Module:Category handler: Difference between revisions

From Guild of Archivists
m (1 revision)
(allow invocations specifying the page parameter to use the mw.loadData optimisations, and don't call mw.title.new every time)
Line 1: Line 1:
----------------------------------------------------------------------------------------------------------
--------------------------------------------------------------------------------
--                                                                                                     --
--                                                                           --
--                                         CATEGORY HANDLER                                           --
--                             CATEGORY HANDLER                             --
--                                                                                                     --
--                                                                           --
--      This module implements the {{category handler}} template in Lua, with a few improvements: all   --
--      This module implements the {{category handler}} template in Lua,     --
--      namespaces and all namespace aliases are supported, and namespace names are detected           --
--      with a few improvements: all namespaces and all namespace aliases    --
--      automatically for the local wiki. This module requires [[Module:Namespace detect]] and          --
--      are supported, and namespace names are detected automatically for    --
--      [[Module:Yesno]] to be available on the local wiki. It can be configured for different wikis    --
--      the local wiki. This module requires [[Module:Namespace detect]]     --
--      by altering the values in the "cfg" table.                                                      --
--      and [[Module:Yesno]] to be available on the local wiki. It can be     --
--                                                                                                     --
--      configured for different wikis by altering the values in             --
----------------------------------------------------------------------------------------------------------
--     [[Module:Category handler/config]], and pages can be blacklisted      --
--     from categorisation by using [[Module:Category handler/blacklist]].  --
--                                                                           --
--------------------------------------------------------------------------------


----------------------------------------------------------------------------------------------------------
-- Load required modules
--                                          Configuration data                                          --
local yesno = require('Module:Yesno')
--                      Language-specific parameter names and values can be set here.                  --
----------------------------------------------------------------------------------------------------------


local cfg = {}
-- Lazily load things we don't always need
local mShared, mappings


-- The following config values set the names of parameters that suppress categorisation. They are used
local p = {}
-- with Module:Yesno, and work as follows:
--
-- cfg.nocat:
-- Result of yesno(args[cfg.nocat])        Effect
-- true                                    Categorisation is suppressed
-- false                                    Categorisation is allowed, and the blacklist check is skipped
-- nil                                      Categorisation is allowed
--
-- cfg.categories:
-- Result of yesno(args[cfg.categories])    Effect
-- true                                    Categorisation is allowed, and the blacklist check is skipped
-- false                                    Categorisation is suppressed
-- nil                                      Categorisation is allowed
cfg.nocat = 'nocat'   
cfg.categories = 'categories'


-- The parameter name for the legacy "category2" parameter. This skips the blacklist if set to the
--------------------------------------------------------------------------------
-- cfg.category2Yes value, and suppresses categorisation if present but equal to anything other than
-- Helper functions
-- cfg.category2Yes or cfg.category2Negative.
--------------------------------------------------------------------------------
cfg.category2 = 'category2'
cfg.category2Yes = 'yes'
cfg.category2Negative = '¬'


-- cfg.subpage is the parameter name to specify how to behave on subpages. cfg.subpageNo is the value to
local function trimWhitespace(s, removeBlanks)
-- specify to not categorise on subpages; cfg.only is the value to specify to only categorise on subpages.
if type(s) ~= 'string' then
cfg.subpage = 'subpage'
return s
cfg.subpageNo = 'no'
end
cfg.subpageOnly = 'only'
s = s:match('^%s*(.-)%s*$')
if removeBlanks then
if s ~= '' then
return s
else
return nil
end
else
return s
end
end


-- The parameter for data to return in all namespaces.
--------------------------------------------------------------------------------
cfg.all = 'all'
-- CategoryHandler class
--------------------------------------------------------------------------------


-- The parameter name for data to return if no data is specified for the namespace that is detected. This
local CategoryHandler = {}
-- must be the same as the cfg.other parameter in [[Module:Namespace detect]].
CategoryHandler.__index = CategoryHandler
cfg.other = 'other'


