Module:scripts: Difference between revisions

From Linguifex
Jump to navigation Jump to search
(Undo revision 320504 by Sware (talk))
Tag: Undo
No edit summary
 
(One intermediate revision by the same user not shown)
Line 1: Line 1:
local m_str_utils = require("Module:string utilities")
local concat = table.concat
local explode = m_str_utils.explode_utf8
local gsplit = m_str_utils.gsplit
local match = string.match
local select = select
local split = m_str_utils.split
local toNFC = mw.ustring.toNFC
local toNFD = mw.ustring.toNFD
local toNFKC = mw.ustring.toNFKC
local toNFKD = mw.ustring.toNFKD
local type = type
local ugsub = m_str_utils.gsub
local umatch = m_str_utils.match
local export = {}
local export = {}
local Script = {}
local Script = {}


--[==[Returns the script code of the script. Example: {{lua|"Cyrl"}} for Cyrillic.]==]
function Script:getCode()
function Script:getCode()
return self._code
return self._code
end
end


--[==[Returns the canonical name of the script. This is the name used to represent that script on Wiktionary. Example: {{lua|"Cyrillic"}} for Cyrillic.]==]
function Script:getCanonicalName()
function Script:getCanonicalName()
return self._rawData.canonicalName
return self._rawData[1] or self._rawData.canonicalName
end
end


--[==[Returns the display form of the script. For scripts, this is the same as the value returned by <code>:getCategoryName("nocap")</code>, i.e. it reads "NAME script" (e.g. {{lua|"Arabic script"}}). The displayed text used in <code>:makeCategoryLink</code> is always the same as the display form.]==]
function Script:getDisplayForm()
function Script:getDisplayForm()
return self:getCategoryName("nocap")
return self:getCategoryName("nocap")
end
--[==[Returns the text direction, if any. Currently, left-to-right scripts are unmarked, while most right-to-left scripts have direction specified as {{code|lua|"rtl"}} and Mongolian as {{code|lua|"down"}}.]==]
function Script:getDirection()
return self._rawData.direction
end
end


Line 31: Line 46:
end
end


--[==[Returns the {{w|IETF language tag#Syntax of language tags|IETF subtag}} used for the script, which should always be a valid {{w|ISO 15924}} script code. This is used when constructing HTML {{code|html|lang{{=}}}} tags. The {{lua|ietf_subtag}} value from the script's data file is used, if present; otherwise, the script code is used. For script codes which contain a hyphen, only the part after the hyphen is used (e.g. {{lua|"fa-Arab"}} becomes {{lua|"Arab"}}).]==]
function Script:getIETFSubtag()
local code = self._ietf_subtag
if code == nil then
code = self._rawData.ietf_subtag or match(self._code, "[^%-]+$")
self._ietf_subtag = code
end
return code
end
--[==[Returns the parent of the script. Example: {{lua|"Arab"}} for {{lua|"fa-Arab"}}. It returns {{lua|"top"}} for scripts without a parent, like {{lua|"Latn"}}, {{lua|"Grek"}}, etc.]==]
function Script:getParent()
function Script:getParent()
return self._rawData.parent
return self._rawData.parent
end
function Script:getSystemCodes()
if not self._systemCodes then
local system_codes = self._rawData[2]
if type(system_codes) == "table" then
self._systemCodes = system_codes
elseif type(system_codes) == "string" then
self._systemCodes = split(system_codes, "%s*,%s*", true)
else
self._systemCodes = {}
end
end
return self._systemCodes
end
end


Line 40: Line 80:
self._systemObjects = {}
self._systemObjects = {}
for _, sys in ipairs(self._rawData.systems or {}) do
for _, ws in ipairs(self:getSystemCodes()) do
table.insert(self._systemObjects, m_systems.getByCode(sys))
table.insert(self._systemObjects, m_systems.getByCode(ws))
end
end
end
end
return self._systemObjects
return self._systemObjects
end
--[==[Check whether the script is of type `system`, which can be a writing system code or object. If multiple systems are passed, return true if the script is any of the specified systems.]==]
function Script:isSystem(...)
for _, system in ipairs{...} do
if type(system) == "table" then
system = system:getCode()
end
for _, s in ipairs(self:getSystemCodes()) do
if system == s then
return true
end
end
end
return false
end
end


