local export = {}
local data = mw.loadData("Module:wuu-pron/data")

--[=[
TODO:
- do IPA for glottalised nasal intials (currently the glottal stop is dropped)
- FIND DATA FOR 3+ SYLLABLE SANDHI AND RPS!!! (we can settle for trisyllabics right)
- northern wu data: FIRST PRIORITY should be
  - CHUANSHA 川沙 (sandhi found in 當代, needs. understanding)
  - NEW KUNSHAN 崑山新派 (trisyllabic sandhi needed)
  - HUZHOU 湖州 (trisyllabic sandhi needed)
  - XIAOSHAN 蕭山 (大西博子 source doesn't exactly line up with wugniu 市志)
  - CIXI 慈溪 (慈溪方言研究 PDF exists but currently unretrieved, please inform @ND381 if a working PDF is in your possession ASAP!! zia-zia)
- any other northern wu points are good!! pls note the source u used in the EDIT SUMMARY if u do add one zia-zia
- MODULES for TAIZHOU & WENZHOU
→ and then we're done (probably)
]=]--

local loc_names = {     		-- Basic 聲韻調    Disyll LPS    Trisyll LPS    Quad+ LPS    RPS    Logic
	--NORTHERN
	  = 'Shanghai',		--      +               +             +             +         +       +
	  = 'Jiading',		--      +               +             +             -         -       +
	  = 'Songjiang',		--      +               +             +             -         -       +
	  = 'Chuansha',		--		+				+			  +				-		  -		  -
	  = 'Chongming',		--      +               +             +             -         +       +
	  = 'Suzhou', 		--      +               +             +             +         +       +
	  = 'Kunshan',		--      +               +             -             -         -       -    Wugniu "新派"
	  = 'Yixing',			--		+				-			  -				-		  -		  -
	  = 'Changzhou',		--      +               +             +             -         -       +
	  = 'Jingjiang',		--		+				+			  -				-		  -		  -
	  = 'Jiaxing',		--      +               +             +             -         -       +
	  = 'Tongxiang',		--      +               +             +             -         -       +
	  = 'Haining', 		--      +               +             +             -         +       +    Xiashi locality
	  = 'Haiyan',  		--      +               +             +             -         -       +
	  = 'Deqing',			--		+				-			  -				-		  -		  -
	  = 'Hangzhou',		--      +               +             +             +         +       +
	  = 'Xiaoshan',		--		+				-			  -				-		  -		  -
	  = 'Shaoxing',		--      +               +             +             -         +       +
	  = 'Cixi',			--		+				+			  -				-		  -		  -
	  = 'Ningbo',  		--      +               +             +             -         +       +
		= 'Zhoushan',		--		+				-			  -				-		  -		  -
	
--            Jinhua        			+               +             +             +         +       +

--	TAIZHOU
--			  Huangyan					+				+			  -			    -				  -
--			  Linhai					+				+			  -			    -				  -
--			  Sanmen					+				-			  -			    -				  -
--			  Tiantai				   +/-				+			  -			    -				  -
--			  Wenling					+				+			  -			    -				  -

--            Wenzhou   			    need to reconcile differences between sources
}

-- default to "<loc_name> dialect" if empty
-- specifies the name of the Wikipedia article of the lect
local wiki_names = {
	jd = 'Shanghainese#Classification',
	sj = 'Shanghainese#Classification',
	cs = 'Shanghainese#Classification',
	yx = 'Taihu Wu',
	jj = 'Taihu Wu',
	jx = 'Taihu Wu',
	tx = 'Taihu Wu',
	hn = 'Taihu Wu',
	hy = 'Taihu Wu',
	dq = 'Taihu Wu',
	xs = 'Taihu Wu',
	cx = 'Taihu Wu',
	zs = 'Taihu Wu',
}

local minidict = {
	 = true,
	 = true,
	 = true,
	 = true,
	 = true,
	 = true,
	 = true,
	 = true,
	 = true
}

local order = {'sh', 'jd', 'sj', 'cm', 'cs', 'sz', 'ks', 'yx', 'cz', 'jj', 'jx', 'tx', 'hn', 'hy', 'dq', 'hz', 'xs', 'sx', 'cx', 'nb', 'zs'}

