ค'] = 'kh',
+	['ฅ'] = 'kh',
+	['ฆ'] = 'kh',
 	['ง'] = 'ng',
 	['จ'] = 'ch',
-	['ฉ'] = 'ch', ['ช'] = 'ch',
+	['ฉ'] = 'ch',
+	['ช'] = 'ch',
 	['ซ'] = 's',
 	['ฌ'] = 'ch',
@@ Line 18: / Line 24: @@
 	['ฎ'] = 'd',
 	['ฏ'] = 't',
-	['ฐ'] = 'th', ['ฑ'] = 'th', ['ฒ'] = 'th',
+	['ฐ'] = 'th',
+	['ฑ'] = 'th',
+	['ฒ'] = 'th',
 	['ณ'] = 'n',
 	['ด'] = 'd',
 	['ต'] = 't',
-	['ถ'] = 'th', ['ท'] = 'th', ['ธ'] = 'th',
+	['ถ'] = 'th',
+	['ท'] = 'th',
+	['ธ'] = 'th',
 	['น'] = 'n',
 	['บ'] = 'b',
@@ Line 36: / Line 46: @@
 	['ล'] = 'l',
 	['ว'] = 'w',
-	['ศ'] = 's', ['ษ'] = 's', ['ส'] = 's',
+	['ศ'] = 's',
+	['ษ'] = 's',
+	['ส'] = 's',
 	['ห'] = 'h',
 	['ฬ'] = 'l',
-	['ฮ'] = 'h'
+	['ฮ'] = 'h',
+	['อ'] = ''
 }
-local FINAL = {
+local finals = {
 	['ก'] = 'k',
-	['ข'] = 'k', ['ฃ'] = 'k', ['ค'] = 'k', ['ฅ'] = 'k', ['ฆ'] = 'k',
+	['ข'] = 'k',
-	['ง'] = 'ng',
+	['ฃ'] = 'k',
+	['ค'] = 'k',
+	['ฅ'] = 'k',
+	['ฆ'] = 'k',
 	['จ'] = 't',
-	['ฉ'] = 't', ['ช'] = 't', ['ซ'] = 't',
+	['ฉ'] = 't',
-	['ฌ'] = '',
+	['ช'] = 't',
-	['ญ'] = 'n',
+	['ซ'] = 't',
-	['ฎ'] = 't', ['ฏ'] = 't',
+	['ฌ'] = 't',
-	['ฐ'] = 't', ['ฑ'] = 't', ['ฒ'] = 't',
+	['ฎ'] = 't',
-	['ณ'] = 'n',
+	['ฏ'] = 't',
+	['ฐ'] = 't',
+	['ฑ'] = 't',
+	['ฒ'] = 't',
 	['ด'] = 't',
 	['ต'] = 't',
-	['ถ'] = 't', ['ท'] = 't', ['ธ'] = 't',
+	['ถ'] = 't',
-	['น'] = 'n',
+	['ท'] = 't',
+	['ธ'] = 't',
+	['ศ'] = 't',
+	['ษ'] = 't',
+	['ส'] = 't',
 	['บ'] = 'p',
 	['ป'] = 'p',
-	['ผ'] = '', ['ฝ'] = '',
+	['ผ'] = 'p',
-	['พ'] = 'p', ['ฟ'] = 'p', ['ภ'] = 'p',
+	['ฝ'] = 'p',
-	['ม'] = 'm',
+	['พ'] = 'p',
-	['ย'] = '',
+	['ฟ'] = 'p',
+	['ภ'] = 'p',
+	['ง'] = 'ng',
+	['ญ'] = 'n',
+	['ณ'] = 'n',
+	['น'] = 'n',
 	['ร'] = 'n',
 	['ล'] = 'n',
-	['ว'] = '',
-	['ศ'] = 't', ['ษ'] = 't', ['ส'] = 't',
-	['ห'] = '',
 	['ฬ'] = 'n',
-	['ฮ'] = ''
+	['ม'] = 'm'
 }
-local CONSONANTS = 'กขฃคฅฆงจฉชซฌญฎฏฐฑฒณดตถทธนบปผฝพฟภมยรลวศษสหฬฮ'
+local vowels = {
+	-- Longest forms first.