Line 52: Line 107:
--end
--end


--[==[Given a list of types as strings, returns true if the script has all of them.


function Script:getType()
Currently the only possible type is {script}; use {{lua|hasType("script")}} to determine if an object that
return "script"
may be a language, family or script is a script.
]==]
function Script:hasType(...)
local types = self._types
if types == nil then
types = {script = true}
local rawtypes = self._rawData.type
if rawtypes then
for rawtype in gsplit(rawtypes, "%s*,%s*", true) do
types[rawtype] = true
end
end
self._types = types
end
for i = 1, arg.n do
if not types[arg[i]] then
return false
end
end
return true
end
end


 
--[==[Returns the name of the main category of that script. Example: {{lua|"Cyrillic script"}} for Cyrillic, whose category is at [[:Category:Cyrillic script]].
Unless optional argument <code>nocap</code> is given, the script name at the beginning of the returned value will be capitalized. This capitalization is correct for category names, but not if the script name is lowercase and the returned value of this function is used in the middle of a sentence. (For example, the script with the code <code>Semap</code> has the name <code>"flag semaphore"</code>, which should remain lowercase when used as part of the category name [[:Category:Translingual letters in flag semaphore]] but should be capitalized in [[:Category:Flag semaphore templates]].) If you are considering using <code>getCategoryName("nocap")</code>, use <code>getDisplayForm()</code> instead.]==]
function Script:getCategoryName(nocap)
function Script:getCategoryName(nocap)
local name = self._rawData.canonicalName
local name = self:getCanonicalName()
-- If the name already has "code" or "semaphore" in it, don't add it.
-- If the name already has "script", "code" or "semaphore" at the end, don't add it.
-- No names contain "script".
if not (
if not name:find("[Cc]ode$") and not name:find("[Ss]emaphore$") then
name:find("[ %-][Ss]cript$") or
name:find("[ %-][Cc]ode$") or
name:find("[ %-][Ss]emaphore$")
) then
name = name .. " script"
name = name .. " script"
end
end
Line 71: Line 150:
return name
return name
end
end


function Script:makeCategoryLink()
function Script:makeCategoryLink()
Line 77: Line 155:
end
end


 
--[==[Returns the {{lua|wikipedia_article}} item in the script's data file, or else calls {{lua|Script:getCategoryName()}}.]==]
function Script:getWikipediaArticle()
function Script:getWikipediaArticle()
return self._rawData.wikipedia_article or self:getCategoryName()
return self._rawData.wikipedia_article or self:getCategoryName()
end
end


 
--[==[Returns the charset defining the script's characters from the script's data file.
This can be used to search for words consisting only of this script, but see the warning above.]==]
function Script:getCharacters()
function Script:getCharacters()
if self._rawData.characters then
return self.characters or nil
return self._rawData.characters
else
return nil
end
end
end


 
--[==[Returns the number of characters in the text that are part of this script.
'''Note:''' You should never assume that text consists entirely of the same script. Strings may contain spaces, punctuation and even wiki markup or HTML tags. HTML tags will skew the counts, as they contain Latin-script characters. So it's best to avoid them.]==]
function Script:countCharacters(text)
function Script:countCharacters(text)
if not self._rawData.characters then
local charset = self._rawData.characters
if charset == nil then
return 0
return 0
else
local _, num = mw.ustring.gsub(text, "[" .. self._rawData.characters .. "]", "")
return num
end
end
return select(2, ugsub(text, "[" .. charset .. "]", ""))
end
function Script:hasCapitalization()
return not not self._rawData.capitalized
end
function Script:hasSpaces()
return self._rawData.spaces ~= false
end
function Script:isTransliterated()
return self._rawData.translit ~= false
end
--[==[Returns true if the script is (sometimes) sorted by scraping page content, meaning that it is sensitive to changes in capitalization during sorting.]==]
function Script:sortByScraping()
return not not self._rawData.sort_by_scraping
end
end