local ipa_initial = {
	 = "p",  = "pʰ",  = "b",  = "m",  = "f",  = "v",
	 = "t",  = "tʰ",  = "d",  = "n",  = "l",
	 = "t͡s",  = "t͡sʰ",  = "s",  = "z",  = "t͡ɕ",  = "t͡ɕʰ",
	 = "d͡z",  = "d͡ʑ",  = "n̠ʲ",  = "ɕ",  = "ʑ",
	 = "k",  = "kʰ",  = "ɡ",  = "ŋ",  = "h",  = "ɦ",
	 = "",
}

local ipa_initial_override = {
	--this always takes priority over the table above
	--additional unique initials can also be defined here
	--a question mark means the initial does not exist
	--there must be empty tables for all locations, even if there is nothing there
	 = {
		 = "?"
	},
	 = {
		 = "?",  = "?"
	},
	 = {
		 = "ɓ",  = "ɗ",
		 = "ɸ",  = "β",
		 = "cʰ",  = "c",  = "ɟ",  = "ç",
		 = "?",  = "?"
	},
	 = {
		 = "fv",  = "sz",  = "ɕʑ",  = "hɦ",
	},
	 = {
		 = "ɓ",  = "ɗ",
		 = "ɸ",  = "β",
		 = "?"
	},
	 = {
		 = "?",  = "?"
	},
	 = {
		 = "?",  = "x"	
	},
	 = {
		
	},
	 = {
		
	},
	 = {
		
	},
	 = {
		 = "?",  = "ʔv"
	},
	 = {
		
	},
	 = {
		
	},
	 = {
		 = "?"
	},
	 = {
		 = "?"	
	},
	 = {
		 = "?"
	},
	 = {
		
	},
	 = {
		
	},
	 = {
		
	},
	 = {
		
	},
	 = {
		
	},
}

local function get_initial(initial, loc)
	return ipa_initial_override or ipa_initial or error('Invalid initial: "' .. initial .. '"')
end

local function get_final(final, loc)
	return data.ipa_final or error('Invalid final for ' .. loc .. ' : "' .. final .. '"')
end

local ipa_syllabic = {
	 = "m̩",  = "n̩",  = "ŋ̍",
}

-- diagnose tone error
local function diagnose_tones(word_length, loc, text, tone, tone2, tone3)
	-- the cap on number of syllables
	local syl_cap = ({sh=5,sj=3,cm=3,cs=3,sz=4,ks=2,yx=1,cz=3,jj=2,jx=3,tx=3,hn=3,hy=3,dq=1,hz=5,xs=1,sx=3,cx=2,nb=3,zs=1})
	if syl_cap and word_length > syl_cap then
		error(("Maximum %d syllables supported for %s."):format(syl_cap, loc))
	end
	-- the cap on number of specified tones
	local tone_cap = ({sj=3,ks=2,cz=3,jx=3,tx=3,hn=3,hy=2})
	if tone_cap then
		local expected = math.min(tone_cap, word_length)
		local received = 1 + (tone2 ~= '' and 1 or 0) + (tone3 ~= '' and 1 or 0)
		if received ~= expected then
			error(('Expected %d tones, but received %d: "%s:%s".'):format(expected, received, loc, text))
		end
	elseif loc == 'sz' or loc == 'sx' then
		-- sz: tone is 7 or 8, but second tone not provided
		error("For " .. loc .. ", second tone must be specified.")
	end
	error(('Incorrect tone notation "%s" for %s. See ].'):format(tone..tone2..tone3, loc))
end

local function tone_superscript(text)
	return text:gsub('',{='¹',='²',='³',='⁴',='⁵'})
end

