Fra Wikipedia, den frie encyklopædi
Spring til navigation Spring til søgning
Documentation icon Moduldokumentation[opret]
local p = {}
local mRedirect = require('Module:Redirect')

local errors
-- Return blank text, or an error message if requested
local function err(text)
	if errors then error(text, 2) end
	return ""

-- In text, match pre..list[1] or pre..list[2] or ...
local function matchany(text, pre, list, post)
	local match
	for i = 1, #list do
		match = mw.ustring.match(text, pre .. list[i] .. post)
		if match then return match end
	return nil

-- Get a page's content, following redirects, and processing file description pages for files.
-- Also returns the page name, or the target page name if a redirect was followed, or false if t
local function getContent(page, frame)
	local title = -- Read description page (for :File:Foo rather than File:Foo)
	if not title then return false, false end

	local redir = mRedirect.getTarget(title)
	if redir then title = end

	if title.namespace == 6 then
		frame = frame or mw.getCurrentFrame()
		return frame:preprocess("{{" .. title.prefixedText .. "}}"), redir or title.prefixedText
		return title:getContent(), redir or title.prefixedText

-- Check image for suitability
local function checkimage(image)
	local page = matchany(image, "", {"[Ff]il", "[Bb]illede", "[Ff]ile", "[Ii]mage"}, "%s*:[^|%]]*") -- match File:(name) or Image:(name)
	if not page then return false end

	-- Limit to image types: .gif, .jpg, .jpeg, .png, .svg, .tiff, .xcf (exclude .ogg audio etc.)
	if not matchany(page, "%.", {"[Gg][Ii][Ff]", "[Jj][Pp][Ee]?[Gg]", "[Pp][Nn][Gg]", "[Ss][Vv][Gg]", "[Tt][Ii][Ff][Ff]", "[Xx][Cc][Ff]"}, "%s*$") then
		return false

	local desc = getContent(page)
	return ( desc and desc ~= "" and not mw.ustring.match(desc, "[Nn]on%-free") ) and true or false -- hide non-free image

-- Attempt to parse [[File:...]] or [[Image:...]], either anywhere (start=false) or at the start only (start=true)
local function parseimage(text, start)
	local startre = ""
	if start then startre = "^" end -- a true flag restricts search to start of string
	local image = matchany(text, startre .. "%[%[%s*", {"[Ff]il", "[Bb]illede", "[Ff]ile", "[Ii]mage"}, "%s*:.*") -- [[File: or [[Image: ...
	if image then
		image = mw.ustring.match(image, "%b[]%s*") -- matching [[...]] to handle wikilinks nested in caption
	return image

-- Parse a caption, which ends at a | (end of parameter) or } (end of infobox) but may contain nested [..] and {..}
local function parsecaption(caption)
	if not caption then return nil end
	local len = mw.ustring.len(caption)
	local pos = 1
	while pos <= len do
		local linkstart, linkend = mw.ustring.find(caption, "%b[]", pos)
		linkstart = linkstart or len + 1 -- avoid comparison with nil when no link
		local templatestart, templateend = mw.ustring.find(caption, "%b{}", pos)
		templatestart = templatestart or len + 1 -- avoid comparison with nil when no template
		local argend = mw.ustring.find(caption, "[|}]", pos) or len + 1
		if linkstart < templatestart and linkstart < argend then
			pos = linkend + 1 -- skip wikilink
		elseif templatestart < argend then
			pos = templateend + 1 -- skip template
		else -- argument ends before the next wikilink or template
			return mw.ustring.sub(caption, 1, argend - 1)

