Open main menu


This module contains four functions, three of which are called by other modules.

standardDiacritics takes spacing or nonstandard diacritics and converts them to standard combining diacritics. This function is used by pronunciationOrder.

reorderDiacritics takes the diacritics, removes them from the letter (mw.ustring.toNFD), and reorders them so that macrons or breves are first; diaeresis or breathing mark is second; acute, grave, or circumflex is third; and iota subscript is last. Aside from the iota subscript part, this is the only order in which the diacritics can display correctly, as explained elsewhere. This function is used by Module:typing-aids and {{chars}}.

  • ά̓̆νερ (α◌́◌̓◌̆νερ) → ᾰ̓́νερ (α◌̆◌̓◌́νερ)

pronunciationOrder does the same thing, except it puts the macron or breve and iota subscript last and recombines the diacritics (mw.ustring.toNFC) after reordering them. The diaeresis or breathing mark and accent mark will recombine, while the macron and breve remains uncombined as a combining character. This function is used by Module:grc-pronunciation and {{grc-IPA}}.

Module:grc-utilities/data holds the diacritic definitions and substitutions that are used by this module.


The function tokenize breaks the text into meaningful units of a single consonant or monophthong letter, or diphthong, with any diacritics, as shown below. This function is used by Module:grc-translit and Module:grc-accent, and by the sandbox module Module:grc-pronunciation/sandbox.

The first argument is the word to be tokenized. The second is a boolean: if true, the function will group εω together as a diphthong, for instance in πόλεως (póleōs), genitive of πόλῐς (pólis, city state).

word tokens
ἡμεῖς , μ, εῖ, ς
οἷαι οἷ, αι
ἀναῡ̈τέω , ν, α, ῡ̈, τ, έ, ω
δαΐφρων δ, α, ΐ, φ, ρ, ω, ν
τούτῳ τ, ού, τ,
ὑϊκός , ϊ, κ, ό, ς
ἡ Ἑλήνη , , , λ, ή, ν, η
νηῦς ν, ηῦ, ς
υἱός υἱ, ό, ς
ὄργυιᾰ , ρ, γ, υι,
οὐ δοκεῖν ἀλλ᾽ εἶναι ἀγαθὸν οὐ, , δ, ο, κ, εῖ, ν, , , λ, λ, , , εἶ, ν, αι, , , γ, α, θ, , ν



local export = {}

local m_script_utils = require("Module:script utilities")
local m_links = require("Module:links")
local lang = require("Module:languages").getByCode("grc")
local sc = require("Module:scripts").getByCode("polytonic")

local m_data = mw.loadData("Module:grc-utilities/data")
local groups = m_data.groups
local diacritic_order = m_data.diacritic_order
local conversions = m_data.conversions
local diacritics = m_data.diacritics
local diacritic = m_data.diacritic
local macron = diacritics.macron
local breve = diacritics.breve
local spacing_macron = diacritics.spacing_macron
local spacing_breve = diacritics.spacing_breve
local rough = diacritics.rough
local smooth = diacritics.smooth
local diaeresis = diacritics.diaeresis
local acute = diacritics.acute
local grave = diacritics.grave
local circumflex = diacritics.circum
local subscript = diacritics.subscript
local combining_diacritic = m_data.combining_diacritic

local UTF8_char = "[\1-\127\194-\244][\128-\191]*"
local basic_Greek = "[\206-\207][\128-\191]" -- excluding first line of Greek and Coptic block: ͰͱͲͳʹ͵Ͷͷͺͻͼͽ;Ϳ

local find = mw.ustring.find
local match = mw.ustring.match
local gmatch = mw.ustring.gmatch
local sub = mw.ustring.sub
local gsub = mw.ustring.gsub
local toNFC = mw.ustring.toNFC
local decompose = mw.ustring.toNFD

local info = {}
-- The tables are shared among different characters so that they can be checked
-- for equality if needed, and to use less space.
local vowel_t = { vowel = true }
local iota_t = { vowel = true, offglide = true }
local upsilon_t = { vowel = true, offglide = true }
-- These don't need any contents.
local rho_t = {}
-- local consonant_t = {}
local diacritic_t = { diacritic = true }
-- Needed for equality comparisons.
local breathing_t = { diacritic = true }

