Documentation for this module may be created at မဝ်ဂျူ:ilo-pron/doc

-- Based on [[Module:es-pronunc]] by Benwing2. 
-- Adaptation by TagaSanPedroAko. Some code based on [[Module:ast-IPA]].

local export = {}

local m_IPA = require("Module:IPA")

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

local u = mw.ustring.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 vowel = "aeëiouɯ" -- vowel
local V = "[" .. vowel .. "]"
local W = "[jw]" -- glide
local accent = AC .. GR .. CFLEX
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({ --feel free to add more unstressed words
	"ti", "si", "ni", -- case markers
	"ket", -- conjunctions
	"Dios", -- monosyllabic nouns
	"de", "del" --particles in Spanish-derived surnames
})

-- 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, ɟ and ć are used internally to represent [d͡ʒ], [j] and [t͡ʃ]

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

	text = ulower(text or mw.title.getCurrentTitle().text)
	-- decompose everything but ë, ñ and ü
	text = mw.ustring.toNFD(text)
	text = rsub(text, "." .. "[" .. TILDE .. DIA .. "]", {
		["e" .. DIA] = "ë",
		["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. [[ti]], [[ket]], [[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 CFLEX to the last vowel not the first one
			-- adding the CFLEX after the 'u'
			words[i] = rsub(word, "^(.*" .. V .. ")", "%1" .. CFLEX)
		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, "([#])([aëeiou])", "%1ʔ%2")

	--determining whether "y" or "w" is a consonant or a vowel
	text = rsub(text, "y(" .. V .. ")", "ɟ%1") -- not the real sound
	text = rsub(text, "y#", "i")
	text = rsub(text, "w(" .. V .. ")","w%1")
	text = rsub(text, "w#","u")

	-- handle certain combinations; ch ng and sh handling needs to go first
	text = rsub(text, "ch", "ts") --not the real sound
	text = rsub(text, "ng", "ŋ")
	text = rsub(text, "sh", "ʃ")

	--x
	text = rsub(text, "x", "ks")

	--c, g, q
	text = rsub(text, "c([ie])", "s%1")
	text = rsub(text, "gu(" .. V .. ")", "ɡw%1")
	text = rsub(text, "qu([ieë])", "k%1")
	text = rsub(text, "ü", "u") 

	table.insert(debug, text)

	--alphabet-to-phoneme
	text = rsub(text, "[cgjñrvyz7]",
	--["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", ["y"] = "ɟ", ["z"] = "s", ["7"] = "ʔ" })

	-- trill in rr (as in [[serrek]])
	text = rsub(text, "ɾɾ", "r")

    -- ë
	text = rsub(text, "ë", "ɯ")

    -- ts
	text = rsub(text, "ts", "ĉ") --not the real sound

	text = rsub(text, "n([# .]*[bpm])", "m%1")

	table.insert(debug, text)

	--syllable division
	local vowel_to_glide = { ["i"] = "j", ["o"] = "w", ["u"] = "w" }
	-- i, o and u between vowels -> j and w. Usually in proper nouns (e.g. [[Agcaoili]])
	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)([iou])(" .. V .. ")",
			function(v1, iou, v2)
				return v1 .. vowel_to_glide[iou] .. v2
			end
	)
	text = rsub_repeatedly(text, "(" .. V .. accent_c .. "*)(" .. C .. W .. "?" .. 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 .. stress_c .. ")", "%1.%2")
	text = rsub(text, "([iu]" .. stress_c .. ")([aeo])", "%1.%2")
	text = rsub_repeatedly(text, "([iu]" .. stress_c .. ")(" .. V .. stress_c .. ")", "%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)

	--diphthongs
	text = rsub(text, "i([aeouɯ])", "j%1")
	text = rsub(text, "u([aeioɯ])", "w%1")

	table.insert(debug, text)

	local accent_to_stress_mark = { [AC] = "ˈ", [GR] = "ˌ", [CFLEX] = "" }

	local function accent_word(word, syllables)
		-- Now stress the word. If any accent exists in the word (including ^ 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, "[^aeëioubcdfghjklmnŋñpqrstvwxzɡɯĉɟĵ#]#") or #syllables == 1 and rfind(word, "[aeëiouɯ]") 		then
				syllables[#syllables] = "ˈ" .. syllables[#syllables]
			elseif #syllables > 1 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, phonetic and "." or "")
	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)

