Modül:IPA

local cikart = {}
-- ]

local m_veri = mw.loadData('Modül:IPA/veri') -- ]
local m_semboller = mw.loadData('Modül:IPA/veri/semboller') -- ]
local m_heceleme -- ]; loaded below if needed

local sub = mw.ustring.sub
local find = mw.ustring.find
local gsub = mw.ustring.gsub
local match = mw.ustring.match
local gmatch = mw.ustring.gmatch
local U = mw.ustring.char

function cikart.bicim_IPA_tam(dil, ogeler, hata, ayirici, siraAnahtari, sayim_yok)
	local IPA_ana, ana_bag, hata_yazi, on_ek, IPAlar, kategori
	local varmiAnahtar = m_veri.bilgi_sayfali_diller
	local alanadi = mw.title.getCurrentTitle().nsText
	
	if hata then
		hata_yazi = '<span class="error">' .. hata .. '</span>'
	else
		if varmiAnahtar then
			IPA_ana = "Ek:" .. dil:getirAsilAd() .. " söylenişi"
		else
			IPA_ana = "wikipedia:" .. dil:getirAsilAd()
		end
		
		ana_bag = "]"
	end

	
	local on_ek = "]<sup>(" .. ( ana_bag or hata_yazi ) .. ")</sup>:&#32;"
	
	IPAlar = cikart.bicim_IPA_coklu(dil, ogeler, ayirici, sayim_yok)
	
	if dil and (alanadi == "" or alanadi == "YeniKurum") then
		siraAnahtari = siraAnahtari or dil:yapSiraAnahtari(mw.title.getCurrentTitle().text)
		siraAnahtari = siraAnahtari and ("|" .. siraAnahtari) or ""
		kategori = "]"
	else
		kategori = ""
	end

	return on_ek .. IPAlar .. kategori
end

local function determine_repr(soyl)
	local repr_mark = {}
	local repr, reconstructed
	
	-- remove initial asterisk before representation marks, used on some Reconstruction pages
	if find(soyl, "^%*") then
		reconstructed = true
		soyl = sub(soyl, 2)
	end
	
	local representation_types = {
		 = { right = '/', type = 'phonemic', },
		 = { right = ']', type = 'phonetic', },
		 = { right = '⟩', type = 'orthographic', },
		 = { type = 'rhyme' },
	}
	
	repr_mark.i, repr_mark.f, repr_mark.left, repr_mark.right = find(soyl, '^(.).-(.)$')
	
	local representation_type = representation_types
	
	if representation_type then
		if representation_type.right then
			if repr_mark.right == representation_type.right then
				repr = representation_type.type
			end
		else
			repr = representation_type.type
		end
	else
		repr = nil
	end
	
	return repr, reconstructed
end

local function varmiGecersizAyirici(transcription)
	if find(transcription, "%.") then
		return true
	else
		return false
	end
end

function cikart.bicim_IPA_coklu(dil, ogeler, ayirici, sayim_yok)
	local notes = {}
	local kategoriler = {}
	ayirici = ayirici or ', '
	
	-- Format
	if not ogeler then
		if mw.title.getCurrentTitle().nsText == "Şablon" then
			table.insert(ogeler, {soyl = "/aɪ piː ˈeɪ/"})
		else
			table.insert(kategoriler, "]")
		end
	end
	
	local bits = {}
	
	for _, oge in ipairs(ogeler) do
		local bit = cikart.bicim_IPA(dil, oge.soyl)
		
		if oge.belirticiler and oge.belirticiler then
			bit = require("Modül:belirtici").bicimlendir_belirtici(oge.belirticiler) .. " " .. bit
		end
		
		if oge.notlar then
			bit = bit .. mw.getCurrentFrame():extensionTag("ref", oge.notlar)
		end
		
		table.insert(bits, bit)
		
		-- The length or gemination symbol should not appear after a syllable break or stress symbol.
		if find(oge.soyl, "") then
			--
		end
		
		if dil then
			-- Eğer dilin ikili ünlüleri ]de listelendiyse hece sayısı ekle.
			if not sayim_yok and mw.title.getCurrentTitle().namespace == 0 then
				m_heceleme = m_heceleme or require('Modül:heceleme')
				if m_heceleme.varmiIkiliunlu(dil) then
					if determine_repr(oge.soyl) == "phonemic" or dil:getirKod() == "ru" then
						local syllable_count = m_heceleme.getirUnluler(oge.soyl, dil)
						if syllable_count then
							table.insert(kategoriler, "]")
						end	
					end
				end
			end

			if dil:getirKod() == "en" then
				if varmiGecersizAyirici(oge.soyl) then
					table.insert(kategoriler, "]")
				end
			end
		end
	end

	return table.concat(bits, ayirici) .. table.concat(kategoriler)
end

-- Takes an IPA pronunciation and formats it and adds cleanup kategoriler.
function cikart.bicim_IPA(dil, soyl, split_output)
	local hata = {}
	local kategoriler = {}
	
	-- Remove wikilinks, so that wikilink brackets are not misinterpreted as
	-- indicating phonemic transcription
	local str_gsub = string.gsub
	local without_links = str_gsub(soyl, '%]+|(]+)%]%]', '%1')
	without_links = str_gsub(without_links, '%]+%]%]', '%1')
	