+	['เียะ'] = 'ia',
+	['เีย'] = 'ia',
+	['เือะ'] = 'uea',
+	['เือ'] = 'uea',
+	['เือย'] = 'ueai',
+	['เียว'] = 'iao',
+	['ัวะ'] = 'ua',
+	['ัว'] = 'ua',
+	['เิ็ย'] = 'oei',
+	['เิย'] = 'oei',
+	['เย'] = 'oei',
+	['เอย'] = 'oei',
+	['็อย'] = 'oi',
+	['อย'] = 'oi',
+	['โย'] = 'oi',
+	['ุย'] = 'ui',
+	['วย'] = 'uai',
+	['ัย'] = 'ai',
+	['ไย'] = 'ai',
+	['าย'] = 'ai',
+	['ใ'] = 'ai',
+	['ไ'] = 'ai',
+	['เ็ว'] = 'eo',
+	['แ็ว'] = 'aeo',
+	['เา'] = 'ao',
+	['าว'] = 'ao',
+	['เอว'] = 'oeu',
+	['แว'] = 'aeo',
+	['เว'] = 'eo',
+	['โว'] = 'ou',
+	['ิว'] = 'io',
+	['ีว'] = 'io',
+	['ำ'] = 'am',
+	['ือ'] = 'ue',
+	['ื'] = 'ue',
+	['ึ'] = 'ue',
+	['ิ'] = 'i',
+	['ี'] = 'i',
+	['ุ'] = 'u',
+	['ู'] = 'u',
+	['เ็'] = 'e',
+	['เ'] = 'e',
+	['แ็'] = 'ae',
+	['แ'] = 'ae',
+	['โะ'] = 'o',
+	['โ'] = 'o',
+	['เาะ'] = 'o',
+	['อ'] = 'o',
+	['เอะ'] = 'oe',
+	['เอ'] = 'oe',
+	['เะ'] = 'e',
+	['ะ'] = 'a',
+	['ั'] = 'a',
+	['า'] = 'a'
+}
+local vowel_order = {
+	'เือย',
+	'เียว',
+	'เือะ',
+	'เียะ',
+	'เีย',
+	'เือ',
+	'ัวะ',
+	'ัว',
+	'เิ็ย',
+	'เิย',
+	'็อย',
+	'เอว',
+	'แ็ว',
+	'เ็ว',
+	'เอย',
+	'เา',
+	'าว',
+	'เว',
+	'แว',
+	'โว',
+	'าย',
+	'ไย',
+	'ัย',
+	'ุย',
+	'วย',
+	'อย',
+	'โย',
+	'เย',
+	'ำ',
+	'ิ',
+	'ี',
+	'ึ',
+	'ื',
+	'ุ',
+	'ู',
+	'เ็',
+	'แ็',
+	'เาะ',
+	'เอะ',
+	'เะ',
+	'ะ',
+	'ั',
+	'า',
+	'เ',
+	'แ',
+	'โ',
+	'อ',
+	'ใ',
+	'ไ'
+}
+local vowels_with_front = {
+	['เ'] = true,
+	['แ'] = true,
+	['โ'] = true,
+	['ใ'] = true,
+	['ไ'] = true
+}
+local consonant_set = {}
+for c in pairs(consonants) do
+	consonant_set[c] = true
+end
+local valid_clusters = {
+	['กร'] = true,
+	['กล'] = true,
+	['กว'] = true,
+	['ขร'] = true,
+	['ขล'] = true,
+	['ขว'] = true,
+	['คร'] = true,
+	['คล'] = true,
+	['คว'] = true,
+	['ฆร'] = true,
+	['ฆล'] = true,
+	['จร'] = true,
+	['ตร'] = true,
+	['ปร'] = true,
+	['ปล'] = true,
+	['พร'] = true,
+	['พล'] = true,
+	['ผล'] = true,
+	['ทร'] = true,
+	['ธร'] = true,
+	['ศร'] = true,
+	['สร'] = true,
+	['สล'] = true,
+	['สว'] = true,
+	['ซร'] = true,
+	['ฝร'] = true,
+	['ฟร'] = true,
+	['หน'] = true,