local function get_tone(text, loc)
	local word_length = text:gsub("+", ""):len() + 1
	local tone, tone2, tone3 = text:match("^(.%u*)%w+ ?(%d?%u?)%w* ?(%d?%u?)")
	if loc == "jx" and tone == "3" then
		tone = text:find("^3s?h") and "3B" or "3A"
	elseif loc == "cm" then
		local result = nil
		if tone:find("") then -- Verb + Motion / Verb + Pronoun
			if word_length ~= 2 then error("cm: Unsupported word length.") end
			result = data.tone_contours or error("cm: Wrong motion/pronoun format.")
		elseif tone:find("R",1,true) then -- Reduplication
			local main_tone, redup_type, word, sub_tone = text:match("^(%d)R()(%l+) (%d)%3$")
			main_tone, sub_tone = tonumber(main_tone), tonumber(sub_tone)
			local conv_tone = (redup_type == "N" and main_tone%2==0 and word:find("^g?") and main_tone-1) or main_tone
			if sub_tone ~= conv_tone then error("cm: Wrong reduplication format.") end
			result = data.tone_contours
		end
		if result then
			return tone_superscript(result)
		end
	elseif loc == "sx" and tone:find("^%dA$") then
		return tone_superscript(data.tone_contours)
	elseif loc == "cs" and word_length == 3 then
		tone2 = ""
	end
	local result = data.tone_contours
		or data.tone_contours
		or data.tone_contours
	return result and tone_superscript(result) or diagnose_tones(word_length, loc, text, tone, tone2, tone3)
end

local function RPS_tone_determ(word_length, tone, loc)
	local result
	if word_length == 1 then
		result = data.tone_contours or data.tone_contours
	else
		result = data.tone_contours
	end
	return tone_superscript(result)
end

local function rom_check(text, locs) --this checks wugniu
	if text:match("%f") or text:match('ny') or text:match('hh') or text:match("h$") then
		error('Invalid syllable: ' .. text ..'. Wugniu expected, but another romanisation is being provided.')
	end
	if text:match('ghi') and locs ~= 'cm' then
		error('Invalid initial "ghi". Use "yi" instead.')
	end
	if text:match('ghu') and locs ~= 'cm' then
		error('Invalid initial "ghu". Use "wu" instead.')
	end
	if text:match('%fy%f') then
		error('Invalid syllable "y"')
	end
	if text:match('%fy') then
		error('Invalid syllable "yn" or "yq"')
	end
	if text:match('gn') then
		error('Palatalization expected. Insert an "i" after the "gn".')
	end
	if text:match('uw') then
		error(('Invalid syllable in "%s".'):format(text))
	end
	if locs:find('cm') and (text:find('ueu') or text:find('uon') or text:find('ui')) then
		error('cm: Mutation-only final found.')
	end
	if locs:find('sh') and text:match('') then
		require("Module:debug").track("wuu-pron/sh-tone-234")
		--error('sh: Incorrect tone number used.')
	end
	for syl in text:gmatch("+") do
		if not syl:match("%d") then
			require("Module:debug").track("wuu-pron/no-tone")
			if locs ~= "sh" then
			require("Module:debug").track("wuu-pron/no-tone-other")
			end
		end
	end
	return nil
end

function export.ipa_syl_conv(text, loc, initials, finals, syllabics, i, main_tone, tone)
	-- get ipa from tables
	local initial, final = text:match("^(??)(.+)$")
	local if_syllabic = syllabics
	if loc == 'sx' and text == 'gn' then if_syllabic = "ɲ̩" end
	if not initial or if_syllabic then
		initial, final = '', text
	end
	if loc == 'cm' then -- mutation
		local mutated_initial = i > 1 and initial == "z" and "z"
		local preglottal = ""
		if tone ~= "0" and (mutated_initial or initial:find("^g?") or initial == "") then
			preglottal = (i > 1 or main_tone:find("^$")) and "ʔ" or "ɦ"
		end
		return preglottal
			.. (mutated_initial or initials(initial,loc))
			.. (if_syllabic or finals(final,loc))
	end
	return initials(initial,loc) .. (if_syllabic or finals(final,loc))
end