--[==[Returns the text direction. Horizontal scripts return {{lua|"ltr"}} (left-to-right) or {{lua|"rtl"}} (right-to-left), while vertical scripts return {{lua|"vertical-ltr"}} (vertical left-to-right) or {{lua|"vertical-rtl"}} (vertical right-to-left).]==]
function Script:getDirection()
function Script:getDirection()
local direction = self._rawData.direction
return self._rawData.direction or "ltr"
if not direction then
return nil
else
return direction
end
end
end


function Script:getRawData()
function Script:getRawData()
Line 115: Line 202:
end
end


--[==[Returns {{lua|true}} if the script contains characters that require fixes to Unicode normalization under certain circumstances, {{lua|false}} if it doesn't.]==]
function Script:hasNormalizationFixes()
return not not self._rawData.normalizationFixes
end
--[==[Corrects discouraged sequences of Unicode characters to the encouraged equivalents.]==]
function Script:fixDiscouragedSequences(text)
if self:hasNormalizationFixes() then
local norm_fixes = self._rawData.normalizationFixes
local to = norm_fixes.to
if to then
for i, v in ipairs(norm_fixes.from) do
text = ugsub(text, v, to[i] or "")
end
end
end
return text
end
do
local combiningClasses
-- Implements a modified form of Unicode normalization for instances where there are identified deficiencies in the default Unicode combining classes.
local function fixNormalization(text, self)
if not self:hasNormalizationFixes() then
return text
end
local norm_fixes = self._rawData.normalizationFixes
local new_classes = norm_fixes.combiningClasses
if not (new_classes and umatch(text, "[" .. norm_fixes.combiningClassCharacters .. "]")) then
return text
end
-- Obtain the list of default combining classes.
combiningClasses = combiningClasses or mw.loadData("Module:scripts/data/combiningClasses")
text = explode(text)
-- Manual sort based on new combining classes.
-- We can't use table.sort, as it compares the first/last values in an array as a shortcut, which messes things up.
for i = 2, #text do
local char = text[i]
local class = new_classes[char] or combiningClasses[char]
if class then
repeat
i = i - 1
local prev = text[i]
if (new_classes[prev] or combiningClasses[prev] or 0) < class then
break
end
text[i], text[i + 1] = char, prev
until i == 1
end
end
return concat(text)
end
function Script:toFixedNFC(text)
return fixNormalization(toNFC(text), self)
end
function Script:toFixedNFD(text)
return fixNormalization(toNFD(text), self)
end
function Script:toFixedNFKC(text)
return fixNormalization(toNFKC(text), self)
end
function Script:toFixedNFKD(text)
return fixNormalization(toNFKD(text), self)
end
end


function Script:toJSON()
function Script:toJSON()
if not self._types then
self:hasType()
end
local types = {}
for type in pairs(self._types) do
table.insert(types, type)
end
local ret = {
local ret = {
canonicalName = self:getCanonicalName(),
canonicalName = self:getCanonicalName(),
Line 124: Line 289:
aliases = self:getAliases(),
aliases = self:getAliases(),
varieties = self:getVarieties(),
varieties = self:getVarieties(),
type = self:getType(),
type = types,
direction = self:getDirection(),
direction = self:getDirection(),
characters = self:getCharacters(),
characters = self:getCharacters(),
parent = self:getParent(),
parent = self:getParent(),
systems = self._rawData.systems or {},
systems = self:getSystemCodes(),
wikipediaArticle = self._rawData.wikipedia_article,
wikipediaArticle = self._rawData.wikipedia_article,
}
}
Line 134: Line 299:
return require("Module:JSON").toJSON(ret)
return require("Module:JSON").toJSON(ret)
end
end


Script.__index = Script
Script.__index = Script
 
 
function export.makeObject(code, data, useRequire)
function export.makeObject(code, data)
return data and setmetatable({
return data and setmetatable({ _rawData = data, _code = code }, Script) or nil
_rawData = data,
_code = code,
characters = data.characters
}, Script) or nil
end
end