+	['หม'] = true,
+	['หย'] = true,
+	['หร'] = true,
+	['หล'] = true,
+	['หว'] = true,
+	['หญ'] = true
+}
+local h_silent_clusters = {
+	['หง'] = true,
+	['หน'] = true,
+	['หม'] = true,
+	['หย'] = true,
+	['หร'] = true,
+	['หล'] = true,
+	['หว'] = true,
+	['หญ'] = true
+}
+local r_silent_clusters = {
+	['จร'] = true,
+	['ซร'] = true,
+	['ศร'] = true,
+	['สร'] = true
+}
+local function chars(text)
+	local result = {}
+	for c in mw.text.gsplit(text, '') do
+		table.insert(result, c)
+	end
+	return result
+end
 local function isConsonant(c)
-	return c and match(c, '[' .. CONSONANTS .. ']') ~= nil
+	return c and consonant_set[c] == true
 end
 local function isThai(c)
-	return c and match(c, '[ก-๛]') ~= nil
+	return c and mw.ustring.find(c, THAI) ~= nil
 end
-local function isVowelMark(c)
+local function isTone(c)
-	return c and match(c, '[ะาิีึืุูัำเแโใไอ]') ~= nil
+	return c == '่' or c == '้' or c == '๊' or c == '๋' or c == '์'
 end
-local function clean(str)
+local function isFrontVowel(c)
-	str = mw.text.unstrip(str or '')
+	return c and vowels_with_front[c] == true
-	str = gsub(str, '<br%s*/?>', ' ')
-	str = gsub(str, '<[^>]+>', '')
-	return str
 end
-local function removeMarks(str)
+local function getVowel(chars_, pos)
-	-- Tone marks are not represented in RTGS.
+	local remaining = table.concat(chars_, '', pos)
-	str = gsub(str, '[่้๊๋]', '')
-	-- Mai thanthakhat makes the marked consonant silent.
+	for _, form in ipairs(vowel_order) do
-	str = gsub(str, '([ก-ฮ])์', '')
+		if mw.ustring.sub(remaining, 1, mw.ustring.len(form)) == form then
+			return form, vowels[form]
+		end
+	end
-	return str
+	return nil, nil
 end
-local function getVowel(str, pos)
+local function skipMarks(chars_, pos)
-	local s = sub(str, pos, -1)
+	while chars_[pos] and (
+		chars_[pos] == '็' or
+		chars_[pos] == '่' or
+		chars_[pos] == '้' or
+		chars_[pos] == '๊' or
+		chars_[pos] == '๋' or
+		chars_[pos] == '์' or
+		chars_[pos] == 'ํ' or
+		chars_[pos] == '๎'
+	) do
+		pos = pos + 1
+	end
-	-- Longest / most specific patterns first.