function export.wugniu_to_ipa(original_text, loc, initials, finals, syllabics, tones)
	local text, conv_text = "", ""
	local tone_number = ""
	original_text = original_text:gsub(" (%l+)(%d%u?)", ' %2%1')
	if loc == 'cm' then
		original_text = original_text:gsub("%fyi?","i"):gsub("%fwu?","u")
	else
		original_text = original_text:gsub("%fyi?","ghi"):gsub("%fwu?","ghu")
	end
	local reading = mw.text.split(original_text, ",", true)
	local syllable = {}
	local syl_tone = {}
	for reading_index = 1, #reading, 1 do
		local components = mw.text.split(reading, "&", true)
		for component_index = 1, #components do
			local indep_words = mw.text.split(components, "+", true)
			for indep_index = 1, #indep_words do
				text = indep_words
				tone_number = text:sub(1, 1)
				local tone = tones(text, loc)
				text = text:gsub("+", "")
				local syllable = mw.text.split(text, " ", true)
				local syl_tone = mw.text.split(tone, " ", true)
				for i = 1, #syllable, 1 do
					--RPS
					if i == #syllable and indep_words and tone ~= "³³" then
						syl_tone = RPS_tone_determ(#syllable, tone_number, loc)
					end
					syllable = (syllable ~= "" and export.ipa_syl_conv(syllable, loc, initials, finals, syllabics, i, tone_number, syl_tone) or "")
						.. (syl_tone == "0" and "" or syl_tone)
				end
				indep_words = table.concat(syllable, " ")
			end
			components = table.concat(indep_words, " &nbsp;")
		end
		reading = table.concat(components, " ")
	end
	return table.concat(reading, "/, /")
end

function export.wikt_to_wugniu(text)
	require("Module:debug").track("wuu-pron/legacy")
	if type(text) == "table" then text = text.args end
	return text
	--initials
		:gsub("'+", {=""})
		:gsub("%f?", {j="c", jj="j", q="ch", x="sh", xx="zh"})
		:gsub("%fny", "gn")
		:gsub("%fhh", "gh")

	--vowels
		:gsub("un", "uen")
		:gsub("yoe", "ioe")
		:gsub("y", "iu")
		:gsub("aan", "aon")
		:gsub("%fr", "y")

	--syllabics
		:gsub("g?h?mm", "m")
		:gsub("g?h?ngg", "ng")

	--tones	
		:gsub("", {='5', ='6', ='7', ='8'})
		
	--gh rules
		:gsub("ghi", "yi")
		:gsub("yi%f", "y")
		:gsub("ghu", "wu")
		:gsub("wu%f", "w")
end

local function wugniu_to_wikt(text)
	if type(text) == "table" then text = text.args end
	--initials
	return export.wugniu_format(text
		:gsub("%f?", {c="j", ch="q", j="jj", sh="x", zh="xx", gn="ny", gh="hh"})
		:gsub("%fyi?", "hhi")
		:gsub("wu?", "hhu")

	--vowels
		:gsub("y%f", "r")
		:gsub("uen", "un")
		:gsub("ioe", "yoe")
		:gsub("iu", "y")
		:gsub("aon", "aan")

	--syllabics
		:gsub("%fg?%f", {m="mm",n="nn",ng="ngg"})

	--initial hh and '
		:gsub("()()", "%1'%2")
		:gsub("()(g?)%f", "%1hh%2")

	--tones	
		:gsub("", {='2', ='3', ='4', ='5'}))
end

--[[學堂拼音 → 錢拼
local function wugniu_to_qian(text)
	if type(text) == "table" then text = text.args end
	--initials
	return export.wugniu_format(text
	gn → n
	sh zh → x xh
	gh w y → hh wh yh
	b d g m n l j → bh dh gh mh nh lh jh
	ng → nhg
	v z → fh sh
	ts tsh c ch → z c j q
	'mh 'nh 'lh 'nhg → m n l ng
	
	--rimes
	q → k
	aon on en → ang ong eng
	au eu → ao ou
	(ae/e split, but no way to do this on enwikt)
	word-initial i and u → yi wu (eg. 一 iq → yik)
	iuk → yuik
	rules regarding the i in iu behaves exactly like the diaresis in Pinyin ü
	
	--syllabics: NO CHANGE: m n ng er → m n ng er
	
	--tones: 5, 6, 7, 8 UNMARKED, 1: APOSTROPHE IN FRONT
	天: 'ti, 去 qu, 定 dhin, 不 bek, 日 nik
	