 
--[==[Finds the script whose code matches the one provided. If it exists, it returns a {{lua|Script}} object representing the script. Otherwise, it returns {{lua|nil}}, unless <span class="n">paramForError</span> is given, in which case an error is generated. If <code class="n">paramForError</code> is {{lua|true}}, a generic error message mentioning the bad code is generated; otherwise <code class="n">paramForError</code> should be a string or number specifying the parameter that the code came from, and this parameter will be mentioned in the error message along with the bad code.]==]
function export.getByCode(code, paramForError, disallowNil)
function export.getByCode(code, paramForError, disallowNil, useRequire)
if code == nil and not disallowNil then
if code == nil and not disallowNil then
return nil
return nil
end
end
local retval = export.makeObject(code, mw.loadData("Module:scripts/data")[code])
if not retval and paramForError then
require("Module:languages").err(code, paramForError, "script code", nil, "not real lang")
end
return retval
end
function export.getByCanonicalName(name)
local code = mw.loadData("Module:scripts/by name")[name]
if not code then
local data
return nil
if useRequire then
end
data = require("Module:scripts/data")[code]
else
return export.makeObject(code, mw.loadData("Module:scripts/data")[code])
data = mw.loadData("Module:scripts/data")[code]
end
 
-- Find the best script to use, based on the characters of a string.
-- If forceDetect is set, run the detection algorithm even if there's only one
-- possible script; in that case, if the text isn't in the script, the return
-- value will be None.
function export.findBestScript(text, lang, forceDetect)
if not text or not lang or not lang.getScripts then
return export.getByCode("None")
end
end
local scripts = lang:getScripts()
local retval = export.makeObject(code, data, useRequire)
if not scripts[2] and not forceDetect then
if not retval and paramForError then
return scripts[1]
require("Module:languages/error")(code, paramForError, "script code", nil, "not real lang")
end
end
--[=[
return retval
Remove any HTML entities; catfix function in [[Module:utilities]]
adds tagging to a no-break space (&nbsp;), which contains Latin characters;
hence Latin was returned as the script if "Latn" is one of the language's scripts.
]=]
text = string.gsub(text, "&[a-zA-Z0-9]+;", "")
-- Try to match every script against the text,
-- and return the one with the most matching characters.
local bestcount = 0
local bestscript = nil
-- Get length of text minus any spacing or punctuation characters.
-- Counting instances of UTF-8 character pattern is faster than mw.ustring.len.
local _, length = string.gsub(mw.ustring.gsub(text, "[%s%p]+", ""), "[\1-\127\194-\244][\128-\191]*", "")
if length == 0 then
return export.getByCode("None")
end
for i, script in ipairs(scripts) do
local count = script:countCharacters(text)
if count >= length then
return script
end
if count > bestcount then
bestcount = count
bestscript = script
end
end
if bestscript then
return bestscript
end
-- No matching script was found. Return "None".
return export.getByCode("None")
end
end


-- Copied from [[Module:Unicode data]].
function export.getByCanonicalName(name, useRequire)
local floor = math.floor
local code
local function binaryRangeSearch(codepoint, ranges)
if useRequire then
local low, mid, high
code = require("Module:scripts/by name")[name]
low, high = 1, ranges.length or require "Module:table".length(ranges)
else
while low <= high do
code = mw.loadData("Module:scripts/by name")[name]
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
end
return nil, mid
end
return export.getByCode(code, nil, nil, useRequire)
 
-- Copied from [[Module:Unicode data]].
local function linearRangeSearch(codepoint, ranges)
for i, range in ipairs(ranges) do
if codepoint < range[1] then
break
elseif codepoint <= range[2] then
return range
end
end
end
 
local function compareRanges(range1, range2)
return range1[1] < range2[1]
end
end


-- Save previously used codepoint ranges in case another character is in the
--[==[
-- same range.
local rangesCache = {}
 
--[=[
Takes a codepoint or a character and finds the script code (if any) that is
Takes a codepoint or a character and finds the script code (if any) that is
appropriate for it based on the codepoint, using the data module
appropriate for it based on the codepoint, using the data module
Line 268: Line 352:
is in the list of individual characters, or if it is in one of the defined
is in the list of individual characters, or if it is in one of the defined
ranges in the 4096-character block that it belongs to, else returns "None".
ranges in the 4096-character block that it belongs to, else returns "None".
]=]
]==]
local charToScriptData
function export.charToScript(char)
function export.charToScript(char)
charToScriptData = charToScriptData or mw.loadData("Module:scripts/recognition data")
return require("Module:scripts/charToScript").charToScript(char)
local t = type(char)
local codepoint
if t == "string" then
local etc
codepoint, etc = mw.ustring.codepoint(char, 1, 2)
if etc then
error("bad argument #1 to 'charToScript' (expected a single character)")
end
elseif t == "number" then
codepoint = char
else
error(("bad argument #1 to 'charToScript' (expected string or a number, got %s)")
:format(t))
end
 
