Module:grc-translit

local export = {}

local m_data = require("Module:grc-utilities/data")
local m_str_utils = require("Module:string utilities")

local tokenize = require('Module:grc-utilities').tokenize

local ufind = m_str_utils.find
local ugsub = m_str_utils.gsub
local U = m_str_utils.char
local ulower = m_str_utils.lower
local uupper = m_str_utils.upper

-- Diacritics
local diacritics = m_data.named

-- Greek
local acute = diacritics.acute
local grave = diacritics.grave
local circumflex = diacritics.circum
local diaeresis = diacritics.diaeresis
local smooth = diacritics.smooth
local rough = diacritics.rough
local macron = diacritics.macron
local breve = diacritics.breve
local subscript = diacritics.subscript

-- Latin
local hat = diacritics.Latin_circum

local macron_diaeresis = macron .. diaeresis .. "?" .. hat
local a_subscript = '^.*' .. subscript .. '$'
local velar = 'κγχξ'

local tt = {
	-- Vowels
	 = "a",
	 = "e",
	 = "e" .. macron,
	 = "i",
	 = "o",
	 = "u",
	 = "o" .. macron,

	-- Consonants
	 = "b",
	 = "g",
	 = "d",
	 = "z",
	 = "th",
	 = "k",
	 = "l",
	 = "m",
	 = "n",
	 = "x",
	 = "p",
	 = "r",
	 = "s",
	 = "s",
	 = "t",
	 = "ph",
	 = "kh",
	 = "ps",
	
	-- Archaic letters
	 = "st",
	 = "w",
	 = "ś",
	 = "q",
	 = "š",
	 = "v",
	
	-- Incorrect characters: see ].
	-- These are tracked by ].
	 = "b",
	 = "th",
	 = "k",
	 = "r",
	 = "s",
	 = "ph",
	
	-- Diacritics
	-- unchanged: macron, diaeresis, grave, acute
	 = '',
	 = '',
	 = '',
	 = hat,
	 = 'i',
}

function export.tr(text, lang, sc)
	if text == '῾' then
		return 'h'
	end
	
	--[[
		Replace semicolon or Greek question mark with regular question mark,
		except after an ASCII alphanumeric character (to avoid converting
		semicolons in HTML entities).
	]]
	text = ugsub(text, "()", "%1?")
	
	-- Handle the middle dot. It is equivalent to semicolon or colon, but semicolon is probably more common.
	text = text:gsub("·", ";")
	
	local tokens = tokenize(text)

	--now read the tokens
	local output = {}
	for i, token in pairs(tokens) do
		-- Convert token to lowercase and substitute each character
		-- for its transliteration
		local translit = ulower(token):gsub(".*", tt)
		
		local next_token = tokens
		
		if token == 'γ' and next_token and velar:find(next_token, 1, true) then
			-- γ before a velar should be <n>
			translit = 'n'
		elseif token == 'ρ' and tokens == 'ρ' then
			-- ρ after ρ should be <rh>
			translit = 'rh'
		elseif ufind(token, a_subscript) then
			-- add macron to ᾳ
			translit = ugsub(translit, '()', '%1' .. macron)
		end
		
		if token:find(rough) then
			if ufind(token, '^') then
				translit = translit .. 'h'
			else -- vowel
				translit = 'h' .. translit
			end
		end
		
		-- Remove macron from a vowel that has a circumflex.
		if ufind(translit, macron_diaeresis) then
			translit = translit:gsub(macron, '')
		end
		
		-- Capitalize first character of transliteration.
		if token ~= ulower(token) then
			translit = translit:gsub("^" .. ".*", uupper)
		end
		
		table.insert(output, translit)
	end
	output = table.concat(output)
	
	return output
end

return export