	SPACES between syllables, not dashes
]]

function export.wugniu_format(text, loc)
	-- 1a a 1a 1a3 a1 -> ^1a-a-a_1-^1a_3-a_1
	-- 1a3-3a5 -> ^1a_3-^3a_5
	return text
		:gsub("", {="", ="-", =" ", =" ", ="; "})
		:gsub("(%-?)(%d?%u?)('?%l+)(%d?%u?)", function(dash, tone1, main, tone2)
			if dash == '-' and tone2 == '' then
				tone1, tone2 = tone2, tone1
			end
			if tone1 ~= '' then
				tone1 = '<sup>' .. tone1 .. '</sup>'
			end
			if tone2 ~= '' then
				tone2 = '<sub>' .. tone2 .. '</sub>'
			end
			return dash .. tone1 .. main .. tone2
		end)
end

local function wikt_format(text)
	return export.wugniu_format(text)
end

local function minidict_format(text)
	-- 1A3 3B5 3C D3 E -> A^3 B^5 C^3 D^3 E
	-- 1A B -> A^1 B
	return text
		:gsub("-", "")
		:gsub("", " ")
		:gsub(",", "; ")
		:gsub("0", "")
		:gsub("?(%l+)()", '%1<sup>%2</sup>')
		:gsub("()(%l+)", '%2<sup>%1</sup>')
		:gsub("%f(%l*)(<sup></sup>)", "'%1%2")
		:gsub("",{
			="平",="平",
			="上",="上",
			="去",="去",
			="入",="入",
		})
end

function export.wugniu_to_minidict(text, loc)
	if type(text) == "table" then text = text.args end
	text = text:gsub('%f', {y = 'yi', w = 'wu'})
	if loc == 'sx' then
		text = text:gsub("+",{een="en",en="eon",iq="ieq"})
	elseif loc == 'hz' then -- are we dealing with mergers?
		text = text:gsub("+q?%f",{eu="ei",ieu="iu",aq="eq",iaq="ieq",iq="ieq",uaq="ueq"})
	elseif loc == 'sz' or loc == 'cz' then
		text = text:gsub("%fyie%f", "yiie") -- ye > yie
	elseif loc == 'nb' then
		text = text:gsub("yu%f", "oe")
	elseif loc == 'sh' then
		text = text:gsub("ie%f", "iae")
	elseif loc == 'cm' then
		text = text:gsub("<sup>→%l+</sup>", "")
	elseif loc == 'jd' then
		text = text:gsub("ue%f", "uie")
	end
	return minidict_format(text
		--finals & syllabic
		:gsub("iu()", "iui%1")
		:gsub("gher", "r")
		:gsub("er", "r")
		:gsub("q", "h"))

		--initials
		--Glottal stops? text = text:gsub("", "'")
		:gsub("gn", "ny")
		:gsub("nyi%f", "ny")
		:gsub('yi()', 'y%1')
		:gsub('wu()', 'w%1')
end

-- various boilerplates
function export.name_boilerplate(name, wiki)
	return '<i>]</i>'
end

function export.consolas(text)
	return '<span class="zhpron-monospace">' .. text .. '</span>'
end

function export.wugniu_boilerplate(text)
	return '\n*** <small><i>]</i></small>: '
		.. export.consolas(text)
end

function export.minidict_boilerplate(text)
	return '\n*** <small><i>]</i></small>: '
		.. export.consolas(text)
end

function export.wikt_boilerplate(text)
	return '\n*** <small><i>] (Shanghai)</i></small>: '
		.. export.consolas(text)
end

function export.IPA_boilerplate(text, name, wiki)
	text = text:gsub("(/?*/*/?)", '<span style="white-space: nowrap;">%1</span>')
	return '\n*** <small>Sinological ]'
		.. ' (' .. export.name_boilerplate(name, wiki) .. ')</small>: '
		.. '<span class="IPA">/' .. text .. '/</span>'
end

local function preprocess_IPA(text, loc)
	if loc == 'hz' then
		return text:gsub("%f(?h?u)%f", "%1w")
	elseif loc == 'sx' then
		return text:gsub("+%++", function(chain)
			local tone1,mode,word1,tone2,word2 = chain:match("^(%d)(?)(%l+)%+(%d)(%l+)$")
			if not tone1 then error("sx: Wrong chain format.") end
			if mode == '' then mode = 'O' end
			if mode == 'A' then
				return tone1..'A'..word1..'&'..tone2..word2
			end
			return tone1..word1..' '..tone2..mode..word2
		end):gsub("#(%d)","%1N")
	elseif loc == 'cm' then
		return text:gsub("%f%l+<(%l*)>","%1")
	end
	return text
end

