Module:ceb-pron

-- Based on [[Module:es-pronunc]] by Benwing2. 
-- Adaptation by TagaSanPedroAko based on [[Module:tl-pron]].

local export = {}

local pron_utilities_module = "Module:pron utilities"

local lang = require("Module:languages").getByCode("ceb")

local u = require("Module:string/char")
local rfind = mw.ustring.find
local rsubn = mw.ustring.gsub
local rsplit = mw.text.split
local ulower = mw.ustring.lower

local AC = u(0x0301) -- acute =  ́
local GR = u(0x0300) -- grave =  ̀
local CFLEX = u(0x0302) -- circumflex =  ̂
local TILDE = u(0x0303) -- tilde =  ̃
local DIA = u(0x0308) -- diaeresis =  ̈
local MACRON = u(0x0304) -- macron 

local vowel = "aeiou" -- vowel
local V = "[" .. vowel .. "]"
local accent = AC .. GR .. CFLEX .. MACRON
local accent_c = "[" .. accent .. "]"
local stress_c = "[" .. AC .. GR .. "]"
local ipa_stress = "ˈˌ"
local ipa_stress_c = "[" .. ipa_stress .. "]"
local separator = accent .. ipa_stress .. "# ."
local separator_c = "[" .. separator .. "]"
local C = "[^" .. vowel .. separator .. "]" -- consonant

local unstressed_words = require("Module:table").listToSet({
	"ang", "sa", "og", "si", "ni", "kang", -- case markers. "Nang" here is for written "ng", but can also work with nang as in the contraction na'ng and the conjunction "nang"
	"ko", "mo", "ka", --single-syllable personal pronouns
	"nga",-- linker, also temporal particle
    "gyud", "jud", "pa", -- particles
	"di7", -- negation words
	"kon", -- subordinating conjunctions
	"ug", "o", -- coordinating conjunctions
	"hay", -- interjections
	"de", "del", -- in some Spanish-derived terms and names
	"-an", "-han", "-hon", "i-",  "-i", "in-", "-in-", "mag-", "mang-", "mo-", "-on", "pag-", "pang-" -- monosyllabic affixes
})

-- version of rsubn() that discards all but the first return value
local function rsub(term, foo, bar)
	local retval = rsubn(term, foo, bar)
	return retval
end

-- version of rsubn() that returns a 2nd argument boolean indicating whether
-- a substitution was made.
local function rsubb(term, foo, bar)
	local retval, nsubs = rsubn(term, foo, bar)
	return retval, nsubs > 0
end

-- apply rsub() repeatedly until no change
local function rsub_repeatedly(term, foo, bar)
	while true do
		local new_term = rsub(term, foo, bar)
		if new_term == term then
			return term
		end
		term = new_term
	end
end

-- ĵ and ɟ are used internally to represent [d͡ʒ] and [j]
--

function export.IPA(text)
	local debug = {}

	text = ulower(text)
	-- decompose everything but ñ and ü
	text = mw.ustring.toNFD(text)
	text = rsub(text, "." .. "[" .. TILDE .. DIA .."]", {
		["n" .. TILDE] = "ñ",
		["u" .. TILDE] = "ü",
	})
	-- convert commas and en/en dashes to IPA foot boundaries
	text = rsub(text, "%s*[,–—]%s*", " | ")
	-- question mark or exclamation point in the middle of a sentence -> IPA foot boundary
	text = rsub(text, "([^%s])%s*[!?]%s*([^%s])", "%1 | %2")

	-- canonicalize multiple spaces and remove leading and trailing spaces
	local function canon_spaces(text)
		text = rsub(text, "%s+", " ")
		text = rsub(text, "^ ", "")
		text = rsub(text, " $", "")
		return text
	end

	text = canon_spaces(text)

