Module:pag-pron

-- Based on ] by Benwing2. 
-- Adaptation by TagaSanPedroAko. Some code based on ].

local export = {}

local m_IPA = require("Module:IPA")

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

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 vowel = "aeëiɨou" -- vowel
local V = ""
local W = "" -- glide
local accent = AC .. GR .. CFLEX
local accent_c = ""
local stress_c = ""
local ipa_stress = "ˈˌ"
local ipa_stress_c = ""
local separator = accent .. ipa_stress .. "# ."
local separator_c = ""
local C = "" -- consonant

local unstressed_words = require("Module:table").listToSet({ --feel free to add more unstressed words
	"say", "ëd", "si", "ni", -- case markers
	"no", -- 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 ,  and 
--

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, "." .. "", {
		 = "ë",
		 = "ñ",
		 = "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*()", "%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. ], ], ], ], 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 then
			-- add CFLEX to the last vowel not the first one
			-- adding the CFLEX after the 'u'
			words = 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)

	--determining whether "y" 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()", "s%1")
	text = rsub(text, "gu(" .. V .. ")", "ɡw%1")
	text = rsub(text, "qu()", "k%1")
	text = rsub(text, "ü", "u") 

	table.insert(debug, text)

	--alphabet-to-phoneme
	text = rsub(text, "ch", "ts") --not the real sound
	text = rsub(text, "#p()", "#%1") -- ], ]
	text = rsub(text, "",
	--="ɡ":  U+0067 LATIN SMALL LETTER G → U+0261 LATIN SMALL LETTER SCRIPT G
		{  = "k",  = "ɨ",  = "p",  = "ɡ",  = "ĵ",  = "nj",  = "k",  = "ɾ",  = "b",  = "j",  = "s" })

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

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

	text = rsub(text, "n(*)", "m%1")

	table.insert(debug, text)

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

	table.insert(debug, text)

	local accent_to_stress_mark = {  = "ˈ",  = "ˌ",  = "" }

	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 = rsub(syllables, "^(.*)(" .. accent_c .. ")(.*)$",
						function(pre, accent, post)
							return accent_to_stress_mark .. 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, "#") or #syllables == 1 and rfind(word, "") 		then
				syllables = "ˈ" .. syllables
			elseif #syllables > 1 then
				syllables = "ˈ" .. syllables
			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 = 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,"()i","%1j")
	text = rsub(text,"()u","%1w")

    table.insert(debug, text)

	--phonetic transcription
	if phonetic then

        --phonemic diphthongs
		text = rsub(text, "()j", "%1ɪ̯")
		text = rsub(text, "()w", "%1ʊ̯")

        --i or u, and /e/ and /o/
	    text = rsub(text,"()","ɪ")
	    text = rsub(text,"()","ɪ")
	    text = rsub(text,"()","ʊ")
	    text = rsub(text,"()","ʊ")

       	table.insert(debug, text)

        text = rsub(text,"d(?)j","%1d͡ʒ") --/d/ before /j/
        text = rsub(text,"n(?)k","ŋ%1k") -- /n/ before /k/ (some proper nouns)
        text = rsub(text,"n(?)ɡ","ŋ%1ɡ") -- /n/ before /ɡ/ (some proper nouns and loanwords)
        text = rsub(text,"n(?)h","ŋ%1h") -- /n/ before /h/ (some proper nouns)
        text = rsub(text,"n(?)m","m%1m") -- /n/ before /m/
        text = rsub(text,"s(?)j","%1ʃ") -- /s/ before /j/
        text = rsub(text,"t(?)j","%1t͡ʃ") -- /t/ before /j/
        text = rsub(text,"(?)d()()","%1 ĵ%3") -- /dj/ before any vowel following stress
        text = rsub(text,"(?)s()()","%1ʃ%3") -- /sj/ before any vowel following stress
        text = rsub(text,"(?)t()()","%1t͡ʃ%3") -- /tj/ before any vowel following stress
        text = rsub(text,"()()(?)()","u%2%3%4") -- /o/ and /ʊ/ before /mb/ or /mp/
	    
        table.insert(debug, text)
        
       	--Change /ɪ/ and /ʊ/ to /i~e/ and /u~o/ depending on surrounding phonemes
	    text = rsub(text,"()(?)ɪ#","%1%2e")
	    text = rsub(text,"()ɪ#","%1e")
	    text = rsub(text,"()ɪ#","%1e")
	    text = rsub(text,"()()(?)ɪ()","%1%2%3i%4")
	    text = rsub(text,"()()ɪ()","%1%2i%3")
	    text = rsub(text,"()()ɪ()","%1%2i%3")
        text = rsub(text,"()(?)ʊ#","%1%2o")
	    text = rsub(text,"()ʊ#","%1o")
	    text = rsub(text,"()ʊ#","%1o")
	    text = rsub(text,"()()(?)ʊ ()","%1%2%3u%4")
	    text = rsub(text,"()()ʊ ()","%1%2u%3")
	    text = rsub(text,"()()ʊ ()","%1%2u%3")
	end

	table.insert(debug, text)

	-- convert fake symbols to real ones
    	local final_conversions = {
		 = "t͡ʃ", -- fake "ch" to real "ch"
		 =  "j", -- fake "y" to real "y"
         = "d͡ʒ" -- fake "j" to real "j"
	}

    	local final_conversions_phonetic = {
		 = "t͡ʃ", -- fake "ch" to real "ch"
		 =  "j", -- fake "y" to real "y"
         = "d͡ʒ" -- fake "j" to real "j"
	}

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

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

	return mw.ustring.toNFC(text)
end

function export.show(frame)
	local params = {
		 = {},
		 = {},
	}

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

	local results = {}

	local text = args or mw.title.getCurrentTitle().text

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

	local pre = args.pre and args.pre .. " " or ""

	return "* " .. pre .. m_IPA.format_IPA_full { lang = lang, items = results }
end

return export