-- Attempt to construct a [[File:...]] block from {{infobox ... |image= ...}}
local function argimage(text)
	local token = nil
	if mw.ustring.match(text, "{{%s*[Ii]nfoboks") then
		local image = matchany(text, "|%s*", {"billede_fil", "billede_navn", "billede_flag", "billede", "image", "PD_image", "image_flag", "Ship image", "Cover", "static_image_name", "static_image", "image_skyline", "image_shield", "static_image2_name", "static_image2"}, "%s*=%s*(.*)")

		if image then -- add in relevant optional parameters: caption, alt text and image size
			token = parseimage(image, true) -- look for image=[[File:...]] etc.
			if not token then
				image = mw.ustring.match(image, "^[^}|]*") -- remove later arguments
				token = "[[" -- Add File: unless name already begins File: or Image:
				if not matchany(image, "^", {"[Ff]il", "[Bb]illede", "[Ff]ile", "[Ii]mage"}, "%s*:") then
					token = token .. "File:"
				token = token .. image
				local caption = parsecaption(matchany(text, "|%s*", {"[Cc]aption", "Ship caption"}, "%s*=%s*(.*)"))
				if caption and mw.ustring.match(caption, "%S") then token = token .. "|" .. caption end
				local alt = mw.ustring.match(text, "|%s*alt%s*=%s*([^}|]*)")
				if alt then token = token .. "|alt=" .. alt end
				local image_size = mw.ustring.match(text, "|%s*image_size%s*=%s*([^}|]*)")
				if image_size and mw.ustring.match(image_size, "%S") then token = token .. "|" .. image_size end
				token = token .. "]]"
			token = mw.ustring.gsub(token, "\n","") .. "\n"

	return token

-- Help gsub to remove unwanted templates
local function striptemplate(t)
	-- If template is unwanted then return "" (gsub will replace by nothing), else return nil (gsub will keep existing string)
	local unwanted = {"[Kk]ilde mangler", "[Kk]m", "[Ee]fn", "[Ee]fn%-[lu]a", "[Ee]l[mn]", "[Rr]p?", "[Ss]fn[bmp]", "[Ss]f[bn]", "NoteTag", "#[Tt]ag:%s*[Rr]ef", "[Rr]efn?",
	 "[CcDd]n", "[Cc]itation needed", "[Dd]isambiguation needed", "[Ff]eatured article", "[Gg]ood article",
	 -- aliases for Clarification needed
	 "[Cc]f[ny]", "[Cc]larification[ _]+inline", "[Cc]larification[%- _]*needed", "[Cc]larification", "[Cc]larify%-inline", "[Cc]larify%-?me",
	 "[Cc]larify[ _]+inline", "[Cc]larify", "[Cc]LARIFY", "[Cc]onfusing%-inline", "[Cc]onfusing%-short", "[Ee]xplainme", "[Hh]uh[ _]*%??", "[Ww]hat%?",
	 "[Ii]nline[ _]+[Uu]nclear", "[Ii]n[ _]+what[ _]+sense", "[Oo]bscure", "[Pp]lease[ _]+clarify", "[Uu]nclear[ _]+inline", "[Ww]hat's[ _]+this%?",
	 -- aliases for Primary source inline
	 "[Pp]s[ci]", "[Nn]psn", "[Nn]on%-primary[ _]+source[ _]+needed", "[Ss]econdary[ _]+source[ _]+needed",
	 "[Pp]rimary[ _]+source[ _]+claim", "[Pp]rimary[%- _]+source[%- _]+inline", "[Pp]rimary[%- _]+inline",
	 -- aliases for Disambiguation (page) and similar
	 "[Bb]egriffsklärung", "[Dd][Aa][Bb]", "[Dd]big", "[%w%s]-%f[%w][Dd]isam[%w%s]-", "[Hh][Nn][Dd][Ii][Ss]"}

	if matchany(t, "^{{%s*", unwanted, "%s*%f[|}]") then return "" end

	-- If template is wanted but produces an unwanted reference then return the string with |shortref or |ref removed
	local noref = mw.ustring.gsub(t, "|%s*shortref%s*%f[|}]", "")
	noref = mw.ustring.gsub(noref, "|%s*ref%s*%f[|}]", "")
	if noref ~= t then return noref end

	return nil -- not an unwanted template: keep

-- Convert a comma-separated list of numbers or min-max ranges into a list of booleans, e.g. "1,3-5" → {1=true,2=false,3=true,4=true,5=true}
local function numberflags(str)
	local ranges = mw.text.split(str, ",") -- parse ranges, e.g. "1,3-5" → {"1","3-5"}
	local flags = {}
	for _, r in pairs(ranges) do
		local min, max = mw.ustring.match(r, "^%s*(%d+)%s*%-%s*(%d+)%s*$") -- "3-5" → min=3 max=5
		if not max then	min, max = mw.ustring.match(r, "^%s*((%d+))%s*$") end -- "1" → min=1 max=1
		if max then
			for p = min, max do flags[p] = true end
	return flags