-- The parameter name used to specify a page other than the current page; used for testing and
function CategoryHandler.new(data, args)
-- demonstration. This must be the same as the cfg.page parameter in [[Module:Namespace detect]].
local obj = setmetatable({ _data = data, _args = args }, CategoryHandler)
cfg.page = 'page'
-- Set the title object
do
local pagename = obj:parameter('demopage')
local success, titleObj
if pagename then
success, titleObj = pcall(mw.title.new, pagename)
end
if success and titleObj then
obj.title = titleObj
if titleObj == mw.title.getCurrentTitle() then
obj._usesCurrentTitle = true
end
else
obj.title = mw.title.getCurrentTitle()
obj._usesCurrentTitle = true
end
end


-- The categorisation blacklist. Pages that match Lua patterns in this list will not be categorised.
-- Set suppression parameter values
-- (However, see the explanation of cfg.nocat, cfg.categories and cfg.category2 for some exceptions.)
for _, key in ipairs{'nocat', 'categories'} do
-- If the namespace name has a space in, it must be written with an underscore, e.g. "Wikipedia_talk".
local value = obj:parameter(key)
-- Other parts of the title can have either underscores or spaces.
value = trimWhitespace(value, true)
cfg.blacklist = {
obj['_' .. key] = yesno(value)
    '^Main Page$', -- don't categorise the main page.
end
   
do
    -- Don't categorise the following pages or their subpages.
local subpage = obj:parameter('subpage')
    '^Wikipedia:Cascade%-protected items$',
local category2 = obj:parameter('category2')
    '^Wikipedia:Cascade%-protected items/.*$',
if type(subpage) == 'string' then
    '^User:UBX$', -- The userbox "template" space.
subpage = mw.ustring.lower(subpage)
    '^User:UBX/.*$',
end
    '^User_talk:UBX$',
if type(category2) == 'string' then
    '^User_talk:UBX/.*$',
subpage = mw.ustring.lower(category2)
   
end
    -- Don't categorise subpages of these pages, but allow
obj._subpage = trimWhitespace(subpage, true)
    -- categorisation of the base page.
obj._category2 = trimWhitespace(category2) -- don't remove blank values
    '^Wikipedia:Template messages/.+$',
end
   
return obj
    '/[aA]rchive' -- Don't categorise archives.
end
}


-- This is a table of namespaces to categorise by default. They should be in the format of parameter
function CategoryHandler:parameter(key)
-- names accepted by [[Module:Namespace detect]].
local parameterNames = self._data.parameters[key]
cfg.defaultNamespaces = {
local pntype = type(parameterNames)
    'main',
if pntype == 'string' or pntype == 'number' then
    'file',
return self._args[parameterNames]
    'help',
elseif pntype == 'table' then
    'category'
for _, name in ipairs(parameterNames) do
}
local value = self._args[name]
if value ~= nil then
return value
end
end
return nil
else
error(string.format(
'invalid config key "%s"',
tostring(key)
), 2)
end
end


----------------------------------------------------------------------------------------------------------
function CategoryHandler:isSuppressedByArguments()
--                                          End configuration data                                      --
return
----------------------------------------------------------------------------------------------------------
-- See if a category suppression argument has been set.
self._nocat == true
or self._categories == false
or (
self._category2
and self._category2 ~= self._data.category2Yes
and self._category2 ~= self._data.category2Negative
)


-- Get dependent modules
-- Check whether we are on a subpage, and see if categories are
local nsDetect = require('Module:Namespace detect')
-- suppressed based on our subpage status.
local yesno = require('Module:Yesno')
or self._subpage == self._data.subpageNo and self.title.isSubpage
or self._subpage == self._data.subpageOnly and not self.title.isSubpage
end
 
function CategoryHandler:shouldSkipBlacklistCheck()
-- Check whether the category suppression arguments indicate we
-- should skip the blacklist check.
return self._nocat == false
or self._categories == true
or self._category2 == self._data.category2Yes
end
 
function CategoryHandler:matchesBlacklist()
if self._usesCurrentTitle then
return self._data.currentTitleMatchesBlacklist
else
mShared = mShared or require('Module:Category handler/shared')
return mShared.matchesBlacklist(
self.title.prefixedText,
mw.loadData('Module:Category handler/blacklist')
)
end
end


