Difference between revisions of "Module:Ancient Greek"

From blackwiki
Jump to navigation Jump to search
blackwiki>Erutuon
m (unnecessary now)
blackwiki>Erutuon
(faster tokenize function from wikt:Module:grc-utilities)
Line 1: Line 1:
 
local p = {}
 
local p = {}
 
local m_table = require "Module:TableTools"
 
local checkType = require "libraryUtil".checkType
 
  
 
local ustring = mw.ustring
 
local ustring = mw.ustring
 
local U = ustring.char
 
local U = ustring.char
local len = ustring.len
 
local sub = ustring.sub
 
 
local find = ustring.find
 
local find = ustring.find
 
local gsub = ustring.gsub
 
local gsub = ustring.gsub
Line 13: Line 8:
 
local lower = ustring.lower
 
local lower = ustring.lower
 
local upper = ustring.upper
 
local upper = ustring.upper
 +
 +
local str_gmatch = string.gmatch
  
 
local macron = U(0x304)
 
local macron = U(0x304)
Line 24: Line 21:
 
local Latin_circumflex = U(0x302)
 
local Latin_circumflex = U(0x302)
 
local subscript = U(0x345)
 
local subscript = U(0x345)
local i_diphthong = "[ΑΕΗΟΥΩαεηουω][Ιι]"
 
local u_diphthong = "[ΑΕΗΟΩαεηοω][Υυ]"
 
 
local macron_circumflex = macron .. diaeresis .. '?' .. Latin_circumflex
 
local macron_circumflex = macron .. diaeresis .. '?' .. Latin_circumflex
  
 
local is_velar = { ['κ'] = true, ['γ'] = true, ['χ'] = true, ['ξ'] = true, }
 
local is_velar = { ['κ'] = true, ['γ'] = true, ['χ'] = true, ['ξ'] = true, }
  
local is_diacritic = {
+
local UTF8_char = "[%z\1-\127\194-\244][\128-\191]*"
[macron] = true, [breve] = true,
+
local basic_Greek = "[\206-\207][\128-\191]" -- excluding first line of Greek and Coptic block: ͰͱͲͳʹ͵Ͷͷͺͻͼͽ;Ϳ
[rough] = true, [smooth] = true, [diaeresis] = true,
+
 
[acute] = true, [grave] = true, [circumflex] = true,
+
local info = {}
[subscript] = true,
+
 
}
+
-- The tables are shared among different characters so that they can be checked
 +
-- for equality if needed, and to use less space.
 +
local vowel = { vowel = true, diacritic_seat = true }
 +
local iota = { vowel = true, diacritic_seat = true, offglide = true }
 +
local upsilon = { vowel = true, diacritic_seat = true, offglide = true }
 +
-- Technically rho is only a seat for rough or smooth breathing.
 +
local rho = { consonant = true, diacritic_seat = true }
 +
local consonant = { consonant = true }
 +
local diacritic = { diacritic = true }
 +
-- Needed for equality comparisons.
 +
local breathing = { 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
 +
end
 +
else
 +
for _, character in ipairs(characters) do
 +
info[character] = t
 +
end
 +
end
 +
end
 +
 
 +
add_info({ macron, breve,
 +
diaeresis,
 +
acute, grave, circumflex,
 +
subscript,
 +
}, diacritic)
 +
 
 +
add_info({rough, smooth}, breathing)
 +
add_info("ΑΕΗΟΩαεηοω", vowel)
 +
add_info("Ιι", iota)
 +
add_info("Υυ", upsilon)
 +
add_info("ΒΓΔΖΘΚΛΜΝΞΠΡΣΤΦΧΨϜϘϺϷͶϠβγδζθκλμνξπρσςτφχψϝϙϻϸͷϡ", consonant)
 +
add_info("Ρρ", rho)
  
local function _check(funcName)
+
local not_recognized = {}
return function(argIndex, arg, expectType, nilOk)
+
setmetatable(info, { __index =
return checkType(funcName, argIndex, arg, expectType, nilOk)
+
function()
 +
return not_recognized
 
end
 
end
 +
})
 +
 +
local function quote(str)
 +
return "“" ..  str .. "”"
 
end
 
end
  
Line 88: Line 123:
 