local individualMatch = charToScriptData.individual[codepoint]
if individualMatch then
return individualMatch
else
local range
if rangesCache[1] then
range = linearRangeSearch(codepoint, rangesCache)
if range then
return range[3]
end
end
 
local index = floor(codepoint / 0x1000)
 
range = linearRangeSearch(index, charToScriptData.blocks)
if not range and charToScriptData[index] then
range = binaryRangeSearch(codepoint, charToScriptData[index])
if range then
table.insert(rangesCache, range)
table.sort(rangesCache, compareRanges)
end
end
return range and range[3] or "None"
end
end
end


function export.findBestScriptWithoutLang(text)
--[==[
local scripts = {}
Returns the code for the script that has the greatest number of characters in `text`. Useful for script tagging text
for character in text:gmatch("[%z\1-\127\194-\244][\128-\191]*") do
that is unspecified for language. Uses [[Module:scripts/recognition data]] to determine a script code for a character
local script = export.charToScript(character)
language-agnostically. Specifically, it works as follows:
scripts[script] = (scripts[script] or 0) + 1
end
Convert each character to a codepoint. Iterate the counter for the script code if the codepoint is in the list
of individual characters, or if it is in one of the defined ranges in the 4096-character block that it belongs to.
local bestScript
Each script has a two-part counter, for primary and secondary matches. Primary matches are when the script is the
local greatestCount = 0
first one listed; otherwise, it's a secondary match. When comparing scripts, first the total of both are compared
for script, count in pairs(scripts) do
(i.e. the overall number of matches). If these are the same, the number of primary and then secondary matches are
if count > greatestCount then
used as tiebreakers. For example, this is used to ensure that `Grek` takes priority over `Polyt` if no characters
bestScript = script
which exclusively match `Polyt` are found, as `Grek` is a subset of `Polyt`.
greatestCount = count
end
end
return bestScript
If `none_is_last_resort_only` is specified, this will never return {"None"} if any characters in `text` belong to a
script. Otherwise, it will return {"None"} if there are more characters that don't belong to a script than belong to
any individual script. (FIXME: This behavior is probably wrong, and `none_is_last_resort_only` should probably
become the default.)
]==]
function export.findBestScriptWithoutLang(text, none_is_last_resort_only)
return require("Module:scripts/charToScript").findBestScriptWithoutLang(text, none_is_last_resort_only)
end
end


return export
return export

Latest revision as of 21:25, 6 August 2024



local m_str_utils = require("Module:string utilities")

local concat = table.concat
local explode = m_str_utils.explode_utf8
local gsplit = m_str_utils.gsplit
local match = string.match
local select = select
local split = m_str_utils.split
local toNFC = mw.ustring.toNFC
local toNFD = mw.ustring.toNFD
local toNFKC = mw.ustring.toNFKC
local toNFKD = mw.ustring.toNFKD
local type = type
local ugsub = m_str_utils.gsub
local umatch = m_str_utils.match

local export = {}

local Script = {}

--[==[Returns the script code of the script. Example: {{lua|"Cyrl"}} for Cyrillic.]==]
function Script:getCode()
	return self._code
end

--[==[Returns the canonical name of the script. This is the name used to represent that script on Wiktionary. Example: {{lua|"Cyrillic"}} for Cyrillic.]==]
function Script:getCanonicalName()
	return self._rawData[1] or self._rawData.canonicalName
end

--[==[Returns the display form of the script. For scripts, this is the same as the value returned by <code>:getCategoryName("nocap")</code>, i.e. it reads "NAME script" (e.g. {{lua|"Arabic script"}}). The displayed text used in <code>:makeCategoryLink</code> is always the same as the display form.]==]
function Script:getDisplayForm()
	return self:getCategoryName("nocap")
end

function Script:getOtherNames(onlyOtherNames)
	return require("Module:language-like").getOtherNames(self, onlyOtherNames)
end

function Script:getAliases()
	return self._rawData.aliases or {}
end

function Script:getVarieties(flatten)
	return require("Module:language-like").getVarieties(self, flatten)
end