+	return pos
-	local patterns = {
+end
-		{ 'เ', 'ื', 'อ', 'ย', 'ueai' },
-		{ 'เ', 'ี', 'ย', 'ว', 'iao' },
-		{ 'เ', 'ี', 'ย', 'ะ', 'ia' },
-		{ 'เ', 'ี', 'ย', '', 'ia' },
-		{ 'เ', 'ื', 'อ', 'ะ', 'uea' },
-		{ 'เ', 'ื', 'อ', '', 'uea' },
-		{ '', 'ั', 'ว', 'ะ', 'ua' },
+local function findNextVowel(chars_, pos)
-		{ '', 'ั', 'ว', '', 'ua' },
+	local i = pos
-		{ 'แ', '็', 'ว', '', 'aeo' },
+	while chars_[i] do
-		{ 'แ', 'ว', '', '', 'aeo' },
+		if isFrontVowel(chars_[i]) then
-		{ 'เ', '็', 'ว', '', 'eo' },
+			return i
-		{ 'เ', 'ว', '', '', 'eo' },
+		end
-		{ '', 'ิ', 'ว', '', 'io' },
+		for _, form in ipairs(vowel_order) do
-		{ '', 'ุ', 'ย', '', 'ui' },
+			if mw.ustring.sub(table.concat(chars_, '', i), 1, mw.ustring.len(form)) == form then
-		{ '', 'ว', 'ย', '', 'uai' },
+				return i
+			end
+		end
-		{ 'เ', 'ื', 'อ', '', 'ue' },
+		if chars_[i] == '์' then
-		{ 'เ', 'ื', '', '', 'ue' },
+			i = i + 2
+		else
+			i = i + 1
+		end
+	end
+	return nil
+end
+local function hasVowelAt(chars_, pos)
+	local form = getVowel(chars_, pos)
+	return form ~= nil
+end
-		{ 'เ', 'า', '', '', 'ao' },
+local function getInitial(chars_, pos)
-		{ '', 'า', 'ว', '', 'ao' },
+	local c1 = chars_[pos]
+	if not isConsonant(c1) then
+		return '', pos
+	end
-		{ 'โ', 'ย', '', '', 'oi' },
+	local c2 = chars_[pos + 1]
-		{ '', 'อ', 'ย', '', 'oi' },
-		{ 'เ', 'ย', '', '', 'oei' },
+	if c1 == 'ห' and c2 and h_silent_clusters[c1 .. c2] then
+		return consonants[c2], pos + 2, true, c1 .. c2
+	end
-		{ 'ใ', '', '', '', 'ai' },
+	if c2 and valid_clusters[c1 .. c2] then
-		{ 'ไ', '', '', '', 'ai' },
+		if c1 == 'ท' and c2 == 'ร' then
-		{ '', 'ั', 'ย', '', 'ai' },
+			return 's', pos + 2, false, c1 .. c2
-		{ '', 'า', 'ย', '', 'ai' },
+		end
-		{ 'เ', 'าะ', '', '', 'o' },
+		if r_silent_clusters[c1 .. c2] then
-		{ 'โ', 'ะ', '', '', 'o' },
+			return consonants[c1], pos + 2, false, c1 .. c2
-		{ 'โ', '', '', '', 'o' },
+		end
-		{ '', 'อ', '', '', 'o' },
-		{ 'เ', 'อะ', '', '', 'oe' },
+		return consonants[c1] .. consonants[c2], pos + 2, false, c1 .. c2
-		{ 'เ', 'ิ', '', '', 'oe' },
+	end
-		{ 'เ', 'อ', '', '', 'oe' },
-		{ 'แ', 'ะ', '', '', 'ae' },
+	return consonants[c1], pos + 1, false, c1
-		{ 'แ', '', '', '', 'ae' },
+end
-		{ 'เ', 'ะ', '', '', 'e' },
+local function getCoda(chars_, pos)
-		{ 'เ', '็', '', '', 'e' },
+	local c = chars_[pos]
-		{ 'เ', '', '', '', 'e' },
-		{ '', 'ำ', '', '', 'am' },
+	if not c or not isConsonant(c) then
+		return '', pos
+	end
-		{ '', 'ะ', '', '', 'a' },
+	if c == 'อ' or c == 'ห' or c == 'ฮ' then
-		{ '', 'ั', '', '', 'a' },
+		return '', pos
-		{ '', 'า', '', '', 'a' },
+	end
-		{ '', 'ิ', '', '', 'i' },
+	if c == 'ย' or c == 'ว' then
-		{ '', 'ี', '', '', 'i' },
+		return '', pos
+	end
-		{ '', 'ึ', '', '', 'ue' },
+	if finals[c] then
-		{ '', 'ื', '', '', 'ue' },
+		return finals[c], pos + 1
+	end
-		{ '', 'ุ', '', '', 'u' },
+	return '', pos + 1
-		{ '', 'ู', '', '', 'u' }
+end
-	}
-	for _, v in ipairs(patterns) do
+local function removeSilent(chars_)
-		local a = v[1] or ''
+	local result = {}
-		local b = v[2] or ''
+	local i = 1
-		local c = v[3] or ''
-		local d = v[4] or ''
-		local roman = v[5] or ''
-		local pattern = a .. b .. c .. d
+	while chars_[i] do
+		if chars_[i] == '์' then
+			i = i + 1
-		if pattern ~= '' and sub(s, 1, len(pattern)) == pattern then
+			while chars_[i] and (
-			return roman, len(pattern)
+				isConsonant(chars_[i]) or
+				chars_[i] == '็' or
+				chars_[i] == '่' or
+				chars_[i] == '้' or
+				chars_[i] == '๊' or
+				chars_[i] == '๋'
+			) do
+				i = i + 1
+			end
+		else
+			table.insert(result, chars_[i])
+			i = i + 1
 		end
 	end
-	-- Implied a.