----------------------------------------------------------------------------------------------------------
function CategoryHandler:isSuppressed()
--                                          Local functions                                            --
-- Find if categories are suppressed by either the arguments or by
--      The following are internal functions, which we do not want to be accessible from other modules. --
-- matching the blacklist.
----------------------------------------------------------------------------------------------------------
return self:isSuppressedByArguments()
or not self:shouldSkipBlacklistCheck() and self:matchesBlacklist()
end


-- Find whether we need to return a category or not.
function CategoryHandler:getNamespaceParameters()
local function needsCategory(pageObject, args)
if self._usesCurrentTitle then
    -- Don't categorise if the relevant options are set.
return self._data.currentTitleNamespaceParameters
    if yesno(args[cfg.nocat])
else
        or yesno(args[cfg.categories]) == false
if not mappings then
        or (
mShared = mShared or require('Module:Category handler/shared')
            args[cfg.category2]
mappings = mShared.getParamMappings(true) -- gets mappings with mw.loadData
            and args[cfg.category2] ~= cfg.category2Yes
end
            and args[cfg.category2] ~= cfg.category2Negative
return mShared.getNamespaceParameters(
        )
self.title,
    then
mappings
        return false
)
    end
end
    -- If there is no pageObject available, then that either means that we are over
    -- the expensive function limit or that the title specified was invalid. Invalid
    -- titles will probably only be a problem during testing, so we choose the best
    -- fallback for being over the expensive function limit. The fallback behaviour
    -- of the old template was to assume the page was not a subpage, so we will do
    -- the same here.
    if args[cfg.subpage] == cfg.subpageNo and pageObject and pageObject.isSubpage then
        return false
    end
    if args[cfg.subpage] == cfg.subpageOnly
        and (not pageObject or (pageObject and not pageObject.isSubpage))
    then
        return false
    end
    return true
end
end


-- Find whether we need to check the blacklist or not.
function CategoryHandler:namespaceParametersExist()
local function needsBlacklistCheck(args)
-- Find whether any namespace parameters have been specified.
    if yesno(args[cfg.nocat]) == false
-- We use the order "all" --> namespace params --> "other" as this is what
        or yesno(args[cfg.categories]) == true
-- the old template did.
        or args[cfg.category2] == cfg.category2Yes
if self:parameter('all') then
    then
return true
        return false
end
    else
if not mappings then
        return true
mShared = mShared or require('Module:Category handler/shared')
    end
mappings = mShared.getParamMappings(true) -- gets mappings with mw.loadData
end
for ns, params in pairs(mappings) do
for i, param in ipairs(params) do
if self._args[param] then
return true
end
end
end
if self:parameter('other') then
return true
end
return false
end
end


-- Find whether any namespace parameters have been specified.
function CategoryHandler:getCategories()
-- Mappings is the table of parameter mappings taken from
local params = self:getNamespaceParameters()
-- [[Module:Namespace detect]].
local nsCategory
local function nsParamsExist(mappings, args)
for i, param in ipairs(params) do
    if args[cfg.all] or args[cfg.other] then
local value = self._args[param]
        return true
if value ~= nil then
    end
nsCategory = value
    for ns, params in pairs(mappings) do
break
        for i, param in ipairs(params) do
end
            if args[param] then
end
                return true
if nsCategory ~= nil or self:namespaceParametersExist() then
            end
-- Namespace parameters exist - advanced usage.
        end
if nsCategory == nil then
    end
nsCategory = self:parameter('other')
    return false
end
local ret = {self:parameter('all')}
local numParam = tonumber(nsCategory)
if numParam and numParam >= 1 and math.floor(numParam) == numParam then
-- nsCategory is an integer
ret[#ret + 1] = self._args[numParam]
else
ret[#ret + 1] = nsCategory
end
if #ret < 1 then
return nil
else
return table.concat(ret)
end
elseif self._data.defaultNamespaces[self.title.namespace] then
-- Namespace parameters don't exist, simple usage.
return self._args[1]
end
return nil
end
end