[subscript] = 'i',
 
[subscript] = 'i',
 
}
 
}
 
local function is_diphthong(chars, next_chars)
 
return (find(chars, '^' .. i_diphthong .. '$')
 
or find(chars, '^' .. u_diphthong .. '$') )
 
and not find(next_chars, "^[" .. macron .. breve .. "]?" .. diaeresis)
 
end
 
 
-- Concatenates a character onto an existing table value, or inserts it,
 
-- 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.")
 
end
 
if list[index] then
 
list[index] = list[index] .. chars
 
else
 
list[index] = chars
 
end
 
return text:sub(#chars + 1)
 
end
 
  
 
--[=[
 
--[=[
Line 115: Line 130:
 
--]=]
 
--]=]
 
local function tokenize(text)
 
local function tokenize(text)
local tokens = {}
+
local tokens, vowel_info, prev_info = {}, {}, {}
local i = 0
+
local token_i = 1
while #text > 0 do
+
local prev
local char = sub(text, 1, 1) or ""
+
for character in str_gmatch(decompose(text), UTF8_char) do
local chars = sub(text, 1, 2) or ""
+
local curr_info = info[character]
local next_chars = sub(text, 3, 4) or ""
+
-- Split vowels between tokens if not a diphthong.
-- Look for a diacritic and add it to the current token. Remove it from the text.
+
if curr_info.vowel then
if is_diacritic[char] then
+
if prev and (not (curr_info.offglide and prev_info.vowel)
text = add(tokens, i, char, text)
+
-- υυ → υ, υ
--[[ See if the next two characters form a diphthong and if so,
+
-- ιυ → ι, υ
add them to the current token. Remove them from the text.
+
or prev_info.offglide and curr_info == upsilon) then
If there's a diaeresis, it will be immediately after
+
token_i = token_i + 1
the second of the two characters, or after a macron or breve. ]]
+
end
elseif is_diphthong(chars, next_chars) then
+
tokens[token_i] = (tokens[token_i] or "") .. character
i = i + 1
+
table.insert(vowel_info, { index = token_i })
text = add(tokens, i, chars, text)
+
elseif curr_info.diacritic then
 +
tokens[token_i] = (tokens[token_i] or "") .. character
 +
if prev_info.vowel or prev_info.diacritic then
 +
if character == diaeresis then
 +
-- Current token is vowel, vowel, possibly other diacritics,
 +
-- and a diaeresis.
 +
-- Split the current token into two:
 +
-- the first letter, then the second letter plus any diacritics.
 +
local previous_vowel, vowel_with_diaeresis = string.match(tokens[token_i], "^(" .. 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
 +
end
 +
end
 +
elseif prev_info == rho then
 +
if curr_info ~= breathing then
 +
return string.format("The character %s cannot have the accent %s on it.", prev, "◌" .. character)
 +
end
 +
else
 +
error("The character " .. quote(prev) .. " cannot have a diacritic on it.")
 +
end
 +
elseif curr_info == rho then
 +
if prev and not (prev_info == breathing and info[string.match(tokens[token_i], "^" .. basic_Greek)] == rho) then
 +
token_i = token_i + 1
 +
end
 +
tokens[token_i] = (tokens[token_i] or "") .. character
 
else
 
else
-- Add the current character to the next token. Remove it from the text.
+
if prev then
i = i + 1
+
token_i = token_i + 1
text = add(tokens, i, char, text)
+
end
 +
tokens[token_i] = (tokens[token_i] or "") .. character
 
end
 
end
 +
prev = character
 +
prev_info = curr_info
 
end
 
end
+
return tokens
tokens.maxindex = i
 
 
return m_table.compressSparseArray(tokens)
 
 
end
 
end
  
function p.transliterate(text, lang, sc)
+
function p.transliterate(text)
 
text = decompose(text)
 
text = decompose(text)
 
--[[
 
--[[

Revision as of 18:51, 22 November 2017

Usage

This module transliterates Ancient Greek text. It is based on an old version of the Ancient Greek transliteration module on Wiktionary, with minor modifications to make it callable through a template.