	-- Make prefixes unstressed unless they have an explicit stress marker; also make certain
	-- monosyllabic words (e.g. [[ang]], [[sa]], [[si]], [[nga]], etc.) without stress marks be
	-- unstressed.
	local words = rsplit(text, " ")
	for i, word in ipairs(words) do
		if rfind(word, "%-$") and not rfind(word, accent_c) or unstressed_words[word] then
			-- add macron to the last vowel not the first one
			-- adding the macron after the 'u'
			words[i] = rsub(word, "^(.*" .. V .. ")", "%1" .. MACRON)
		end
	end
	text = table.concat(words, " ")
	-- Convert hyphens to spaces
	text = rsub(text, "%-", " ")
	-- canonicalize multiple spaces again, which may have been introduced by hyphens
	text = canon_spaces(text)
	-- now eliminate punctuation
	text = rsub(text, "[!?']", "")
	-- put # at word beginning and end and double ## at text/foot boundary beginning/end
	text = rsub(text, " | ", "# | #")
	text = "##" .. rsub(text, " ", "# #") .. "##"

	table.insert(debug, text)

	-- Add glottal stop for words starting with vowel
	text = rsub(text, "([#])([aeiou])", "%1ʔ%2")

	-- handle certain combinations; ch ng and sh handling needs to go first
	text = rsub(text, "([t]?)ch", "ts") 
	text = rsub(text, "([n]?)g̃", "ŋ") -- Spanish spelling support
	text = rsub(text, "ng", "ŋ")
	text = rsub(text, "sh", "sɟ")

	--x
	text = rsub(text, "([#])x(" .. V .. ")", "%1s%2")
	text = rsub(text, "x", "ks")
	
	--ll
	text = rsub(text, "ll([i]?)(".. V.. ")", "ly%2")

	--c, gü/gu+e or i, q
	text = rsub(text, "c([iey])", "s%1")
	text = rsub(text, "(" .. V .. ")gü([ie])", "%1ɡw%2")
	text = rsub(text, "gü([ie])", "ɡuw%1")
	text = rsub(text, "gui([aeëo])", "ɡy%1")
	text = rsub(text, "gu([ie])", "ɡ%1")
	text = rsub(text, "qu([ie])", "k%1")
	text = rsub(text, "ü", "u")  

	table.insert(debug, text)

	--alphabet-to-phoneme
	text = rsub(text, "[cfgjñqrvz7]",
	--["g"]="ɡ":  U+0067 LATIN SMALL LETTER G → U+0261 LATIN SMALL LETTER SCRIPT G
		{ ["c"] = "k", ["f"] = "p", ["g"] = "ɡ", ["j"] = "ĵ", ["ñ"] = "ɲ", ["q"] = "k", ["r"] = "ɾ", ["v"] = "b", ["z"] = "s", ["7"] = "ʔ"})

	-- trill in rr
	text = rsub(text, "ɾɾ", "r")

	table.insert(debug, text)

	text = rsub_repeatedly(text, "([^" .. vowel ..  "])([u])([" .. AC .. MACRON .. "]?)([aeio])("  .. accent_c .. "?)","%1%2%3.w%4%5")
	text = rsub_repeatedly(text, "(" .. V ..  ")([u])([" .. AC .. MACRON .. "]?)([aeio])("  .. accent_c .. "?)","%1.w%3%4")
	text = rsub(text, "([i])([" .. AC .. MACRON .. "])([aeou])("  .. accent_c .. "?)","%1%2.y%3%4")
	text = rsub(text, "([i])([aeou])(" .. accent_c .. "?)","y%2%3")
	text = rsub(text, "a([".. AC .."]*)o([#.])","a%1w%2")

	--determining whether "y" is a consonant or a vowel
	text = rsub(text, "y(" .. V .. ")", "ɟ%1") -- not the real sound
	text = rsub(text,"y([ˈˌ.]*)([bćĉdfɡhjĵklmnɲŋpɾrsʃtvwɟzʔ#" .. vowel .. "])","i%1%2")
	text = rsub(text, "w(" .. V .. ")","w%1")
	text = rsub(text,"w([ˈˌ]?)([bćĉdfɡjĵklmnɲŋpɾrsʃtvwɟzʔ#])","u%1%2")

	table.insert(debug, text)
	