----------------------------------------------------------------------------------------------------------
--------------------------------------------------------------------------------
--                                          Global functions                                            --
-- Exports
--      The following functions are global, because we want them to be accessible from #invoke and      --
--------------------------------------------------------------------------------
--     from other Lua modules.                                                                        --
----------------------------------------------------------------------------------------------------------


local p = {}
local p = {}


-- Find if a string matches the blacklist. Returns the match if one is found, or nil otherwise.
function p._exportClasses()
-- Input should be a page title with a namespace prefix, e.g. "Wikipedia talk:Articles for deletion".
-- Used for testing purposes.
function p.matchesBlacklist(page)
return {
    if type(page) ~= 'string' then return end
CategoryHandler = CategoryHandler
    for i, pattern in ipairs(cfg.blacklist) do
}
        local match = mw.ustring.match(page, pattern)
        if match then
            return match
        end
    end
end
end


-- The main structure of the module. Checks whether we need to categorise,
function p._main(args, data)
-- and then passes the relevant arguments to [[Module:Namespace detect]].
data = data or mw.loadData('Module:Category handler/data')
function p._main(args)
local handler = CategoryHandler.new(data, args)
    -- Get the page object and argument mappings from
if handler:isSuppressed() then
    -- [[Module:Namespace detect]], to save us from having to rewrite the
return nil
    -- code.
end
    local pageObject = nsDetect.getPageObject(args[cfg.page])
return handler:getCategories()
    local mappings = nsDetect.getParamMappings()
   
    if not needsCategory(pageObject, args) then return end
   
    local ret = ''
    -- Check blacklist if necessary.
    if not needsBlacklistCheck(args) or not p.matchesBlacklist(pageObject.prefixedText) then
        if not nsParamsExist(mappings, args) then
            -- No namespace parameters exist; basic usage. Pass args[1] to
            -- [[Module:Namespace detect]] using the default namespace
            -- parameters, and return the result.
            local ndargs = {}
            for _, ndarg in ipairs(cfg.defaultNamespaces) do
                ndargs[ndarg] = args[1]
            end
            ndargs.page = args.page
            ndargs.demospace = args.demospace
            local ndresult = nsDetect._main(ndargs)
            if ndresult then
                ret = ret .. ndresult
            end
        else
            -- Namespace parameters exist; advanced usage.
            -- If the all parameter is specified, return it.
            local all = args.all
            if type(all) == 'string' then
                ret = ret .. all
            end
           
            -- Get the arguments to pass to [[Module:Namespace detect]].
            local ndargs = {}
            for ns, params in pairs(mappings) do
                for _, param in ipairs(params) do
                    ndargs[param] = args[param] or args[cfg.other] or nil
                end
            end
            ndargs.other = args.other
            ndargs.page = args.page
            ndargs.demospace = args.demospace
           
            local data = nsDetect._main(ndargs)
           
            -- Work out what to return based on the result of the namespace detect call.
            local datanum = tonumber(data)
            if type(datanum) == 'number' then
                -- "data" is a number, so return that positional parameter.
                -- Remove non-positive integer values, as only positive integers
                -- from 1-10 were used with the old template.
                if datanum > 0 and math.floor(datanum) == datanum then
                    local dataArg = args[datanum]
                    if type(dataArg) == 'string' then
                        ret = ret .. dataArg
                    end
                end
            else
                -- "data" is not a number, so return it as it is.
                if type(data) == 'string' then
                    ret = ret .. data
                end
            end
        end
    end
    return ret
end
end


function p.main(frame)
function p.main(frame, data)
    -- If called via #invoke, use the args passed into the invoking
data = data or mw.loadData('Module:Category handler/data')
    -- template, or the args passed to #invoke if any exist. Otherwise
local args = require('Module:Arguments').getArgs(frame, {
    -- assume args are being passed directly in.
wrappers = data.wrappers,
    local origArgs
valueFunc = function (k, v)
    if frame == mw.getCurrentFrame() then
v = trimWhitespace(v)
        origArgs = frame:getParent().args
if type(k) == 'number' then
        for k, v in pairs(frame.args) do
if v ~= '' then
            origArgs = frame.args
return v
            break
else
        end
return nil
    else
end
        origArgs = frame
else
    end
return v
 
end
    -- Trim whitespace and remove blank arguments for the following args:
end
    -- 1, 2, 3 etc., "nocat", "categories", "subpage", and "page".
})
    local args = {}
return p._main(args, data)
    for k, v in pairs(origArgs) do
        if type(v) == 'string' then
            v = mw.text.trim(v) -- Trim whitespace.
        end
        if type(k) == 'number'
            or k == cfg.nocat
            or k == cfg.categories
            or k == cfg.subpage
            or k == cfg.page
        then
            if v ~= '' then
                args[k] = v
            end
        else
            args[k] = v
        end
    end
   