{{#invoke:Ancient Greek|translit|οἷος}}
  • hoîos

The code below uses the basic string functions (for instance, str:gsub(...)) when possible. Ustring functions have to be used when patterns contain sets with multiple-byte characters (for instance, "[αΑ]"), or quantifiers that act on multiple-byte characters ("α+"). And they must be used to correctly get a substring of the ith to the jth Unicode character. In other situations, basic string functions can be used, and are preferred for efficiency's sake, as they don't have to parse the string into codepoints before operating on it.


local p = {}

local ustring = mw.ustring
local U = ustring.char
local find = ustring.find
local gsub = ustring.gsub
local decompose = ustring.toNFD
local lower = ustring.lower
local upper = ustring.upper

local str_gmatch = string.gmatch

local macron = U(0x304)
local breve = U(0x306)
local rough = U(0x314)
local smooth = U(0x313)
local diaeresis = U(0x308)
local acute = U(0x301)
local grave = U(0x300)
local circumflex = U(0x342)
local Latin_circumflex = U(0x302)
local subscript = U(0x345)
local macron_circumflex = macron .. diaeresis .. '?' .. Latin_circumflex

local is_velar = { ['κ'] = true, ['γ'] = true, ['χ'] = true, ['ξ'] = true, }

local UTF8_char = "[%z\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 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 = { vowel = true, diacritic_seat = true }
local iota = { vowel = true, diacritic_seat = true, offglide = true }
local upsilon = { vowel = true, diacritic_seat = true, offglide = true }
-- Technically rho is only a seat for rough or smooth breathing.
local rho = { consonant = true, diacritic_seat = true }
local consonant = { consonant = true }
local diacritic = { diacritic = true }
-- Needed for equality comparisons.
local breathing = { 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
		end
	else
		for _, character in ipairs(characters) do
			info[character] = t
		end
	end
end

add_info({ macron, breve,
		diaeresis,
		acute, grave, circumflex,
		subscript,
	}, diacritic)

add_info({rough, smooth}, breathing)
add_info("ΑΕΗΟΩαεηοω", vowel)
add_info("Ιι", iota)
add_info("Υυ", upsilon)
add_info("ΒΓΔΖΘΚΛΜΝΞΠΡΣΤΦΧΨϜϘϺϷͶϠβγδζθκλμνξπρσςτφχψϝϙϻϸͷϡ", consonant)
add_info("Ρρ", rho)

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

local function quote(str)
	return "“" ..  str .. "”"
end

local tt = {
	-- Vowels
	["α"] = "a",
	["ε"] = "e",
	["η"] = "e" .. macron,
	["ι"] = "i",
	["ο"] = "o",
	["υ"] = "u",
	["ω"] = "o" .. macron,

	-- Consonants
	["β"] = "b",
	["γ"] = "g",
	["δ"] = "d",
	["ζ"] = "z",
	["θ"] = "th",
	["κ"] = "k",
	["λ"] = "l",
	["μ"] = "m",
	["ν"] = "n",
	["ξ"] = "x",
	["π"] = "p",
	["ρ"] = "r",
	["σ"] = "s",
	["ς"] = "s",
	["τ"] = "t",
	["φ"] = "ph",
	["χ"] = "kh",
	["ψ"] = "ps",
	