-- a basic parser to trim down extracted wikitext
--   @param text : Wikitext to be processed
--   @param options : A table of options...
--          options.paraflags : Which number paragraphs to keep, as either a string (e.g. `1,3-5`) or a table (e.g. `{1=true,2=false,3=true,4=true,5=true}`. If not present, all paragraphs will be kept.
--          options.fileflags : table of which files to keep, as either a string (e.g. `1,3-5`) or a table (e.g. `{1=true,2=false,3=true,4=true,5=true}`
--          options.fileargs : args for the [[File:]] syntax, such as `left`
--   @param filesOnly : If set, only return the files and not the prose
local function parse(text, options, filesOnly)
	local allparas = true -- keep all paragraphs?
	if options.paraflags then
		if type(options.paraflags) ~= "table" then options.paraflags = numberflags(options.paraflags) end
		for _, v in pairs(options.paraflags) do
			if v then allparas = false end -- if any para specifically requested, don't keep all
	if filesOnly then
		allparas = false
		options.paraflags = {}

	local maxfile = 0 -- for efficiency, stop checking images after this many have been found
	if options.fileflags then
		if type(options.fileflags) ~= "table" then options.fileflags = numberflags(options.fileflags) end
		for k, v in pairs(options.fileflags) do
			if v and k > maxfile then maxfile = k end -- set maxfile = highest key in fileflags

	local leadstart = nil -- have we found some text yet?
	local t = "" -- the stripped down output text
	local files = 0 -- how many images so far
	local paras = 0 -- how many paragraphs so far

	text = mw.ustring.gsub(text,"^%s*","") -- remove initial white space
	repeat -- loop around parsing a template, image or paragraph
		local token = mw.ustring.match(text, "^%b{}%s*") or false -- {{Template}}
		local line = mw.ustring.match(text, "[^\n]*")
		if token and line and mw.ustring.len(token) < mw.ustring.len(line) then -- template is followed by text (but it may just be other templates)
			line = mw.ustring.gsub(line, "%b{}", "") -- remove all templates from this line
			-- if anything is left, other than an incomplete further template or an image, keep the template: it counts as part of the line
			if mw.ustring.find(line, "%S") and not matchany(line, "^%s*", { "{{", "%[%[%s*[Ff]ile:", "%[%[%s*[Ii]mage:" }, "") then
				token = nil

		if token then -- found a template which is not the prefix to a line of text
			if leadstart then -- lead has already started, so keep the template within the text
				if not filesOnly then t = t .. token end
			elseif files < maxfile then -- discard template, but if we are still collecting images...
				local image = argimage(token) or parseimage(token, false) -- look for embedded [[File:...]], |image=, etc.
				if image and checkimage(image) then -- if image is found and qualifies (not a sound file, non-free, etc.)