    -- Lower-case "nocat", "categories", "category2", and "subpage". These
    -- parameters are put in lower case whenever they appear in the old
    -- template, so we can just do it once here and save ourselves some work.
    local lowercase = {cfg.nocat, cfg.categories, cfg.category2, cfg.subpage}
    for _, v in ipairs(lowercase) do
        local argVal = args[v]
        if type(argVal) == 'string' then
            args[v] = mw.ustring.lower(argVal)
        end
    end
   
    return p._main(args)
end
end


return p
return p

Revision as of 05:08, 22 July 2014

This is the Category handler LUA module which is invoked in a number of MediaWiki-based WIKIs, including the venerable Wikipedia.

Dependencies

The following modules are required:


--------------------------------------------------------------------------------
--                                                                            --
--                              CATEGORY HANDLER                              --
--                                                                            --
--      This module implements the {{category handler}} template in Lua,      --
--      with a few improvements: all namespaces and all namespace aliases     --
--      are supported, and namespace names are detected automatically for     --
--      the local wiki. This module requires [[Module:Namespace detect]]      --
--      and [[Module:Yesno]] to be available on the local wiki. It can be     --
--      configured for different wikis by altering the values in              --
--      [[Module:Category handler/config]], and pages can be blacklisted      --
--      from categorisation by using [[Module:Category handler/blacklist]].   --
--                                                                            --
--------------------------------------------------------------------------------

-- Load required modules
local yesno = require('Module:Yesno')

-- Lazily load things we don't always need
local mShared, mappings

local p = {}

--------------------------------------------------------------------------------
-- Helper functions
--------------------------------------------------------------------------------

local function trimWhitespace(s, removeBlanks)
	if type(s) ~= 'string' then
		return s
	end
	s = s:match('^%s*(.-)%s*$')
	if removeBlanks then
		if s ~= '' then
			return s
		else
			return nil
		end
	else
		return s
	end
end

--------------------------------------------------------------------------------
-- CategoryHandler class
--------------------------------------------------------------------------------

local CategoryHandler = {}
CategoryHandler.__index = CategoryHandler

function CategoryHandler.new(data, args)
	local obj = setmetatable({ _data = data, _args = args }, CategoryHandler)
	
	-- Set the title object
	do
		local pagename = obj:parameter('demopage')
		local success, titleObj
		if pagename then
			success, titleObj = pcall(mw.title.new, pagename)
		end
		if success and titleObj then
			obj.title = titleObj
			if titleObj == mw.title.getCurrentTitle() then
				obj._usesCurrentTitle = true
			end
		else
			obj.title = mw.title.getCurrentTitle()
			obj._usesCurrentTitle = true
		end
	end

	-- Set suppression parameter values
	for _, key in ipairs{'nocat', 'categories'} do
		local value = obj:parameter(key)
		value = trimWhitespace(value, true)
		obj['_' .. key] = yesno(value)
	end
	do
		local subpage = obj:parameter('subpage')
		local category2 = obj:parameter('category2')
		if type(subpage) == 'string' then
			subpage = mw.ustring.lower(subpage)
		end
		if type(category2) == 'string' then
			subpage = mw.ustring.lower(category2)
		end
		obj._subpage = trimWhitespace(subpage, true)
		obj._category2 = trimWhitespace(category2) -- don't remove blank values
	end
	return obj
end

function CategoryHandler:parameter(key)
	local parameterNames = self._data.parameters[key]
	local pntype = type(parameterNames)
	if pntype == 'string' or pntype == 'number' then
		return self._args[parameterNames]
	elseif pntype == 'table' then
		for _, name in ipairs(parameterNames) do
			local value = self._args[name]
			if value ~= nil then
				return value
			end
		end
		return nil
	else
		error(string.format(
			'invalid config key "%s"',
			tostring(key)
		), 2)
	end
end

function CategoryHandler:isSuppressedByArguments()
	return
		-- See if a category suppression argument has been set.
		self._nocat == true
		or self._categories == false
		or (
			self._category2
			and self._category2 ~= self._data.category2Yes
			and self._category2 ~= self._data.category2Negative
		)

