Module:Unicode data

Documentation for this module may be created at Module:Unicode data/doc

local p = {}

local floor = math.floor

local function errorf(level, ...)
	if type(level) == number then
		return error(string.format(...), level + 1)
	else -- level is actually the format string.
		return error(string.format(level, ...), 2)
	end
end

-- For the algorithm used to generate Hangul Syllable names,
-- see "Hangul Syllable Name Generation" in section 3.12 of the
-- Unicode Specification:
-- https://www.unicode.org/versions/Unicode11.0.0/ch03.pdf
local Hangul_data -- loaded if needed

local name_hooks = {
	{     0x00,     0x1F, "<control-%04X>" }, -- C0 control characters
	{     0x7F,     0x9F, "<control-%04X>" }, -- DEL and C1 control characters
	{   0x3400,   0x4DB5, "CJK UNIFIED IDEOGRAPH-%04X" }, -- CJK Ideograph Extension A
	{   0x4E00,   0x9FEF, "CJK UNIFIED IDEOGRAPH-%04X" }, -- CJK Ideograph
	{   0xAC00,   0xD7A3, function (codepoint) -- Hangul Syllables
		Hangul_data = Hangul_data or mw.loadData("Module:Unicode data/Hangul")
		local syllable_index = codepoint - 0xAC00

		return ("HANGUL SYLLABLE %s%s%s"):format(
			Hangul_data.leads[floor(syllable_index / Hangul_data.final_count)],
			Hangul_data.vowels[floor((syllable_index % Hangul_data.final_count)
				/ Hangul_data.trail_count)],
			Hangul_data.trails[syllable_index % Hangul_data.trail_count]
		)
	end },
	-- High Surrogates, High Private Use Surrogates, Low Surrogates
	{   0xD800,   0xDFFF, "<surrogate-%04X>" },
	{   0xE000,   0xF8FF, "<private-use-%04X>" }, -- Private Use
	{   0xF900,   0xFA6D, "CJK COMPATIBILITY IDEOGRAPH-%04X" }, -- CJK Compatibility Ideographs
	{   0xFA70,   0xFAD9, "CJK COMPATIBILITY IDEOGRAPH-%04X" },
	{  0x17000,  0x187F1, "TANGUT IDEOGRAPH-%04X" }, -- Tangut
	{  0x18800,  0x18AF2, function (codepoint)
		return ("TANGUT COMPONENT-%03d"):format(codepoint - 0x187FF)
	end },
	{  0x1B170,  0x1B2FB, "NUSHU CHARACTER-%04X" }, -- Nushu
	{  0x20000,  0x2A6D6, "CJK UNIFIED IDEOGRAPH-%04X" }, -- CJK Ideograph Extension B
	{  0x2A700,  0x2B734, "CJK UNIFIED IDEOGRAPH-%04X" }, -- CJK Ideograph Extension C
	{  0x2A740,  0x2B81D, "CJK UNIFIED IDEOGRAPH-%04X" }, -- CJK Ideograph Extension D
	{  0x2B820,  0x2CEA1, "CJK UNIFIED IDEOGRAPH-%04X" }, -- CJK Ideograph Extension E
	{  0x2CEB0,  0x2EBE0, "CJK UNIFIED IDEOGRAPH-%04X" }, -- CJK Ideograph Extension F
	-- CJK Compatibility Ideographs Supplement (Supplementary Ideographic Plane)
	{  0x2F800,  0x2FA1D, "CJK COMPATIBILITY IDEOGRAPH-%04X" },
	{  0xE0100,  0xE01EF, function (codepoint) -- Variation Selectors Supplement
		return ("VARIATION SELECTOR-%d"):format(codepoint - 0xE0100 + 17)
	end},
	{  0xF0000,  0xFFFFD, "<private-use-%04X>" }, -- Plane 15 Private Use
	{ 0x100000, 0x10FFFD, "<private-use-%04X>" }  -- Plane 16 Private Use
}
name_hooks.length = #name_hooks

local function binary_range_search(codepoint, ranges)
	local low, mid, high
	low, high = 1, ranges.length or require "Module:Table".length(ranges)
	while low <= high do
		mid = floor((low + high) / 2)
		local range = ranges[mid]
		if codepoint < range[1] then
			high = mid - 1
		elseif codepoint <= range[2] then
			return range, mid
		else
			low = mid + 1
		end
	end
	return nil, mid
end

local function linear_range_search(codepoint, ranges)
	for i, range in ipairs(ranges) do
		if range[1] <= codepoint and codepoint <= range[2] then
			return range
		end
	end
end

local name_range_cache

local function generate_name(data, codepoint)
	if type(data) == "string" then
		return data:format(codepoint)
	else
		return data(codepoint)
	end
end

-- https://www.unicode.org/versions/Unicode11.0.0/ch04.pdf, section 4.8
function p.lookup_name(codepoint)
	require 'libraryUtil'.checkType('lookup_name', 1, codepoint, 'number')
	if codepoint < 0 or 0x10FFFF < codepoint then
		errorf("Code point %04X out of range", codepoint)
	end
	