local function add_info(characters, t)
	if type(characters) == "string" then
		for character in string.gmatch(characters, UTF8_char) do
			info[character] = t
		for i, character in ipairs(characters) do
			info[character] = t

add_info({ macron, breve,
		acute, grave, circumflex,
	}, diacritic_t)

add_info({rough, smooth}, breathing_t)
add_info("ΑΕΗΟΩαεηοω", vowel_t)
add_info("Ιι", iota_t)
add_info("Υυ", upsilon_t)
-- add_info("ΒΓΔΖΘΚΛΜΝΞΠΡΣΤΦΧΨϜϘϺϷͶϠβγδζθκλμνξπρσςτφχψϝϙϻϸͷϡ", consonant_t)
add_info("Ρρ", rho_t)

local not_recognized = {}
setmetatable(info, { __index =
	function(t, key)
		return not_recognized

local sparseConcat = require("Module:table").sparseConcat

local checkType = require "libraryUtil".checkType

local function _check(funcName)
	return function(argIndex, arg, expectType, nilOk)
		return checkType(funcName, argIndex, arg, expectType, nilOk)

-- Perform a function on each Unicode character in a string.
local function forEach(str, func)
	for char in string.gmatch(str, UTF8_char) do

-- This concatenates or inserts a character, then removes it from the text.
local function add(list, index, chars, text)
	if not chars then
		error("The function add cannot act on a nil character.")
	if list[index] then
		list[index] = list[index] .. chars
		list[index] = chars
	-- Basic string function works here.
	return text:sub(#chars + 1)

function export.tag(term, face)
	return m_script_utils.tag_text(term, lang, sc, face)

function, face, alt, tr)
	return m_links.full_link( { term = term, alt = alt, lang = lang, sc = sc, tr = tr }, face)

local function linkNoTag(term, alt)
	return m_links.language_link{ term = term, lang = lang, alt = alt }

-- Convert spacing to combining diacritics, and nonstandard to standard polytonic Greek.
function export.standardDiacritics(text)
	text = decompose(text)
	text = text:gsub(UTF8_char, conversions)
	return text

--[=[	This function arranges diacritics in the following order:
			1. macron or breve
			2. breathings or diaeresis
			3. acute, circumflex, or grave
			4. iota subscript
		Used by [[Module:typing-aids]].
		Returns an error if a sequence of diacritics contains more than one
		of each category.
local function reorderDiacriticSequence(diacritics)
	local output = {}
		function (diacritic)
			local index = diacritic_order[diacritic]
			if not output[index] then
				output[index] = diacritic
				-- Place breve after macron.
				if diacritic == breve then
					index = index + 1
				-- The following might have odd results when there
				-- are three or more diacritics.
				table.insert(output, index, diacritic)
				-- [[Special:WhatLinksHere/Template:tracking/grc-utils/too many diacritics]]
				require("Module:debug").track("grc-utils/too many diacritics")
				local m_templates = require("Module:grc-utilities/templates")
				error("There are two diacritics, " ..
						m_templates.addDottedCircle(output[index]) .. " and " ..
						m_templates.addDottedCircle(diacritic) ..
						" that belong in the same position. There should be only one."
	return sparseConcat(output)

function export.reorderDiacritics(text)
	local d = diacritics
	return (gsub(decompose(text),
		combining_diacritic .. combining_diacritic .. "+",