	--vowels with grave/circumflex to vowel+glottal stop
	text = rsub(text, CFLEX, AC .. GR)
	text = rsub(text, "(" .. V .. ")([" .. AC .. "]?)" .. GR .. "([#" .. vowel .. "])", "%1%2ʔ%3")
	text = rsub(text, "(" .. V .. ")([" .. AC .. "]?)" .. GR, "%1%2")
	
	-- Add glottal stop for words starting with vowel
	text = rsub(text, "([#])(" .. V .. ")", "%1ʔ%2")
	text = rsub(text, "◌", "")

	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C .. V .. ")", "%1.%2")
	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*" .. C .. ")(" .. C .. V .. ")", "%1.%2")
	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*" .. C .. "+)(" .. C .. C .. V .. ")", "%1.%2")
	text = rsub_repeatedly(text, "(" .. C .. ")%.s(" .. C .. ")", "%1s.%2")
	-- Any aeo, or stressed iu, should be syllabically divided from a following aeo or stressed iu.
	text = rsub_repeatedly(text, "([aeo]" .. accent_c .. "*)([aeo])", "%1.%2")
	text = rsub_repeatedly(text, "([aeo]" .. accent_c .. "*)(" .. V .. AC .. ")", "%1.%2")
	text = rsub(text, "([iuə]" .. AC .. ")([aeo])", "%1.%2")
	text = rsub_repeatedly(text, "([iuə]" .. AC .. ")(" .. V .. AC .. ")", "%1.%2")
	text = rsub_repeatedly(text, "i(" .. accent_c .. "*)i", "i%1.i")
	text = rsub_repeatedly(text, "u(" .. accent_c .. "*)u", "u%1.u")

	table.insert(debug, text)

	local accent_to_stress_mark = { [AC] = "ˈ", [MACRON] = "" }

	local function accent_word(word, syllables)
		-- Now stress the word. If any accent exists in the word (including macron indicating an unaccented word),
		-- put the stress mark(s) at the beginning of the indicated syllable(s). Otherwise, apply the default
		-- stress rule.
		if rfind(word, accent_c) then
			for i = 1, #syllables do
				syllables[i] = rsub(syllables[i], "^(.*)(" .. accent_c .. ")(.*)$",
					function(pre, accent, post)
						return accent_to_stress_mark[accent] .. pre .. post
					end
				)
			end
		else
			-- Default stress rule. Words without vowels (e.g. IPA foot boundaries) don't get stress.
			if #syllables > 1 and rfind(word, "[^aeiouəʔbcĉdfɡghjɟĵklmnñŋpqrɾstvwxz#]#") or #syllables == 1 and rfind(word, V) 		then
				syllables[#syllables] = "ˈ" .. syllables[#syllables]
			elseif #syllables >= 2 then
				syllables[#syllables - 1] = "ˈ" .. syllables[#syllables - 1]
			end
		end
	end

	local words = rsplit(text, " ")
	for j, word in ipairs(words) do
		-- accentuation
		local syllables = rsplit(word, "%.")
		accent_word(word, syllables)
		-- Reconstruct the word.
		words[j] = table.concat(syllables, ".")
	end

	text = table.concat(words, " ")

	-- suppress syllable mark before IPA stress indicator
	text = rsub(text, "%.(" .. ipa_stress_c .. ")", "%1")
	--make all primary stresses but the last one be secondary
	text = rsub_repeatedly(text, "ˈ(.+)ˈ", "ˌ%1ˈ")

    table.insert(debug,text)

    --correct final glottal stop placement
    text = rsub(text,"([ˈˌ])ʔ([#]*)([ʔbĉćdfɡhĵɟklmnŋɲpɾrsʃtvwz])(" .. V .. ")","%1%2%3%4ʔ")

    table.insert(debug,text)