local function preprocess_wugniu(text, loc)
	if loc == 'jx' then
		return text:gsub("3","3")
	elseif loc == 'cm' then
		return text:gsub("?","")
			:gsub("%f(%l*)<(%l*)>(%l*)(%d?)","%1%3%4<sup>→%2%3</sup>")
	elseif loc == 'sx' then
		return text:gsub("","")
	end
	return text
end

local function preprocess_mutation(text, locs)
	if locs:find('cm') then
		text = text:gsub(" (%d?C?)(h?)(+)", function(tone, initial, final)
			local mutated_initial = ({v="u",zh="",gh=""})
			if mutated_initial == "u" and final:find("^u") then
				mutated_initial = ""
			elseif initial == "d" and final:find("^i") then
				mutated_initial = "l"
			end
			if final:find("<") or not mutated_initial then
				return " "..tone..initial..final
			end
			return " "..tone..initial.."<"..mutated_initial..">"..final
		end)
	end
	if text:find("<") and locs ~= "cm" then
		error("cm: Mutation is incompatible with collapsing.")
	end
	return text
end

function export.make(text, w_count)
	if not text:match(':') then -- assume Shanghainese
		text = 'sh:'..text
	end
	local show = ""
	local hide = ""
	local roms = {}
	local input_seen, duplicated = {}, false
	text = mw.text.split(text, ';', true)
	local show_name = "<i>]</i>"
	if #text == 1 and text:find("^..:") then -- single locality
		local loc = text:sub(1,2)
		show_name = export.name_boilerplate(loc_names, wiki_names)
	end
	for i = 1,#text,1 do
		local s = mw.text.split(text, ':', true)
		if not duplicated then
			if input_seen] then duplicated = true end
			input_seen] = true
		end
		if #s ~= 2 or #s == 0 then
			error("Wugniu: prefix is required or too many prefixes")
		end
		local locs, t = mw.text.split(s, ',', true), s
		rom_check(t, s)
		
		t = preprocess_mutation(t, s)
		local list = {}
		local format_text = t
		for _, loc in ipairs(locs) do
			if loc_names then
				list = true
			else
				error('Wugniu: prefix "' .. loc .. '" is not recognized')
			end
			format_text = preprocess_wugniu(format_text, loc)
		end
		local wugniu_text = export.wugniu_format(format_text, locs)
		table.insert(roms,wugniu_text)
		local names = {}
		local minidicts = {}
		local minidicts_seen = {}
		local IPAs = {}
		for _, loc in ipairs(order) do if list then
			table.insert(names, export.name_boilerplate(loc_names, wiki_names))
			if minidict then
				local minidict_result = export.wugniu_to_minidict(format_text, loc)
				if not minidicts_seen then
					table.insert(minidicts, minidict_result)
					minidicts_seen = true
				end
			end
			local ipa_text = preprocess_IPA(t, loc)
			ipa_text = export.wugniu_to_ipa(ipa_text, loc, get_initial, get_final, ipa_syllabic, get_tone)
			table.insert(IPAs,export.IPA_boilerplate(ipa_text, loc_names, wiki_names))
		end end
		hide = hide .. '\n** <small>(<i>]</i>: ' .. table.concat(names,', ') .. ')</small>'
		hide = hide .. export.wugniu_boilerplate(wugniu_text)
		for _,minidict_text in ipairs(minidicts) do
			hide = hide .. export.minidict_boilerplate(minidict_text)
		end
		if list.sh then
			hide = hide .. export.wikt_boilerplate(wugniu_to_wikt(format_text))
		end
		hide = hide .. table.concat(IPAs, '')
	end
	if not w_count or w_count > 1 then
		show = '\n** <small>('..show_name..')</small>: ' .. export.consolas(table.concat(roms, ' / '))
	else
		show = ' <small>('..show_name..', <i>]</i>)</small>: ' .. export.consolas(table.concat(roms, ' / '))
	end
	if duplicated then 
		require("Module:debug").track("wuu-pron/duplicated")
	end
	return show, hide
end

return export