					files = files + 1 -- count the file, whether displaying it or not
					if options.fileflags and options.fileflags[files] then -- if displaying this image
						image = mw.ustring.gsub(image, "|%s*frameless%s*%f[|%]]", "") -- make image a thumbnail, not frameless etc.
						image = mw.ustring.gsub(image, "|%s*framed?%s*%f[|%]]", "")
						if not matchany(image, "|%s*", {"thumb", "thumbnail"}, "%s*%f[|%]]") then
							image = mw.ustring.gsub(image, "(%]%]%s*)$", "|thumb%1")
						if options.fileargs then image = mw.ustring.gsub(image, "(%]%]%s*)$", "|" .. options.fileargs .. "%1") end
						t = t .. image
		else -- the next token in text is not a template
			token = parseimage(text, true)
			if token then -- the next token in text looks like an image
				if files < maxfile and checkimage(token) then -- if more images are wanted and this is a wanted image
					files = files + 1
					if options.fileflags and options.fileflags[files] then
						local image = token -- copy token for manipulation by adding |right etc. without changing the original
						if options.fileargs then image = mw.ustring.gsub(image, "(%]%]%s*)$", "|" .. options.fileargs .. "%1") end
						t = t .. image
			else -- got a paragraph, which ends at a file, image, blank line or end of text
				local afterend = mw.ustring.len(text) + 1
				local blankpos = mw.ustring.find(text, "\n%s*\n") or afterend -- position of next paragraph delimiter (or end of text)
				local endpos = math.min( -- find position of whichever comes first: [[File:, [[Image: or paragraph delimiter
				 mw.ustring.find(text, "%[%[%s*[Ff]ile%s*:") or afterend,
				 mw.ustring.find(text, "%[%[%s*[Ii]mage%s*:") or afterend,
				 mw.ustring.find(text, "%[%[%s*[Ff]il%s*:") or afterend,
				 mw.ustring.find(text, "%[%[%s*[Bb]illede%s*:") or afterend,
				token = mw.ustring.sub(text, 1, endpos-1)
				if blankpos < afterend and blankpos == endpos then -- paragraph ends with a blank line
					token = token .. mw.ustring.match(text, "\n%s*\n", blankpos)
				leadstart = leadstart or mw.ustring.len(t) + 1 -- we got a paragraph, so mark the start of the lead section
				paras = paras + 1
				if allparas or (options.paraflags and options.paraflags[paras]) then t = t .. token end -- add if this paragraph wanted
			end -- of "else got a paragraph"
		end -- of "else not a template"

		if token then text = mw.ustring.sub(text, mw.ustring.len(token)+1) end -- remove parsed token from remaining text
	until not text or text == "" or not token or token == "" -- loop until all text parsed

	text = mw.ustring.gsub(t, "\n+$", "") -- remove trailing line feeds, so "{{Transclude text excerpt|Foo}} more" flows on one line
	return text

local function cleanupText(text, leadOnly)
	text = mw.ustring.gsub(text, "<!%-%-.-%-%->","") -- remove HTML comments
	if leadOnly then
		text = mw.ustring.gsub(text, "%c%s*==.*","") -- remove first ==Heading== and everything after it
	text = mw.ustring.gsub(text, "<noinclude>.-</noinclude>", "") -- remove noinclude bits
	text = mw.ustring.gsub(text, "<%s*ref[^>]-/%s*>", "") -- remove refs cited elsewhere
	text = mw.ustring.gsub(text, "<%s*ref.->.-<%s*/%s*ref%s*>", "") -- remove refs
	text = mw.ustring.gsub(text, "<%s*imagemap.->.-<%s*/%s*imagemap%s*>", "") -- remove imagemaps
	text = mw.ustring.gsub(text, "%b{}", striptemplate) -- remove unwanted templates such as references
	text = mw.ustring.gsub(text, "\n%s*{{%s*[Tt][Oo][Cc].-}}", "\n") -- remove most common tables of contents
	text = mw.ustring.gsub(text, "\n%s*{{%s*[Pp]p%-.-}}", "\n") -- remove protection templates
	text = mw.ustring.gsub(text, "%s*{{[^{|}]*sidebar%s*}}", "") -- remove most sidebars
	text = mw.ustring.gsub(text, "%s*{{[^{|}]*%-stub%s*}}", "") -- remove most stub templates
	text = mw.ustring.gsub(text, "%s*%[%[%s*:?[Cc]ategory:.-%]%]", "") -- remove categories
        text = mw.ustring.gsub(text, "%s*%[%[%s*:?[Kk]ategori:.-%]%]", "") -- slet kategorier
	return text