    --add temporary macron for /a/, /i/ and /u/ in stressed syllables so they don't get replaced by unstressed form

	text = rsub(text,"([ˈˌ])([#]*)([ʔbćĉdfɡhĵɟklmnŋpɾrstvwz]?)([ɟlnɾst]?)([a])([ʔbdfɡiklmnŋpɾstu]?)([bdɡklmnpɾst]?)","%1%2%3%4ā%6%7")
	text = rsub(text,"([ˈˌ])([#]*)([ʔbćĉdfɡhĵɟklmnŋpɾrstvwz]?)([ɟlnɾst]?)([e])([ʔbdfɡklmnŋpɾstu]?)([bdɡklmnpɾst]?)","%1%2%3%4ē%6%7")
	text = rsub(text,"([ˈˌ])([#]*)([ʔbćĉdfɡhĵɟklmnŋpɾrstvwz]?)([ɟlnɾst]?)([i])([ʔbdfɡklmnŋpɾstu]?)([bdɡklmnpɾst]?)","%1%2%3%4ī%6%7")
	text = rsub(text,"([ˈˌ])([#]*)([ʔbćĉdfɡhĵɟklmnŋpɾrstvwz]?)([ɟlnɾst]?)([o])([ʔbdfɡiklmnŋpɾst]?)([bdɡklmnpɾst]?)","%1%2%3%4ō%6%7")
	text = rsub(text,"([ˈˌ])([#]*)([ʔbćĉdfɡhĵɟklmnŋpɾrstvwz]?)([ɟlnɾst]?)([u])([ʔbdfɡiklmnŋpɾst]?)([bdɡklmnpɾst]?)","%1%2%3%4ū%6%7")

    table.insert(debug, text)

      --Corrections for diphthongs
    text = rsub(text,"([aāeəoō])i","%1j") --ay
    text = rsub(text,"([aāeəiīo])u","%1w") --aw

    table.insert(debug, text)
    
    --remove "ɟ" and "w" inserted on vowel pair starting with "i" and "u"
    text = rsub(text,"([i])([ˈˌ]?)ɟ([aāeəouū])","%1%2%3")
    text = rsub(text,"([u])([ˈˌ]?)w([aāeəiī])","%1%2%3")

    table.insert(debug, text)

    local ceb_IPA_table = {
    	["phonetic"] = text,
    	["phonemic"] = text
    }

	for key, value in pairs(ceb_IPA_table) do
		text = ceb_IPA_table[key]

		--phonetic transcription
		if key == "phonetic" then
	       	table.insert(debug, text)
	
	        --Turn phonemic diphthongs to phonetic diphthongs
			text = rsub(text, "([aāeəouū])j", "%1ɪ̯")
			text = rsub(text, "([aāeəiīo])w", "%1ʊ̯")

        table.insert(debug, text)

        --change a, e, i, o, u to unstressed equivalents (certain forms to restore)
	    text = rsub(text,"a","ɐ")
	    text = rsub(text,"e","e")
	    text = rsub(text,"i","ɪ")
	    text = rsub(text,"o","o")
	    text = rsub(text,"u","ʊ")

        table.insert(debug,text)
        
        --Combine consonants (except H) followed by I/U and certain stressed vowels
	    text = rsub(text,"([bdɡĵklmnŋpɾrst])([ɟlnɾst]?)ɪ([ˈˌ.])ɟ?([āʌēɪōoūʊ])","%3%1%2ɟ%4")
	    text = rsub(text,"([bdɡĵklmnŋpɾrst])([ɟlnɾst]?)ʊ([ˈˌ.])w?([āʌēīɪōʊ])","%3%1%2w%4")

       	table.insert(debug, text)

	    text = rsub(text, "([nŋ])([# .]*[bpm])", "m%2")
        text = rsub(text,"n([ˈˌ.])k","ŋ%1k") -- /n/ before /k/ (some proper nouns)
        text = rsub(text,"n([ˈˌ.])ɡ","ŋ%1ɡ") -- /n/ before /ɡ/ (some proper nouns and loanwords)
        -- text = rsub(text,"n([ˈˌ.])h","ŋ%1h") -- /n/ before /h/ (some proper nouns)
        text = rsub(text,"n([ˈˌ.])m","m%1m") -- /n/ before /m/
        text = rsub(text,"t([ˈˌ.])s","%1ć") -- /t/ before /s/