+	return result
-	return 'a', 0
 end
-local function romanizeSyllable(str, start)
+local function cleanWord(word)
+	word = mw.ustring.gsub(word, '์', '')
+	return word
+end
+local function parseSyllable(chars_, start)
+	local n = #chars_
 	local pos = start
-	local n = len(str)
-	local initial = ''
+	if not chars_[pos] then
-	local initialRom = ''
+		return nil, pos
+	end
-	-- Preposed vowels are written before the initial consonant.
+	local front = ''
-	local pre = sub(str, pos, pos)
-	if pre == 'เ' or pre == 'แ' or pre == 'โ' or pre == 'ใ' or pre == 'ไ' then
+	if isFrontVowel(chars_[pos]) then
+		front = chars_[pos]
 		pos = pos + 1
 	end
-	local first = sub(str, pos, pos)
+	local initial, afterInitial, silentH, cluster =
+		getInitial(chars_, pos)
-	if not isConsonant(first) then
+	if not chars_[pos] or not isConsonant(chars_[pos]) then
-		return first, 1
+		return nil, start + 1
 	end
-	initial = first
+	pos = afterInitial
-	initialRom = INITIAL[first] or ''
-	pos = pos + 1
+	local vowelForm
+	local vowel
+	if front ~= '' then
+		local rest = front
+		for _, form in ipairs(vowel_order) do
+			local candidate = front .. mw.ustring.sub(
+				table.concat(chars_, '', pos),
+,
+				mw.ustring.len(form)
+			)
-	-- ห + sonorant consonant: ห is not romanized.
+			if vowels[candidate] then
-	local second = sub(str, pos, pos)
+				vowelForm = candidate
+				vowel = vowels[candidate]
+				pos = pos + mw.ustring.len(form)
+				break
+			end
+		end
-	if initial == 'ห' and second and match(second, '[งนมยรลวญณฬ]') then
+		if not vowel then
-		initial = second
+			if front == 'เ' then
-		initialRom = INITIAL[second]
+				if chars_[pos] == 'า' then
-		pos = pos + 1
+					vowelForm = 'เา'
+					vowel = 'ao'
+					pos = pos + 1
+				else
+					vowelForm = 'เ'
+					vowel = 'e'
+				end
+			elseif front == 'แ' then
+				vowelForm = 'แ'
+				vowel = 'ae'
+			elseif front == 'โ' then
+				vowelForm = 'โ'
+				vowel = 'o'
+			elseif front == 'ใ' or front == 'ไ' then
+				vowelForm = front
+				vowel = 'ai'
+			end
+		end
+	else
+		vowelForm, vowel = getVowel(chars_, pos)
+		if vowelForm then
+			pos = pos + mw.ustring.len(vowelForm)
+		end
 	end
-	-- จร-, ซร-, ศร-, สร-: ร is not romanized.
+	-- No written vowel: determine an implied vowel.