		-- Check whether we are on a subpage, and see if categories are
		-- suppressed based on our subpage status.
		or self._subpage == self._data.subpageNo and self.title.isSubpage
		or self._subpage == self._data.subpageOnly and not self.title.isSubpage
end

function CategoryHandler:shouldSkipBlacklistCheck()
	-- Check whether the category suppression arguments indicate we
	-- should skip the blacklist check.
	return self._nocat == false
		or self._categories == true
		or self._category2 == self._data.category2Yes
end

function CategoryHandler:matchesBlacklist()
	if self._usesCurrentTitle then
		return self._data.currentTitleMatchesBlacklist
	else
		mShared = mShared or require('Module:Category handler/shared')
		return mShared.matchesBlacklist(
			self.title.prefixedText,
			mw.loadData('Module:Category handler/blacklist')
		)
	end
end

function CategoryHandler:isSuppressed()
	-- Find if categories are suppressed by either the arguments or by
	-- matching the blacklist.
	return self:isSuppressedByArguments()
		or not self:shouldSkipBlacklistCheck() and self:matchesBlacklist()
end

function CategoryHandler:getNamespaceParameters()
	if self._usesCurrentTitle then
		return self._data.currentTitleNamespaceParameters
	else
		if not mappings then
			mShared = mShared or require('Module:Category handler/shared')
			mappings = mShared.getParamMappings(true) -- gets mappings with mw.loadData
		end
		return mShared.getNamespaceParameters(
			self.title,
			mappings
		)
	end
end

function CategoryHandler:namespaceParametersExist()
	-- Find whether any namespace parameters have been specified.
	-- We use the order "all" --> namespace params --> "other" as this is what
	-- the old template did.
	if self:parameter('all') then
		return true
	end
	if not mappings then
		mShared = mShared or require('Module:Category handler/shared')
		mappings = mShared.getParamMappings(true) -- gets mappings with mw.loadData
	end
	for ns, params in pairs(mappings) do
		for i, param in ipairs(params) do
			if self._args[param] then
				return true
			end
		end
	end
	if self:parameter('other') then
		return true
	end
	return false
end

function CategoryHandler:getCategories()
	local params = self:getNamespaceParameters()
	local nsCategory
	for i, param in ipairs(params) do
		local value = self._args[param]
		if value ~= nil then
			nsCategory = value
			break
		end
	end
	if nsCategory ~= nil or self:namespaceParametersExist() then
		-- Namespace parameters exist - advanced usage.
		if nsCategory == nil then
			nsCategory = self:parameter('other')
		end
		local ret = {self:parameter('all')}
		local numParam = tonumber(nsCategory)
		if numParam and numParam >= 1 and math.floor(numParam) == numParam then
			-- nsCategory is an integer
			ret[#ret + 1] = self._args[numParam]
		else
			ret[#ret + 1] = nsCategory
		end
		if #ret < 1 then
			return nil
		else
			return table.concat(ret)
		end
	elseif self._data.defaultNamespaces[self.title.namespace] then
		-- Namespace parameters don't exist, simple usage.
		return self._args[1]
	end
	return nil
end

--------------------------------------------------------------------------------
-- Exports
--------------------------------------------------------------------------------

local p = {}

function p._exportClasses()
	-- Used for testing purposes.
	return {
		CategoryHandler = CategoryHandler
	}
end

function p._main(args, data)
	data = data or mw.loadData('Module:Category handler/data')
	local handler = CategoryHandler.new(data, args)
	if handler:isSuppressed() then
		return nil
	end
	return handler:getCategories()
end

function p.main(frame, data)
	data = data or mw.loadData('Module:Category handler/data')
	local args = require('Module:Arguments').getArgs(frame, {
		wrappers = data.wrappers,
		valueFunc = function (k, v)
			v = trimWhitespace(v)
			if type(k) == 'number' then
				if v ~= '' then
					return v
				else
					return nil
				end
			else
				return v
			end
		end
	})
	return p._main(args, data)
end

return p