--[==[Returns the {{w|IETF language tag#Syntax of language tags|IETF subtag}} used for the script, which should always be a valid {{w|ISO 15924}} script code. This is used when constructing HTML {{code|html|lang{{=}}}} tags. The {{lua|ietf_subtag}} value from the script's data file is used, if present; otherwise, the script code is used. For script codes which contain a hyphen, only the part after the hyphen is used (e.g. {{lua|"fa-Arab"}} becomes {{lua|"Arab"}}).]==]
function Script:getIETFSubtag()
	local code = self._ietf_subtag
	if code == nil then
		code = self._rawData.ietf_subtag or match(self._code, "[^%-]+$")
		self._ietf_subtag = code
	end
	return code
end

--[==[Returns the parent of the script. Example: {{lua|"Arab"}} for {{lua|"fa-Arab"}}. It returns {{lua|"top"}} for scripts without a parent, like {{lua|"Latn"}}, {{lua|"Grek"}}, etc.]==]
function Script:getParent()
	return self._rawData.parent
end

function Script:getSystemCodes()
	if not self._systemCodes then
		local system_codes = self._rawData[2]
		if type(system_codes) == "table" then
			self._systemCodes = system_codes
		elseif type(system_codes) == "string" then
			self._systemCodes = split(system_codes, "%s*,%s*", true)
		else
			self._systemCodes = {}
		end
	end
	return self._systemCodes
end

function Script:getSystems()
	if not self._systemObjects then
		local m_systems = require("Module:writing systems")
		self._systemObjects = {}
		
		for _, ws in ipairs(self:getSystemCodes()) do
			table.insert(self._systemObjects, m_systems.getByCode(ws))
		end
	end
	
	return self._systemObjects
end

--[==[Check whether the script is of type `system`, which can be a writing system code or object. If multiple systems are passed, return true if the script is any of the specified systems.]==]
function Script:isSystem(...)
	for _, system in ipairs{...} do
		if type(system) == "table" then
			system = system:getCode()
		end
		for _, s in ipairs(self:getSystemCodes()) do
			if system == s then
				return true
			end
		end
	end
	return false
end

--function Script:getAllNames()
--	return self._rawData.names
--end

--[==[Given a list of types as strings, returns true if the script has all of them. 

Currently the only possible type is {script}; use {{lua|hasType("script")}} to determine if an object that
may be a language, family or script is a script.
]==]	
function Script:hasType(...)
	local types = self._types
	if types == nil then
		types = {script = true}
		local rawtypes = self._rawData.type
		if rawtypes then
			for rawtype in gsplit(rawtypes, "%s*,%s*", true) do
				types[rawtype] = true
			end
		end
		self._types = types
	end
	for i = 1, arg.n do
		if not types[arg[i]] then
			return false
		end
	end
	return true
end

--[==[Returns the name of the main category of that script. Example: {{lua|"Cyrillic script"}} for Cyrillic, whose category is at [[:Category:Cyrillic script]].
Unless optional argument <code>nocap</code> is given, the script name at the beginning of the returned value will be capitalized. This capitalization is correct for category names, but not if the script name is lowercase and the returned value of this function is used in the middle of a sentence. (For example, the script with the code <code>Semap</code> has the name <code>"flag semaphore"</code>, which should remain lowercase when used as part of the category name [[:Category:Translingual letters in flag semaphore]] but should be capitalized in [[:Category:Flag semaphore templates]].) If you are considering using <code>getCategoryName("nocap")</code>, use <code>getDisplayForm()</code> instead.]==]
function Script:getCategoryName(nocap)
	local name = self:getCanonicalName()
	
	-- If the name already has "script", "code" or "semaphore" at the end, don't add it.
	if not (
		name:find("[ %-][Ss]cript$") or
		name:find("[ %-][Cc]ode$") or
		name:find("[ %-][Ss]emaphore$")
	) then
		name = name .. " script"
	end
	if not nocap then
		name = mw.getContentLanguage():ucfirst(name)
	end
	return name
end

function Script:makeCategoryLink()
	return "[[:Category:" .. self:getCategoryName() .. "|" .. self:getDisplayForm() .. "]]"
end