-	second = sub(str, pos, pos)
+	if not vowel then
+		local nextC = chars_[pos]
-	if second == 'ร'
+		if nextC and isConsonant(nextC) then
-		and (initial == 'จ' or initial == 'ซ' or initial == 'ศ' or initial == 'ส') then
+			local nextNext = chars_[pos + 1]
-		pos = pos + 1
+			if nextNext and (
+				isFrontVowel(nextNext) or
+				hasVowelAt(chars_, pos + 1)
+			) then
+				-- Open syllable with an implied a.
+				vowelForm = ''
+				vowel = 'a'
+			else
+				-- Closed syllable with an implied o.
+				vowelForm = ''
+				vowel = 'o'
+			end
+		else
+			vowelForm = ''
+			vowel = 'a'
+		end
 	end
-	-- Read the vowel.
+	pos = skipMarks(chars_, pos)
-	local vowelPos
-	if pre == 'เ' or pre == 'แ' or pre == 'โ' or pre == 'ใ' or pre == 'ไ' then
+	local coda = ''
-		vowelPos = start
+	local codaPos = pos
-	else
-		vowelPos = pos
+	if chars_[pos] and isConsonant(chars_[pos]) then
+		local nextC = chars_[pos]
+		local nextNext = chars_[pos + 1]
+		-- If the following consonant is followed by a vowel,
+		-- it normally starts the next syllable rather than being a coda.
+		if nextNext and (
+			isFrontVowel(nextNext) or
+			hasVowelAt(chars_, pos + 1)
+		) then
+			coda = ''
+		else
+			coda, codaPos = getCoda(chars_, pos)
+		end
 	end
-	local vowel, vowelLen = getVowel(str, vowelPos)
+	pos = codaPos
+	pos = skipMarks(chars_, pos)
+	return {
+		initial = initial,
+		vowel = vowel,
+		vowelForm = vowelForm or '',
+		coda = coda,
+		next = pos,
+		cluster = cluster,
+		silentH = silentH
+	}, pos
+end
+local function shouldHyphen(previous, current)
+	if not previous or not current then
+		return false
+	end
-	-- Advance over the vowel.
+	-- A hyphen is used before a syllable beginning with a vowel.
-	if vowelLen > 0 then
+	if current.initial == '' and current.vowel ~= '' then
-		pos = vowelPos + vowelLen
+		return true
 	end
-	-- If there was a preposed vowel, getVowel consumed from its position,
+	-- A syllable ending in a vowel followed by initial ng.
-	-- so the resulting position already includes it.
+	if previous.coda == '' and current.initial == 'ng' then
-	if vowelLen == 0 and (pre == 'เ' or pre == 'แ' or pre == 'โ' or pre == 'ใ' or pre == 'ไ') then
+		return true
-		pos = start + 1
 	end
-	-- Final consonant.
+	return false
-	local final = sub(str, pos, pos)
+end
-	local finalRom = ''
+local function romanizeWord(word)
+	word = cleanWord(word)
+	local c = chars(word)
+	c = removeSilent(c)
+	local syllables = {}
+	local pos = 1
+	local safety = 0
+	while pos <= #c and safety < 1000 do
+		safety = safety + 1
-	if isConsonant(final) then
+		if not c[pos] then
-		local after = sub(str, pos + 1, pos + 1)
+			break
+		end
-		-- If another vowel begins immediately afterward, this consonant
+		if c[pos] == 'ๆ' then
-		-- belongs to the following syllable.
+			table.insert(syllables, {
-		if not isVowelMark(after) and not (after == 'เ' or after == 'แ' or after == 'โ' or after == 'ใ' or after == 'ไ') then
+				raw = 'ๆ',
-			finalRom = FINAL[final] or ''
+				initial = '',
+				vowel = '',
+				coda = ''
+			})
 			pos = pos + 1
+		elseif isConsonant(c[pos]) or isFrontVowel(c[pos]) then
+			local syllable, nextPos = parseSyllable(c, pos)
-			-- Final -ร after another final consonant.
