Module:mr-IPA

local export = {}

local m_IPA = require("Module:IPA")
local m_str_utils = require("Module:string utilities")

local find = m_str_utils.find
local gcodepoint = m_str_utils.gcodepoint
local gmatch = m_str_utils.gmatch
local gsub = m_str_utils.gsub
local u = m_str_utils.char

local lang = require("Module:languages").getByCode("mr")
local sc = require("Module:scripts").getByCode("Deva")

local correspondences = {
	 = "ŋ",  = "ɡ", 
     = "t͡ɕ",  = "d͡ʑ",  = "t͡s",  = "d͡z",  = "n",
	 = "ʈ",  = "ɖ",  = "ɳ",
	 = "t̪",  = "d̪",
	 = "j",  = "ɾ",  = "ʋ",  = "l̪",  = "ɭ̆",
	 = "ɕ",  = "ʂ",  = "ɦ",
	 = "ɽ",  = "z",  = "ɭ",  = "ɡ",  = "k",  = "kʰ",  = "n",  = "ɾ",

	 = "ə",  = "a",  = "i",
	 = "i",  = "o",  = "e",
	 = "u",  = "u",  = "ɔ",  = "æ",

	 = "ũ",  = "õ",  = "ə̃",  = "ã", 

	 = "oːm",  = "ʰ",   = "",
}

local deaspirate = {
     = "क",  = "ग",
     = "च",  = "ज",
     = "ट",  = "ड",
     = "त",  = "द",
     = "प",  = "ब"
}

local vowels = "aāiīuūoŏĕɔɛeæ"
local weak_h_c = "gjdḍṇbṛnmrṟlv"
local weak_h = "()h"
local aspirate = "()"
local syllabify_pattern = "(+)(+)(+)"

local function find_consonants(text)
	local current = ""
	local cons = {}
	for cc in gcodepoint(text .. " ") do
		local ch = u(cc)
		if find(current .. ch, "^̈?$") or find(current .. ch, "^̈?h$") then
			current = current .. ch
		else
			table.insert(cons, current)
			current = ch
		end
	end
	return cons
end

local function syllabify(text)
	for count = 1, 2 do
		text = gsub(text, syllabify_pattern, function(a, b, c)
			b_set = find_consonants(b)
			table.insert(b_set, #b_set > 1 and 2 or 1, ".")
			return a .. table.concat(b_set) .. c end)
	end
	return text
end

local identical = "knlsfzθ"
for character in gmatch(identical, ".") do
	correspondences = character
end

local function transliterate(text)
	return (lang:transliterate(text))
end

function export.link(term)
	return require("Module:links").full_link{ term = term, lang = lang, sc = sc }
end

function export.toIPA(text, phonetic)
	--fix how aspiration is represented
	text = gsub(text, '()(़?्)()', function(a, b, c)
		if a == c then return deaspirate .. b .. c end
		return a .. b .. c
	end)

	local translit = transliterate(text)
	if not translit then
		error('The term "' .. text .. '" could not be transliterated (is it in the correct script?).')
	end
	if phonetic then
		translit = gsub(translit, 'ts', 'ċh') -- ts realization
		translit = gsub(translit, '()aha()', '%1a%2') -- schwa dropping
		translit = gsub(translit, '()ah()', '%1h%2') -- schwa dropping
		translit = gsub(translit, '()()h()', '%1h%2%3') -- murmur + aspiration rules
		translit = gsub(translit, '()ah()', '%1%2') -- schwa + h dropping
		translit = gsub(translit, '()()ha()', '%1%2%3') -- schwa + h dropping 2
		translit = gsub(translit, '()()h()', '%1%2%3') -- h dropping
	end
	
	-- vowels
	translit = gsub(translit, "͠", "̃")
	translit = gsub(translit, 'a(̃?)i', 'əi%1')
	translit = gsub(translit, 'a(̃?)u', 'əu%1')
	translit = gsub(translit, "%-", ".")
	translit = gsub(translit, "ŕ", "ru")
	
	-- schwa force
	translit = gsub(translit, "%*", "a")
	
	translit = syllabify(translit)
	
	-- clusters
	translit = gsub(translit, 'ndny', 'ndñ')
	translit = gsub(translit, 'dny', 'dñ')
	if phonetic then translit = gsub(translit, 'ts', 't͡sʰ') end
	
	if phonetic then
		translit = gsub(translit, '()āh', '%1hā')
		translit = gsub(translit, aspirate .. "h", '%1ʰ')
		translit = gsub(translit, weak_h, '%1ʱ')
		translit = gsub(translit, '()%.h', '.%1ʱ')
		translit = gsub(translit, aspirate .. '%.h', '.%1ʰ')
	end
	
	if phonetic then
		translit = gsub(translit, '()()h%.', '%1%2.') -- more h dropping
		translit = gsub(translit, '()()h$', '%1%2') -- more h dropping
	end
	
	translit = gsub(translit, "%.ː", "ː.")

	-- aspiration
	translit = gsub(translit, "()h", "%1ʰ")
	translit = gsub(translit, "()ʰ", "%1ʱ")
	translit = gsub(translit, "()%.h", ".%1ʱ")
	
	local result = gsub(translit, ".̈?", correspondences)
	
	-- formatting
	result = gsub(result, "ː̃", "̃ː")
	result = gsub(result, "ː.̃", "̃ː.")
	result = gsub(result, "%. ", " ")
    result = gsub(result, "%.$", "")

	if phonetic then
		-- lengthening of i/u in final syllables (Dhongde & Wali p. 9)
		result = gsub(result, "i(*)$", "iː%1")
		result = gsub(result, "u(*)$", "uː%1")
	end
	
	return result
end

function export.make(frame)
	local args = frame:getParent().args
	local pagetitle = mw.title.getCurrentTitle().text
	
	local p, results = {}, {}
	
	if args then
		for index, item in ipairs(args) do
			table.insert(p, (item ~= "") and item or nil)
		end
	else
		p = { pagetitle }
	end
	
	for _, Marathi in ipairs(p) do
		local broad, narrow = export.toIPA(Marathi, false), export.toIPA(Marathi, true)
		table.insert(results, { pron = "/" .. broad .. "/" })
		if broad ~= narrow then
			table.insert(results, { pron = "" })
		end
	end
	
	return m_IPA.format_IPA_full { lang = lang, items = results }
end

return export