--[==[Returns the {{lua|wikipedia_article}} item in the script's data file, or else calls {{lua|Script:getCategoryName()}}.]==]
function Script:getWikipediaArticle()
	return self._rawData.wikipedia_article or self:getCategoryName()
end

--[==[Returns the charset defining the script's characters from the script's data file.
This can be used to search for words consisting only of this script, but see the warning above.]==]
function Script:getCharacters()
	return self.characters or nil
end

--[==[Returns the number of characters in the text that are part of this script.
'''Note:''' You should never assume that text consists entirely of the same script. Strings may contain spaces, punctuation and even wiki markup or HTML tags. HTML tags will skew the counts, as they contain Latin-script characters. So it's best to avoid them.]==]
function Script:countCharacters(text)
	local charset = self._rawData.characters
	if charset == nil then
		return 0
	end
	return select(2, ugsub(text, "[" .. charset .. "]", ""))
end

function Script:hasCapitalization()
	return not not self._rawData.capitalized
end

function Script:hasSpaces()
	return self._rawData.spaces ~= false
end

function Script:isTransliterated()
	return self._rawData.translit ~= false
end

--[==[Returns true if the script is (sometimes) sorted by scraping page content, meaning that it is sensitive to changes in capitalization during sorting.]==]
function Script:sortByScraping()
	return not not self._rawData.sort_by_scraping
end

--[==[Returns the text direction. Horizontal scripts return {{lua|"ltr"}} (left-to-right) or {{lua|"rtl"}} (right-to-left), while vertical scripts return {{lua|"vertical-ltr"}} (vertical left-to-right) or {{lua|"vertical-rtl"}} (vertical right-to-left).]==]
function Script:getDirection()
	return self._rawData.direction or "ltr"
end

function Script:getRawData()
	return self._rawData
end

--[==[Returns {{lua|true}} if the script contains characters that require fixes to Unicode normalization under certain circumstances, {{lua|false}} if it doesn't.]==]
function Script:hasNormalizationFixes()
	return not not self._rawData.normalizationFixes
end

--[==[Corrects discouraged sequences of Unicode characters to the encouraged equivalents.]==]
function Script:fixDiscouragedSequences(text)
	if self:hasNormalizationFixes() then
		local norm_fixes = self._rawData.normalizationFixes
		local to = norm_fixes.to
		if to then
			for i, v in ipairs(norm_fixes.from) do
				text = ugsub(text, v, to[i] or "")
			end
		end
	end
	return text
end

do
	local combiningClasses
	
	-- Implements a modified form of Unicode normalization for instances where there are identified deficiencies in the default Unicode combining classes.
	local function fixNormalization(text, self)
		if not self:hasNormalizationFixes() then
			return text
		end
		local norm_fixes = self._rawData.normalizationFixes
		local new_classes = norm_fixes.combiningClasses
		if not (new_classes and umatch(text, "[" .. norm_fixes.combiningClassCharacters .. "]")) then
			return text
		end
		-- Obtain the list of default combining classes.
		combiningClasses = combiningClasses or mw.loadData("Module:scripts/data/combiningClasses")
		text = explode(text)
		-- Manual sort based on new combining classes.
		-- We can't use table.sort, as it compares the first/last values in an array as a shortcut, which messes things up.
		for i = 2, #text do
			local char = text[i]
			local class = new_classes[char] or combiningClasses[char]
			if class then
				repeat
					i = i - 1
					local prev = text[i]
					if (new_classes[prev] or combiningClasses[prev] or 0) < class then
						break
					end
					text[i], text[i + 1] = char, prev
				until i == 1
			end
		end
		return concat(text)
	end
	
	function Script:toFixedNFC(text)
		return fixNormalization(toNFC(text), self)
	end
	
	function Script:toFixedNFD(text)
		return fixNormalization(toNFD(text), self)
	end
	
	function Script:toFixedNFKC(text)
		return fixNormalization(toNFKC(text), self)
	end
	