-- Main function returns a string value: text of the lead of a page
local function main(pagenames, options)
	if not pagenames or #pagenames < 1 then return err("No page names given") end
	local pagename
	local text
	local pagecount = #pagenames
	local firstpage = pagenames[1] or "(nil)" -- save for error message, as it the name will be deleted

	-- read the page, or a random one if multiple pages were provided
	if pagecount > 1 then math.randomseed(os.time()) end
	while not text and pagecount > 0 do
		local pagenum = 1
		if pagecount > 1 then pagenum = math.random(pagecount) end -- pick a random title
		pagename = pagenames[pagenum]
		if pagename and pagename ~= "" then
			pagename = mw.ustring.match(pagename, "%[%[%s*(.-)[]|#]") or pagename -- "[[Foo|Bar]]" → "Foo"
			pagename = mw.ustring.gsub(pagename, "^%s+", "") -- strip leading ...
			pagename = mw.ustring.gsub(pagename, "%s+$", "") -- ...and trailing white space

			if pagename and pagename ~= "" then
				text, noramliesdPagename = getContent(pagename)
				if not noramliesdPagename then
					return err("No title for page name " .. pagename)
					pagename = noramliesdPagename
		if not text then table.remove(pagenames, pagenum) end -- this one didn't work; try another
		pagecount = pagecount - 1 -- ensure that we exit the loop after at most #pagenames iterations
	if not text then return err("Cannot read a valid page: first name is " .. firstpage) end

	text = cleanupText(text, true)
	text = parse(text, options)

	-- replace the bold title or synonym near the start of the article by a wikilink to the article
	local lang = mw.language.getContentLanguage()
	local pos = mw.ustring.find(text, "'''" .. lang:ucfirst(pagename) .. "'''", 1, true) -- look for "'''Foo''' is..." (uc) or "A '''foo''' is..." (lc)
	 or mw.ustring.find(text, "'''" .. lang:lcfirst(pagename) .. "'''", 1, true) -- plain search: special characters in pagename represent themselves
	if pos then
		local len = mw.ustring.len(pagename)
		text = mw.ustring.sub(text, 1, pos + 2) .. "[[" .. mw.ustring.sub(text, pos + 3, pos + len + 2) .. "]]" .. mw.ustring.sub(text, pos + len + 3, -1) -- link it
	else -- look for anything unlinked in bold, assumed to be a synonym of the title (e.g. a person's birth name)
		text = mw.ustring.gsub(text, "(.-''')(.-'*)'''", function(a, b)
			if mw.ustring.len(a) < 100 + (leadstart or 0) and not mw.ustring.find(b, "%[") then ---if early in article and not wikilinked
				return a .. "[[" .. pagename .. "|" .. b .. "]]'''" -- replace '''Foo''' by '''[[pagename|Foo]]
				return nil -- instruct gsub to make no change
		 end, 1) -- "end" here terminates the anonymous replacement function(a, b) passed to gsub

	if options.more then text = text .. " '''[[" .. pagename .. "|" .. options.more .. "]]'''" end -- wikilink to article for more info
	return text

-- Parse a ==Section== from a page
local function getsection(text, section)
	local level, content = mw.ustring.match(text .. "\n", "\n(==+)%s*" .. section .. "%s*==.-\n(.*)")
	if not content then return nil end -- no such section
	local nextsection = "\n==" .. mw.ustring.rep("=?", #level - 2) .. "[^=].*" -- "===" → "\n===?[^=].*", matching "==" or "===" but not "===="
	content = mw.ustring.gsub(content, nextsection, "") -- remove later sections with headings at this level or higher
	return content