+			if syllable then
-			if final == 'ร' then
+				table.insert(syllables, syllable)
-				finalRom = 'n'
+				if nextPos <= pos then
+					pos = pos + 1
+				else
+					pos = nextPos
+				end
+			else
+				-- Unknown Thai structure: consume one character so that
+				-- the parser can never become stuck.
+				table.insert(syllables, {
+					raw = c[pos],
+					initial = consonants[c[pos]] or '',
+					vowel = 'a',
+					coda = ''
+				})
+				pos = pos + 1
 			end
+		else
+			pos = pos + 1
 		end
 	end
-	-- ทร- exception.
+	local output = {}
-	if initial == 'ท' and sub(str, pos, pos) == 'ร' then
-		initialRom = 's'
-		pos = pos + 1
-	end
-	-- -รร-
+	for i, s in ipairs(syllables) do
-	-- With a final consonant, รร = a.
+		if s.raw == 'ๆ' then
-	-- Without a final consonant, รร = an.
+			table.insert(output, 'ๆ')
-	if sub(str, vowelPos, vowelPos + 1) == 'รร' then
-		if finalRom ~= '' then
-			vowel = 'a'
 		else
-			vowel = 'an'
+			local result = (s.initial or '') .. (s.vowel or '') .. (s.coda or '')
+			if result == '' then
+				result = 'a'
+			end
+			if shouldHyphen(syllables[i - 1], s) then
+				table.insert(output, '-')
+			end
+			table.insert(output, result)
 		end
 	end
-	return initialRom .. vowel .. finalRom, math.max(1, pos - start)
+	return table.concat(output)
 end
-local function romanizeWord(word)
+local function romanizeThaiText(text)
 	local result = {}
-	local pos = 1
-	local n = len(word)
-	while pos <= n do
+	local parts = mw.text.split(
-		local c = sub(word, pos, pos)
+		text,
+		'(\1MANUALHYPHEN\1|\1MANUALSPACE\1)',
+		true
+	)
+	for _, part in ipairs(parts) do
+		if part == '\1MANUALHYPHEN\1' then
+			table.insert(result, '-')
+		elseif part == '\1MANUALSPACE\1' then
+			table.insert(result, ' ')
+		else
+			local last = 1
-		if c == '<' then
+			while true do
-			local close = mw.ustring.find(word, '>', pos + 1, true)
+				local startPos, endPos =
+					mw.ustring.find(part, THAI .. '+', last)
-			if close then
+				if not startPos then
-				local marker = sub(word, pos, close)
+					table.insert(result, mw.ustring.sub(part, last))
+					break
+				end
-				if marker == '< >' then
+				if startPos > last then
-					table.insert(result, ' ')
+					table.insert(
-				elseif marker == '<->' then
+						result,
-					table.insert(result, '-')
+						mw.ustring.sub(part, last, startPos - 1)
-				else
+					)
-					table.insert(result, marker)
 				end
-				pos = close + 1
+				local word =
-			else
+					mw.ustring.sub(part, startPos, endPos)
-				table.insert(result, c)
-				pos = pos + 1
+				table.insert(result, romanizeWord(word))
+				last = endPos + 1
 			end
-		elseif isThai(c) then
-			local roman, consumed = romanizeSyllable(word, pos)
-			table.insert(result, roman)
-			pos = pos + consumed
-		else
-			table.insert(result, c)
-			pos = pos + 1
 		end
 	end
@@ Line 336: / Line 740: @@
 end
-local function capitalizeWords(str)
+local function hasThai(text)
-	local chars = mw.text.split(str, '')
+	return mw.ustring.find(text, THAI) ~= nil
-	local capitalize = true
+end
-	for i, c in ipairs(chars) do
+local function containsThai(text)
-		if match(c, '[A-Za-z]') then
+	return mw.ustring.find(text, '[ก-๛]') ~= nil
-			if capitalize then
+end
-				chars[i] = U.upper(c)
-				capitalize = false
+local function hasOnlyThaiOrFormatting(text)
-			end
+	return not mw.ustring.find(text, '[ก-๛]')
-		elseif c == ' ' or c == '-' then
+end
-			capitalize = true
+local function validateRomanization(text)
+	-- The automatic result must never leak Thai characters.