		 -- final fix for "iy" and "uw" combination
	    text = rsub(text,"([ɪī])([.])ɟ([ʌeoʊ])","%1%2%3")
	    text = rsub(text,"([ʊū])([.])w([ʌeɪ])","%1%2%3")

        --final fix for phonetic diphthongs

	    text = rsub(text,"([aɐ])i̯", "ai̯") --ay
	    text = rsub(text,"([aɐ])u̯", "au̯") --aw
	    text = rsub(text,"([eiɪ])i̯", "i") -- ey
	    text = rsub(text,"([eiɪ])u̯", "iu̯") --iw
	    text = rsub(text,"([oɔ])i̯", "oi̯") --oy or uy

        table.insert(debug,text)
        
        --Change /ʌ/, /ɪ/ and /ʊ/ back to /a/, /i/ and /u/ in penultimate
	    text = rsub(text,"([ˈˌ])([ʔbdɡhklmnŋpɾstwj])([lɾstj]?)ʌ","%1%2%3a")
	    text = rsub(text,"([ˈˌ])([ʔbdɡhklmnŋpɾstwj])([lɾstj]?)ɪ","%1%2%3i")
	    text = rsub(text,"([ˈˌ])([ʔbdɡhklmnŋpɾstwj])([lɾstj]?)ʊ","%1%2%3u")

        table.insert(debug,text)
        
        --Coalesce [wʊ] or [wɔ] to [w]
	    text = rsub(text,"([.])([w])([ɔʊ])","%2")

	    table.insert(debug,text)

 --turn phonemic consonants to Cebuano dental consonants

	    text = rsub(text,"d","d̪")
	    text = rsub(text,"l","l̪")
	    text = rsub(text,"n","n̪")
	    text = rsub(text,"ɾ","ɾ̪")
	    text = rsub(text,"s","s̪")
	    text = rsub(text,"t","t̪")

		table.insert(debug, text)

        --delete temporary macron in /a/, /e/, /i/, /o/ and /u/

	    text = rsub(text,"ā","a")
	    text = rsub(text,"ē","i")
	    text = rsub(text,"ī","i")
	    text = rsub(text,"ō","o")
	    text = rsub(text,"ū","u")

       	table.insert(debug, text)

		else
			text =  rsub(text,"%.","")
		end	

	    table.insert(debug, text)

    --delete temporary macron in vowels

	    text = rsub(text,"ā","a")
	    text = rsub(text,"ē","e")
	    text = rsub(text,"ī","i")
	    text = rsub(text,"ō","o")
	    text = rsub(text,"ū","u")

	-- convert fake symbols to real ones
    local final_conversions = {
		["ć"] = "t͡s", -- fake "t.s" to real "t.s"
		["ɟ"] =  "j", -- fake "y" to real "y"
        ["ĵ"] = "d͡ʒ" -- fake "j" to real "j"
	}

		text = rsub(text, "[ćɟĵ]", final_conversions)

	-- remove # symbols at word and text boundaries
	text = rsub(text, "#([.]?)", "")
	
	-- resuppress syllable mark before IPA stress indicator
	text = rsub(text, "%.(" .. ipa_stress_c .. ")", "%1")

		-- Do not have multiple syllable break consecutively
		text = rsub_repeatedly(text, "([.]+)", ".")
		text = rsub_repeatedly(text, "([.]?)(" .. ipa_stress_c .. ")([.]?)", "%2")
    
    	ceb_IPA_table[key] = mw.ustring.toNFC(text)
	end

	return ceb_IPA_table
end

local function respelling_to_IPA(data)
	local IPA_result = export.IPA(data.respelling)
	return ("/%s/ [%s]"):format(IPA_result["phonemic"], IPA_result["phonetic"])
end

function export.show(frame)
	local parent_args = frame:getParent().args
	return require(pron_utilities_module).format_prons {
		lang = lang,
		respelling_to_IPA = respelling_to_IPA,
		raw_args = parent_args,
		track_module = "ceb-pron",
	}
end

return export