-- Shared template invocation code for lead and random functions
local function invoke(frame, func)
	-- args = { 1,2,... = page names, paragraphs = list e.g. "1,3-5", files = list, more = text}
	local args = {} -- args[k] = frame.args[k] or frame:getParent().args[k] for all k in either (numeric or not)
	for k, v in pairs(frame:getParent().args) do args[k] = v end
	for k, v in pairs(frame.args) do args[k] = v end -- args from a Lua call have priority over parent args from template
	errors = args["errors"] -- set the module level boolean used in local function err

	local articlecount = #args -- must be 1 except with selected=Foo and Foo=Somepage
	if articlecount < 1 and not (func == "selected" and args[func] and args[args[func]]) then
		return err("No articles provided")

	local pagenames = {}
	if func == "lead" then
		pagenames = { args[1] }
	elseif func == "linked" or func == "listitem" then
		-- Read named page and find its wikilinks
		local page = args[1]
		local text, title = getContent(page)
		if not title then
			return err("Ingen titel for siden " .. page)
		elseif not text then
			return err("Intet indhold for siden " .. page)
		if args["section"] or args["sektion"] then -- check relevant section only
			local argsektion = args["section"] or args["sektion"]
			text = getsection(text, argsektion)
			if not text then return err("Afnittet " .. argsektion .. " findes ikke på siden " .. page) end
		if func == "linked" then
			for p in mw.ustring.gmatch(text, "%[%[%s*([^%]|#\n]*)") do table.insert(pagenames, p) end
		else -- listitem: first wikilink on a line beginning *, :#, etc. except in "See also" or later section
			text = mw.ustring.gsub(text, "\n== *Se også.*", "")
			for p in mw.ustring.gmatch(text, "\n:*[%*#][^\n]-%[%[%s*([^%]|#\n]*)") do table.insert(pagenames, p) end
	elseif func == "random" then
		-- accept any number of page names.  If more than one, we'll pick one randomly
		for i, p in pairs(args) do
			if p and type(i) == 'number' then table.insert(pagenames, p) end
	elseif func == "selected" then
		local articlekey = args[func] or args["valgt"] -- dansk parameter hedder "valgt"
		if tonumber(articlekey) then -- normalise article number into the range 1..#args
			articlekey = articlekey % articlecount
			if articlekey == 0 then articlekey = articlecount end
		pagenames = { args[articlekey] }

	local options = args -- pick up miscellaneous options: more, errors, fileargs
	options.paraflags = numberflags(args["paragraphs"] or args["afsnit"] or "") -- parse paragraphs, e.g. "1,3-5" → {"1","3-5"}
	options.fileflags = numberflags(args["files"] or args["filer"] or "") -- parse file numbers
        if options.mere and options.mere == "" then options.mere = "Læs mere..." end
	if options.more and options.more == "" then options.more = "Read more..." end -- more= is short for this default text
        if options.mere then options.more = options.mere end -- oversætter parameternavn
        if options.filarg then options.fileargs = options.filarg end -- oversætter parameternavn
        if options.fejl then options.errors = options.fejl end -- oversætter parameternavn
        if options.valgt then options.selected = options.valgt end -- oversætter parameternavn

	local text = main(pagenames, options)
	return frame:preprocess(text)

-- Entry points for template callers using #invoke:
function p.lead(frame) return invoke(frame, "lead") end -- {{Transclude lead excerpt}} reads the first and only article
function p.linked(frame) return invoke(frame, "linked") end -- {{Transclude linked excerpt}} reads a randomly selected article linked from the given page
function p.listitem(frame) return invoke(frame, "listitem") end -- {{Transclude list item excerpt}} reads a randomly selected article listed on the given page
function p.random(frame) return invoke(frame, "random") end -- {{Transclude random excerpt}} reads any article (default for invoke with one argument)
function p.selected(frame) return invoke(frame, "selected") end -- {{Transclude selected excerpt}} reads the article whose key is in the selected= parameter

-- Entry points for other Lua modules
function p.getContent(page, frame) return getContent(page, frame) end
function p.getsection(text, section) return getsection(text, section) end
function p.parse(text, options, filesOnly) return parse(text, options, filesOnly) end
function p.argimage(text) return argimage(text) end
function p.checkimage(image) return checkimage(image) end
function p.parseimage(text, start) return parseimage(text, start) end
function p.cleanupText(text, leadOnly) return cleanupText(text, leadOnly) end

return p