+	if containsThai(text) then
+		return false
+	end
+	-- Thai marks should never survive either.
+	if mw.ustring.find(text, '[ะ-๛]') then
+		return false
+	end
+	return true
+end
+local function capitalizeWords(text)
+	return mw.ustring.gsub(
+		text,
+		'(%a)([%a]*)',
+		function(first, rest)
+			return mw.ustring.upper(first) .. rest
 		end
+	)
+end
+local function capitalizeSentence(text)
+	local first = mw.ustring.match(text, '^[%s%p]*([a-z])')
+	if not first then
+		return text
+	end
+	local upper = mw.ustring.upper(first)
+	return mw.ustring.gsub(
+		text,
+		'^([%s%p]*)' .. first,
+		'%1' .. upper,
+	)
+end
+local function convertThaiDigits(text)
+	local digits = {
+		['๐'] = '0',
+		['๑'] = '1',
+		['๒'] = '2',
+		['๓'] = '3',
+		['๔'] = '4',
+		['๕'] = '5',
+		['๖'] = '6',
+		['๗'] = '7',
+		['๘'] = '8',
+		['๙'] = '9'
+	}
+	for thai, latin in pairs(digits) do
+		text = mw.ustring.gsub(text, thai, latin)
 	end
-	return table.concat(chars)
+	return text
 end
-local function sentenceCase(str)
+function p.romanize(text)
-	local chars = mw.text.split(str, '')
+	if not text or mw.text.trim(text) == '' then
-	local capitalize = true
+		return ''
+	end
-	for i, c in ipairs(chars) do
+	if not hasThai(text) then
-		if match(c, '[A-Za-z]') then
+		return ''
-			if capitalize then
+	end
-				chars[i] = U.upper(c)
-				capitalize = false
+	-- Manual spacing/hyphen controls.
-			end
+	text = text
-		elseif match(c, '[.!?]') then
+		:gsub('<->', '\1')
-			capitalize = true
+		:gsub('<>', '\2')
-		end
+	text = convertThaiDigits(text)
+	local result = romanizeThaiText(text)
+	-- Restore manual separators.
+	result = result
+		:gsub('\1', '-')
+		:gsub('\2', ' ')
+	result = mw.ustring.gsub(result, '%s+', ' ')
+	result = mw.text.trim(result)
+	if not validateRomanization(result) then
+		return ''
 	end
-	return table.concat(chars)
+	return result
 end
 function p.main(frame)
-	local args = require('Module:Arguments').getArgs(frame, {
+	local args = arguments.getArgs(frame, {
 		parentFirst = true,
 		wrappers = {'Template:Th-rm'}
 	})
-	if not args[1] or args[1] == '' then
+	local text = args[1]
+	if not text or mw.text.trim(text) == '' then
 		return ''
 	end
-	local str = clean(args[1])
+	-- If there is no Thai at all, this module should not create
+	-- an automatic romanization.
-	-- Nothing to romanize.
+	if not hasThai(text) then
-	if not match(str, '[ก-๛]') then
 		return ''
 	end
-	str = removeMarks(str)
+	local result = p.romanize(text)
-	local result = romanizeWord(str)
+	if result == '' then
+		return ''
+	end
 	if args.capi then
 		result = capitalizeWords(result)
-	elseif args.sent then
-		result = sentenceCase(result)
 	end
-	return result
+	if args.sent then
+		result = capitalizeSentence(result)
+	end
+	-- Automatically generated romanizations are deliberately marked
+	-- with the tooltip from Module:Tt.
+	return tt._main({
+		[1] = result,
+		[2] = AUTO_TOOLTIP,
+		['class'] = args.class
+	})
 end
 return p

Module:Th-rm: Difference between revisions

ReisuDesign (talk | contribs)