	-- U+FDD0-U+FDEF and all code points ending in FFFE or FFFF are Unassigned
	-- (Cn) and specifically noncharacters:
	-- https://www.unicode.org/faq/private_use.html#nonchar4
	if 0xFDD0 <= codepoint and (codepoint <= 0xFDEF
			or floor(codepoint % 0x10000) >= 0xFFFE) then
		return ("<noncharacter-%04X>"):format(codepoint)
	end

	if name_range_cache -- Check if previously used "name hook" applies to this code point.
			and codepoint >= name_range_cache[1]
			and codepoint <= name_range_cache[2] then
		return generate_name(name_range_cache[3], codepoint)
	end
	
	local range = binary_range_search(codepoint, name_hooks)
	if range then
		name_range_cache = range
		return generate_name(range[3], codepoint)
	end

	local success, data = pcall(mw.loadData,
		('Module:Unicode data/names/%03X'):format(codepoint / 0x1000))
	
	if success and data[codepoint] then
		return data[codepoint]
	
	-- Unassigned (Cn) consists of noncharacters and reserved characters.
	-- The character has been established not to be a noncharacter,
	-- and if it were assigned, its name would already been retrieved,
	-- so it must be reserved.
	else
		return ("<reserved-%04X>"):format(codepoint)
	end
end

function p.lookup_image(codepoint)
	local success, data = pcall(mw.loadData,
		('Module:Unicode data/images/%03X'):format(codepoint / 0x1000)
	)
	
	if success then
		return data[codepoint]
	end
end

function p.template_lookup_name(frame)
	local param = frame.args[1] or frame:getParent().args[1]
	
	local codepoint = tonumber(param, 16)
	if not codepoint then
		errorf("Expected a code point in hexadecimal base, got '%s'", param)
	end
	
	local name = p.lookup_name(codepoint):gsub("<", "&lt;")
	return name
end

local planes = {
	[ 0] = "Basic Multilingual Plane";
	[ 1] = "Supplementary Multilingual Plane";
	[ 2] = "Supplementary Ideographic Plane";
	[13] = "Supplementary Special-purpose Plane";
	[14] = "Supplementary Private Use Area-A";
	[15] = "Supplementary Private Use Area-B";
}

-- Load [[Module:Unicode data/blocks]] if needed and assign it to this variable.
local blocks

local function block_iter(blocks, i)
	i = i + 1
	local data = blocks[i]
	if data then
		return i, unpack(data)
	end
end

-- An ipairs-type iterator generator for the list of blocks.
function p.enum_blocks()
	blocks = blocks or mw.loadData("Module:Unicode data/blocks")
	return block_iter, blocks, 0
end

function p.lookup_plane(codepoint)
	local i = floor(codepoint / 0x10000)
	return planes[i] or ("Plane %u"):format(i)
end

function p.lookup_block(codepoint)
	blocks = blocks or mw.loadData("Module:Unicode data/blocks")
	local range = binary_range_search(codepoint, blocks)
	if range then
		return range[3]
	else
		return "No Block"
	end
end

function p.get_block_info(name)
	blocks = blocks or mw.loadData("Module:Unicode data/blocks")
	
	for i, block in ipairs(blocks) do
		if block[3] == name then
			return block
		end
	end
end

function p.is_valid_pagename(pagename)
	local has_nonws = false

	for cp in mw.ustring.gcodepoint(pagename) do
		if (cp == 0x0023) -- #
		or (cp == 0x005B) -- [
		or (cp == 0x005D) -- ]
		or (cp == 0x007B) -- {
		or (cp == 0x007C) -- |
		or (cp == 0x007D) -- }
		or (cp == 0x180E) -- MONGOLIAN VOWEL SEPARATOR
		or ((cp >= 0x2000) and (cp <= 0x200A)) -- spaces in General Punctuation block
		or (cp == 0xFFFD) -- REPLACEMENT CHARACTER
		then
			return false
		end

		local printable, result = p.is_printable(cp)
		if not printable then
			return false
		end

		if result ~= "space-separator" then
			has_nonws = true
		end
	end

	return has_nonws
end

local function manual_unpack(what, from)
	local result = {}
	from = from or 1
	for i, item in ipairs(what) do
		if i >= from then
			table.insert(result, item)
		end
	end
	return unpack(result)
end