	function Script:toFixedNFKD(text)
		return fixNormalization(toNFKD(text), self)
	end
end

function Script:toJSON()
	if not self._types then
		self:hasType()
	end
	local types = {}
	for type in pairs(self._types) do
		table.insert(types, type)
	end
	
	local ret = {
		canonicalName = self:getCanonicalName(),
		categoryName = self:getCategoryName("nocap"),
		code = self._code,
		otherNames = self:getOtherNames(true),
		aliases = self:getAliases(),
		varieties = self:getVarieties(),
		type = types,
		direction = self:getDirection(),
		characters = self:getCharacters(),
		parent = self:getParent(),
		systems = self:getSystemCodes(),
		wikipediaArticle = self._rawData.wikipedia_article,
	}
	
	return require("Module:JSON").toJSON(ret)
end

Script.__index = Script
	
function export.makeObject(code, data, useRequire)
	return data and setmetatable({
		_rawData = data,
		_code = code,
		characters = data.characters
	}, Script) or nil
end

--[==[Finds the script whose code matches the one provided. If it exists, it returns a {{lua|Script}} object representing the script. Otherwise, it returns {{lua|nil}}, unless <span class="n">paramForError</span> is given, in which case an error is generated. If <code class="n">paramForError</code> is {{lua|true}}, a generic error message mentioning the bad code is generated; otherwise <code class="n">paramForError</code> should be a string or number specifying the parameter that the code came from, and this parameter will be mentioned in the error message along with the bad code.]==]
function export.getByCode(code, paramForError, disallowNil, useRequire)
	if code == nil and not disallowNil then
		return nil
	end
	
	local data
	if useRequire then
		data = require("Module:scripts/data")[code]
	else
		data = mw.loadData("Module:scripts/data")[code]
	end
	
	local retval = export.makeObject(code, data, useRequire)
	
	if not retval and paramForError then
		require("Module:languages/error")(code, paramForError, "script code", nil, "not real lang")
	end
	
	return retval
end

function export.getByCanonicalName(name, useRequire)
	local code
	if useRequire then
		code = require("Module:scripts/by name")[name]
	else
		code = mw.loadData("Module:scripts/by name")[name]
	end
	
	return export.getByCode(code, nil, nil, useRequire)
end

--[==[
	Takes a codepoint or a character and finds the script code (if any) that is
	appropriate for it based on the codepoint, using the data module
	[[Module:scripts/recognition data]]. The data module was generated from the
	patterns in [[Module:scripts/data]] using [[Module:User:Erutuon/script recognition]].

	Converts the character to a codepoint. Returns a script code if the codepoint
	is in the list of individual characters, or if it is in one of the defined
	ranges in the 4096-character block that it belongs to, else returns "None".
]==]
function export.charToScript(char)
	return require("Module:scripts/charToScript").charToScript(char)
end

--[==[
Returns the code for the script that has the greatest number of characters in `text`. Useful for script tagging text
that is unspecified for language. Uses [[Module:scripts/recognition data]] to determine a script code for a character
language-agnostically. Specifically, it works as follows:
	
Convert each character to a codepoint. Iterate the counter for the script code if the codepoint is in the list
of individual characters, or if it is in one of the defined ranges in the 4096-character block that it belongs to.
	
Each script has a two-part counter, for primary and secondary matches. Primary matches are when the script is the
first one listed; otherwise, it's a secondary match. When comparing scripts, first the total of both are compared
(i.e. the overall number of matches). If these are the same, the number of primary and then secondary matches are
used as tiebreakers. For example, this is used to ensure that `Grek` takes priority over `Polyt` if no characters
which exclusively match `Polyt` are found, as `Grek` is a subset of `Polyt`.
	
If `none_is_last_resort_only` is specified, this will never return {"None"} if any characters in `text` belong to a
script. Otherwise, it will return {"None"} if there are more characters that don't belong to a script than belong to
any individual script. (FIXME: This behavior is probably wrong, and `none_is_last_resort_only` should probably
become the default.)
]==]
function export.findBestScriptWithoutLang(text, none_is_last_resort_only)
	return require("Module:scripts/charToScript").findBestScriptWithoutLang(text, none_is_last_resort_only)
end

return export