	-- Archaic letters
	["ϝ"] = "w",
	["ϻ"] = "ś",
	["ϙ"] = "q",
	["ϡ"] = "š",
	["ͷ"] = "v",
	
	-- Diacritics
	-- unchanged: macron, diaeresis, grave, acute
	[breve] = '',
	[smooth] = '',
	[rough] = '',
	[circumflex] = Latin_circumflex,
	[subscript] = 'i',
}

--[=[
		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 tokenize(text)
	local tokens, vowel_info, prev_info = {}, {}, {}
	local token_i = 1
	local prev
	for character in str_gmatch(decompose(text), UTF8_char) do
		local curr_info = info[character]
		-- Split vowels between tokens if not a diphthong.
		if curr_info.vowel then
			if prev and (not (curr_info.offglide and prev_info.vowel)
					-- υυ → υ, υ
					-- ιυ → ι, υ
					or prev_info.offglide and curr_info == upsilon) then
				token_i = token_i + 1
			end
			tokens[token_i] = (tokens[token_i] or "") .. character
			table.insert(vowel_info, { index = token_i })
		elseif curr_info.diacritic then
			tokens[token_i] = (tokens[token_i] or "") .. character
			if prev_info.vowel or prev_info.diacritic then
				if character == diaeresis then
					-- Current token is vowel, vowel, possibly other diacritics,
					-- and a diaeresis.
					-- Split the current token into two:
					-- the first letter, then the second letter plus any diacritics.
					local previous_vowel, vowel_with_diaeresis = string.match(tokens[token_i], "^(" .. 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
					end
				end
			elseif prev_info == rho then
				if curr_info ~= breathing then
					return string.format("The character %s cannot have the accent %s on it.", prev, "◌" .. character)
				end
			else
				error("The character " .. quote(prev) .. " cannot have a diacritic on it.")
			end
		elseif curr_info == rho then
			if prev and not (prev_info == breathing and info[string.match(tokens[token_i], "^" .. basic_Greek)] == rho) then
				token_i = token_i + 1
			end
			tokens[token_i] = (tokens[token_i] or "") .. character
		else
			if prev then
				token_i = token_i + 1
			end
			tokens[token_i] = (tokens[token_i] or "") .. character
		end
		prev = character
		prev_info = curr_info
	end
	return tokens
end

function p.transliterate(text)
	text = decompose(text)
	--[[
	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 = gsub(text, "([^A-Za-z0-9])[;" .. U(0x37E) .. "]", "%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
		-- substitute each character in the token for its transliteration
		local translit = gsub(mw.ustring.lower(token), '.', tt)
		
		if token == 'γ' and is_velar[tokens[i + 1]] then
			-- γ before a velar should be <n>
			translit = 'n'
		elseif token == 'ρ' and tokens[i - 1] == 'ρ' then
			-- ρ after ρ should be <rh>
			translit = 'rh'
		elseif find(token, '^[αΑ].*' .. subscript .. '$') then
			-- add macron to ᾳ
			translit = gsub(translit, '([aA])', '%1' .. macron)
		end
		
		if token:find(rough) then
			if find(token, '[Ρρ]') then
				translit = translit .. 'h'
			else -- vowel
				translit = 'h' .. translit
			end
		end
		
		-- Remove macron from a vowel that has a circumflex.
		if find(translit, macron_circumflex) then
			translit = translit:gsub(macron, '')
		end
		
		-- Capitalize first character of transliteration.
		if token ~= lower(token) then
			translit = gsub(translit, "^.", upper)
		end
		
		table.insert(output, translit)
	end
	
	return table.concat(output)
end

function p.translit(frame)
	local args = mw.getCurrentFrame():getParent().args
	local text = frame.args[1] or args[1]
	local transliteration = p.transliterate(text)
	return '<span title="Ancient Greek transliteration" class="Unicode" style="white-space:normal; text-decoration: none"><i>' .. transliteration .. '</i></span>'
end


return p