	-- Detect whether this is a phonemic or phonetic transcription
	local repr, reconstructed = determine_repr(without_links)
	
	if reconstructed then
		soyl = sub(soyl, 2)
	end
	
	-- If valid, strip the representation marks
	if repr == "phonemic" then
		soyl = sub(soyl, 2, -2)
		without_links = sub(without_links, 2, -2)
	elseif repr == "phonetic" then
		soyl = sub(soyl, 2, -2)
		without_links = sub(without_links, 2, -2)
	elseif repr == "orthographic" then
		soyl = sub(soyl, 2, -2)
		without_links = sub(without_links, 2, -2)
	elseif repr == "rhyme" then
		soyl = sub(soyl, 2)
		without_links = sub(without_links, 2)
	else
		table.insert(kategoriler, "]")
	end
	
	if soyl == "" then
		table.insert(kategoriler, "]")
	end
	
	-- Check for obsolete and nonstandard symbols
	for i, symbol in ipairs(m_veri.standartolmayan) do
		local result
		for nonstandard in gmatch(soyl, symbol) do
			if not result then
				result = {}
			end
			table.insert(result, nonstandard)
			table.insert(kategoriler, "]")
		end
		
		if result then
			table.insert(hata, "obsolete or nonstandard characters (" .. table.concat(result) .. ")")
			break
		end
	end
	
	--[[ Check for invalid symbols after removing the following:
			1. wikilinks (handled above)
			2. paired HTML tags
			3. bolding
			4. italics
			5. HTML entity for space
			6. asterisk at beginning of transcription
			7. comma followed by spacing characters
			8. superscripts enclosed in superscript parentheses		]]
	local found_HTML
	local result = str_gsub(without_links, "<(%a+)*>(+)</%1>",
		function(tagName, content)
			found_HTML = true
			return content
		end)
	result = str_gsub(result, "'''(*)'''", "%1")
	result = str_gsub(result, "''(*)''", "%1")
	result = str_gsub(result, "&+;", "") -- This may catch things that are not valid character entities.
	result = str_gsub(result, "^%*", "")
	result = gsub(result, ",%s+", "")
	result = gsub(result, "⁽ .. "]+⁾", "")
	result = gsub(result, ' .. ']', '')
	if result ~= '' then
		local suggestions = {}
		mw.log(soyl, result)
		local namespace = mw.title.getCurrentTitle().namespace
		local kategori
		if namespace == 0 then
			-- main namespace
			kategori = "Geçersiz IPA karakterli IPA okunuşları"
		elseif namespace == 118 then
			-- reconstruction namespace
			kategori = "Geçersiz IPA karakterli IPA okunuşları/yenikurum"
		else
			kategori = "Geçersiz IPA karakterli IPA okunuşları/madde_dışı"
		end
		for character in gmatch(result, ".") do
			local suggestion = m_semboller
			if suggestion then
				table.insert(suggestions, character .. " with " .. suggestion)
			end
			table.insert(kategoriler, "]")
		end
		table.insert(hata, "geçersiz IPA karakterleri (" .. result .. ")")
		if suggestions then
			table.insert(hata, "replace " .. table.concat(suggestions, ", "))
		end
	end
	
	if found_HTML then
		table.insert(kategoriler, "]")
	end
	
	if repr == "phonemic" or repr == "rhyme" then
		if dil and m_veri.sesbirimleri then
			local gecerli_sesbirimleri = m_veri.sesbirimleri
			local rest = soyl
			local sesbirimleri = {}
			
			while mw.ustring.len(rest) > 0 do
				local longestmatch = ""
				
				if sub(rest, 1, 1) == "(" or sub(rest, 1, 1) == ")" then
					longestmatch = sub(rest, 1, 1)
				else
					for _, sesbirimi in ipairs(gecerli_sesbirimleri) do
						if mw.ustring.len(sesbirimi) > mw.ustring.len(longestmatch) and sub(rest, 1, mw.ustring.len(sesbirimi)) == sesbirimi then
							longestmatch = sesbirimi
						end
					end
				end
				
				if mw.ustring.len(longestmatch) > 0 then
					table.insert(sesbirimleri, longestmatch)
					rest = sub(rest, mw.ustring.len(longestmatch) + 1)
				else
					local sesbirimi = sub(rest, 1, 1)
					table.insert(sesbirimleri, "<span style=\"color: red\">" .. sesbirimi .. "</span>")
					rest = sub(rest, 2)
					table.insert(kategoriler, "]")
				end
			end
			
			soyl = table.concat(sesbirimleri)
		end
		
		if repr == "phonemic" then
			soyl = "/" .. soyl .. "/"
		else
			soyl = "-" .. soyl
		end
	elseif repr == "phonetic" then
		soyl = ""
	elseif repr == "orthographic" then
		soyl = "⟨" .. soyl .. "⟩"
	end
	
	if reconstructed then
		soyl = "*" .. soyl
	end
	
	if hata then
		hata = '<span class="previewonly error" style="font-size: small;>&#32;' .. table.concat(hata, ', ') .. '</span>'
	else
		hata = ""
	end
	
	if split_output then -- for use of IPA in links 
		return '<span class="IPA">' .. soyl .. '</span>', table.concat(kategoriler), hata
	else
		return '<span class="IPA">' .. soyl .. '</span>' .. hata .. table.concat(kategoriler)
	end
end

return cikart