		This breaks a word into meaningful "tokens", which are
		individual letters or diphthongs with their diacritics.
		Used by [[Module:grc-accent]] and [[Module:grc-pronunciation]].
local function make_tokens(text)
	local tokens, prev_info = {}, {}
	local token_i, vowel_count = 1, 0 -- Vowel count tracks .
	local prev
	for character in string.gmatch(decompose(text), UTF8_char) do
		local curr_info = info[character]
		-- Split vowels between tokens if not a diphthong.
		if curr_info.vowel then
			vowel_count = vowel_count + 1
			if prev and (not (vowel_count == 2 and curr_info.offglide and prev_info.vowel)
					-- υυ → υ, υ
					-- ιυ → ι, υ
					or prev_info.offglide and curr_info == upsilon_t) then
				token_i = token_i + 1
				if prev_info.vowel then
					vowel_count = 1
			elseif vowel_count == 2 then
				vowel_count = 0
			tokens[token_i] = (tokens[token_i] or "") .. character
		elseif curr_info.diacritic then
			vowel_count = 0
			tokens[token_i] = (tokens[token_i] or "") .. character
			if prev_info.diacritic or prev_info.vowel then
				if character == diaeresis then
					-- Split the diphthong in the current token if a diaeresis was found:
					-- the first letter, then the second letter plus any diacritics.
					local previous_vowel, vowel_with_diaeresis =
							"^(" .. basic_Greek .. ")(" .. basic_Greek .. ".+)")
					if previous_vowel then
						tokens[token_i], tokens[token_i + 1] = previous_vowel, vowel_with_diaeresis
						token_i = token_i + 1
						-- The vowel preceding the vowel with the diaeresis will already be
						-- placed in the previous token if it has a diacritic:
						-- Περικλῆῐ̈ → Π ε ρ ι κ λ ῆ ῐ̈
						mw.log('Diaeresis was found in ' .. text .. ', but the previous token ' ..
							require("Module:Unicode data").add_dotted_circle(tokens[token_i]) ..
							' couldn’t be split because it does not consist of two Basic Greek characters followed by other characters.')
			elseif prev_info == rho_t then
				if curr_info ~= breathing_t then
					mw.log(string.format("The character %s in %s should not have the accent %s on it.",
						prev, text, require("Module:grc-utilities/templates").addDottedCircle(character)))
				mw.log("The character " .. prev .. " cannot have a diacritic on it.")
			vowel_count = 0
			if prev then
				token_i = token_i + 1
			tokens[token_i] = (tokens[token_i] or "") .. character
		prev = character
		prev_info = curr_info
	return tokens

local cache = {}
function export.tokenize(text)
	local decomposed = decompose(text)
	if not cache[decomposed] then
		cache[decomposed] = make_tokens(text)
	return cache[decomposed]

--[=[	Places diacritics in the following order:
			1. breathings or diaeresis
			2. acute, circumflex, or grave
			3. macron or breve
			4. iota subscript
		Used by [[Module:grc-pronunciation]].		]=]
function export.pronunciationOrder(text)
	text = export.standardDiacritics(text)
	if find(text, groups[1]) then
		text = gsub(text,
			diacritic .. diacritic .. "+",
				-- Put breathing and diaeresis first, then accents, then macron or breve
				return table.concat{
					match(sequence, groups[2]) or "",
					match(sequence, groups[3]) or "",
					match(sequence, groups[1]) or "",
					match(sequence, groups[4]) or ""
		text = gsub(text, macron, spacing_macron) -- combining to spacing macron
		text = gsub(text, breve, spacing_breve) -- combining to spacing breve
	return toNFC(text)

-- Returns a table of any ambiguous vowels in the text, language-tagged.
function export.findAmbig(text, noTag)
	if (not text) or type(text) ~= "string" then
		error("The input to function findAmbig is nonexistent or not a string")
	local lengthDiacritic = "[" .. macron .. breve .. circumflex .. subscript .. "]"
	local aiu_diacritic = "^([" .. "αιυ" .. "])(" .. diacritic .. "*)$"
	-- breaks the word into units
	local output, vowels = {}, {}
	for _, token in ipairs(export.tokenize(text)) do
		if not find(token, m_data.consonant) then
			local vowel, diacritics = match(
			if vowel and (diacritics == "" or
					not find(diacritics, lengthDiacritic)) then
				local diacriticked_vowel
				if not noTag then
					diacriticked_vowel = export.tag(vowel .. diacritics)
					diacriticked_vowel = vowel
				table.insert(output, diacriticked_vowel)
				-- Lists the vowel letters that are ambiguous, for categorization purposes.
				vowels[mw.ustring.lower(vowel)] = true
	return output, vowels

return export