-- Creates a function to look up data in a module that contains "singles" (a
-- code point-to-data map) and "ranges" (an array containing arrays that contain
-- the low and high code points of a range and the data associated with that
-- range).
-- "loader" loads and returns the "singles" and "ranges" tables.
-- "match_func" is passed the code point and either the data or the "dots", and
-- generates the final result of the function.
-- The varargs ("dots") describes the default data to be returned if there wasn't
-- a match.
-- In case the function is used more than once, "cache" saves ranges that have
-- already been found to match, or a range whose data is the default if there
-- was no match.
local function memo_lookup(loader, match_func, ...)
	local dots = { ... }
	local cache = {}
	local singles, ranges

	return function (codepoint)
		if not singles then
			singles, ranges = loader()
		end

		if singles[codepoint] then
			return match_func(codepoint, singles[codepoint])
		end

		local lastlast = -1
		local range = linear_range_search(codepoint, cache)
		if range then
			return match_func(codepoint, unpack(range, 3))
		end
		
		local range, index = binary_range_search(codepoint, ranges)
		if range then
			table.insert(cache, { manual_unpack(range) })
			return match_func(codepoint, manual_unpack(range, 3))
		end
		
		if ranges[index] then
			local dots_range
			if codepoint > ranges[index][2] then
				dots_range = {
					ranges[index][2] + 1,
					ranges[index + 1] and ranges[index + 1][1] - 1 or 0x10FFFF,
					unpack(dots)
				}
			else -- codepoint < range[index][1]
				dots_range = {
					ranges[index - 1] and ranges[index - 1][2] + 1 or 0,
					ranges[index][1] - 1,
					unpack(dots)
				}
			end
			table.insert(cache, dots_range)	
		end
		
		return match_func(codepoint)
	end
end

-- Get a code point's combining class value in [[Module:Unicode data/combining]],
-- and return whether this value is not zero. Zero is assigned as the default
-- if the combining class value is not found in this data module.
-- That is, return true if character is combining, or false if it is not.
-- See http://www.unicode.org/reports/tr44/#Canonical_Combining_Class_Values for
-- more information.
p.is_combining = memo_lookup(function ()
	local m_comb = mw.loadData('Module:Unicode data/combining')
	return m_comb.singles, m_comb.ranges
end, function (codepoint, combining_class)
	return combining_class and combining_class ~= 0
		or false
end, 0)

function p.add_dotted_circle(str)
	return (mw.ustring.gsub(str, ".",
		function(char)
			if p.is_combining(mw.ustring.codepoint(char)) then
				return '◌' .. char
			end
		end))
end

local lookup_control = memo_lookup(function ()
	local m_cc = mw.loadData('Module:Unicode data/control')
	return m_cc.singles, m_cc.ranges
end, function (codepoint, ccc)
	return ccc or "assigned"
end, "assigned")

function p.is_assigned(codepoint)
	return lookup_control(codepoint) ~= "unassigned"
end

function p.is_printable(codepoint)
	local result = lookup_control(codepoint)
	return (result == "assigned") or (result == "space-separator"), result
end

function p.is_whitespace(codepoint)
	local result = lookup_control(codepoint)
	return (result == "space-separator"), result
end

p.lookup_category = memo_lookup(
	function ()
		local category_data =  mw.loadData "Module:Unicode data/category"
		return category_data.singles, category_data.ranges
	end,
	function (codepoint, category)
		return category
	end,
	"Cn")

local script_range_cache
local script_data
local function lookup_script(codepoint)
	script_data = script_data or mw.loadData("Module:Unicode data/scripts")
	if script_data.singles[codepoint] then
		return script_data.singles[codepoint]
	end
	
	if script_range_cache then
		local range = linear_range_search(codepoint, script_range_cache)
		if range then
			return range[3]
		end
	end
	
	local range = binary_range_search(codepoint, script_data.ranges)
	if range then
		script_range_cache = script_range_cache or {}
		table.insert(script_range_cache, range)
		return range[3]
	end
	
	return "Zzzz"
end
p.lookup_script = lookup_script

function p.get_best_script(str)
	local scripts = {}
	for codepoint in mw.ustring.gcodepoint(str) do
		local script = lookup_script(codepoint)
		
		-- Ignore "Inherited", "Undetermined", or "Uncoded" scripts.
		if not (script == "Zyyy" or script == "Zinh" or script == "Zzzz") then
			scripts[script] = true
		end
	end
	
	-- If scripts does not contain two or more keys,
	-- return first and only key (script code) in table.
	if not next(scripts, next(scripts)) then
		return next(scripts)
	end -- else return majority script, or else "Zzzz"?
end

function p.is_Latin(str)
	-- Search for the leading bytes that introduce the UTF-8 encoding of the
	-- code points U+0340-U+10FFFF. See the codepage in the [[UTF-8]] article.
	-- If they are not found, then there is no need to test the scripts in the
	-- string, because the only scripts found below U+0370 (the first code point
	-- of the Greek and Coptic block) are Latn, Zyyy, Zinh, and Zzzz.
	if not str:find "[\205-\244]" then
		return true
	end
	
	for codepoint in mw.ustring.gcodepoint(str) do
		local script = lookup_script(codepoint)
		
		if not (script == "Latn" or script == "Zyyy" or script == "Zinh" or script == "Zzzz") then
			return false
		end
	end
	
	return true
end

return p