      --phonemic diphthongs
    text = rsub(text,"([aeouɯ])i","%1j")
	text = rsub(text,"([aeio])u","%1w")

    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])([#]?)([ʔbĉćdfɡhjĵɟklmnŋɲpɾrsʃtw]?)([jlnɾt]?)a([bdfɡiklmnŋpɾstu]?)([bdɡklmnpɾst]?)","%1%2%3%4ā%5%6")

    table.insert(debug, text)

	--phonetic transcription
	if phonetic then

       	table.insert(debug, text)

        --phonemic diphthongs
		text = rsub(text, "([āaeouɯ])([ɟj])", "%1ɪ̯")
		text = rsub(text, "([āaeio])w", "%1ʊ̯")

        --/a/ to [ɐ]
	    text = rsub(text,"a","ɐ")

       	table.insert(debug, text)

        text = rsub(text,"d([ˈˌ]?)j","%1d͡ʒ") --/d/ before /j/
        text = rsub(text,"l([ˈˌ]?)j","%1ʎ") -- /l/ before /j/
        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,"n([ˈˌ.])j","%1ɲ") -- /n/ before /j/
        text = rsub(text,"s([ˈˌ.])j","%1ʃ") -- /s/ before /j/
        text = rsub(text,"t([ˈˌ.])j","%1ć") -- /t/ before /j/
        text = rsub(text,"([ˈˌ.#])d([j])([aɐeiouɯ])","%1ĵ%3") -- /dj/ at syllable onset
        text = rsub(text,"([ˈˌ.#])l([j])([aɐeiouɯ])","%1ʎ%3") -- /lj/ at syllable onset
        text = rsub(text,"([ˈˌ.#])n([j])([aɐeiouɯ])","%1ɲ%3") -- /nj/ at syllable onset
        text = rsub(text,"([ˈˌ.#])s([j])([aɐeiouɯ])","%1ʃ%3") -- /sj/ at syllable onset
        text = rsub(text,"([ˈˌ.#])t([j])([aɐeiouɯ])","%1ć%3") -- /tj/ at syllable onset

        --final fix for phonetic diphthongs

	    text = rsub(text,"([ɐ])ɪ̯","aɪ̯") --ay
	    text = rsub(text,"([ɐ])ʊ̯","aʊ̯") --aw

       	table.insert(debug, text)
        
       	--Change [ɐ] back to /a/ at end of word

	    text = rsub(text,"([#.])([bdɡhklmnŋprɾsʃtwɟ])([lɾt]?)ɐ([bdɡhklmnŋpɾst]?)([#])","%1%2%3a%4%5")

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

	    text = rsub(text,"ā","a")

	end

	table.insert(debug, text)

    --delete temporary macron in /a/

	    text = rsub(text,"ā","a")

	-- convert fake symbols to real ones
    	local final_conversions = {
		["ĉ"] = "t͡ʃ", -- fake "ch" to real "ch"
		["ɟ"] =  "j", -- fake "y" to real "y"
        ["ĵ"] = "d͡ʒ" -- fake "j" to real "j"
	}

    	local final_conversions_phonetic = {
		["ĉ"] = "t͡ʃ", -- fake "ch" to real "ch"
		["ɟ"] =  "j", -- fake "y" to real "y"
        ["ĵ"] = "d͡ʒ" -- fake "j" to real "j"
	}

	if phonetic then
	text = rsub(text, "[ĉɟĵ]", final_conversions_phonetic)
    	end
	text = rsub(text, "[ĉɟĵ]", final_conversions)

	-- remove # symbols at word and text boundaries
	text = rsub(text, "#", "")

	return mw.ustring.toNFC(text)

end

function export.show(frame)
	local params = {
		[1] = {},
		["pre"] = {},
		["post"] = {},
		["bullets"] = { type = "number", default = 1 },
	}

	local parargs = frame:getParent().args
	local args = require("Module:parameters").process(parargs, params)

	local results = {}

	local text = args[1] or mw.title.getCurrentTitle().text

	table.insert(results, { pron = "/" .. export.IPA(text, false) .. "/" })
	table.insert(results, { pron = "[" .. export.IPA(text, true) .. "]" })

	local bullet = string.rep("* ", args.bullets) .. " "
	local pre = args.pre and args.pre .. " " or ""
    local post = args.post and " " or ""

	return bullet .. pre .. m_IPA.format_IPA_full(lang, results) .. post
end

return export