|
-- Module:Excerpt implements the Excerpt template
-- Get localized data
-- Documentation and master version: https://en.wikipedia.org/wiki/Module:Excerpt
local d = require("Module:Excerpt/i18n")
-- Authors: User:Sophivorus, User:Certes, User:Aidan9382 & others
-- License: CC-BY-SA-3.0
local Transcluder = require( 'Module:Transcluder' )
local p = {}
local yesno = require( 'Module:Yesno' )
-- Helper function to test for truthy and falsy values
local function is(value)
if not value or value == "" or value == "0" or value == "false" or value == "no" then
return false
end
return true
end
local ok, config = pcall( require, 'Module:Excerpt/config' )
-- Error handling function
if not ok then config = {} end
-- Throws a Lua error or returns an empty string if error reporting is disabled
errors = true -- show errors by default
local function luaError(message, value)
if not is(errors) then return "" end -- error reporting is disabled
message = d.errors[message] or message or ""
message = mw.ustring.format(message, value)
error(message, 2)
end
local p = {}
-- Error handling function
-- Returns a wiki friendly error or an empty string if error reporting is disabled
local function wikiError(message, value)
if not is(errors) then return "" end -- error reporting is disabled
message = d.errors[message] or message or ""
message = mw.ustring.format(message, value)
message = d.errors.prefix .. message
if mw.title.getCurrentTitle().isContentPage then
local errorsCategory = mw.title.new(d.errorsCategory, 'Category')
if errorsCategory then message = message .. '[[' .. errorsCategory.prefixedText .. ']]' end
end
message = mw.html.create('div'):addClass('error'):wikitext(message)
return message
end
-- Helper function to matchget from a list regular expressionsarguments
local args
-- Like so: match pre..list[1]..post or pre..list[2]..post or ...
local function matchAnygetArg(text, prekey, list, post,default init)
local matchvalue = {}args[ key ]
if value and mw.text.trim( value ) ~= '' then
for i = 1, #list do
return value
match = { mw.ustring.match(text, pre .. list[i] .. post, init) }
if match[1] then return unpack(match) end
end
return nildefault
end
-- Helper function to converthandle imagemaps into standard imageserrors
local function convertImageMapgetError(imagemap message, value )
if type( message ) == 'string' then
local image = matchAny(imagemap, "[>\n]%s*", d.fileNamespaces, "[^\n]*")
message = Transcluder.getError( message, value )
if image then
return "<!--imagemap-->[[" .. mw.ustring.gsub(image, "[>\n]%s*", "", 1) .. "]]"
else
return "" -- remove entire block if image can't be extracted
end
if config.categories and config.categories.errors and mw.title.getCurrentTitle().isContentPage then
end
message:node( '[[Category:' .. config.categories.errors .. ']]' )
-- Helper function to convert a comma-separated list of numbers or min-max ranges into a list of booleans
-- For example: "1,3-5" to {1=true,2=false,3=true,4=true,5=true}
local function numberFlags(str)
if not str then return {} end
local flags = {}
local ranges = mw.text.split(str, ",") -- parse ranges: "1,3-5" to {"1","3-5"}
for _, r in pairs(ranges) do
local min, max = mw.ustring.match(r, "^%s*(%d+)%s*%-%s*(%d+)%s*$") -- "3-5" to min=3 max=5
if not max then min, max = mw.ustring.match(r, "^%s*((%d+))%s*$") end -- "1" to min=1 max=1
if max then
for p = min, max do flags[p] = true end
end
end
return flagsmessage
end
-- Helper function to convertget templatelocalized arguments into an array of arguments fit for get()messages
local function parseArgsgetMessage(frame key )
local argsok, TNT = {}pcall( require, 'Module:TNT' )
if not ok then return key end
for key, value in pairs(frame:getParent().args) do args[key] = value end
return TNT.format( 'I18n/Module:Excerpt.tab', key )
for key, value in pairs(frame.args) do args[key] = value end -- args from a Lua call have priority over parent args from template
args.paraflags = numberFlags(args["paragraphs"] or "") -- parse paragraphs: "1,3-5" to {"1","3-5"}
args.fileflags = numberFlags(args["files"] or "") -- parse file numbers
return args
end
-- Main entry point for templates
-- Helper function to remove unwanted templates and pseudo-templates such as #tag:ref and DEFAULTSORT
local function stripTemplatep.main(t frame )
args = Transcluder.parseArgs( frame )
-- If template is unwanted then return "" (gsub will replace by nothing), else return nil (gsub will keep existing string)
if matchAny(t, "^{{%s*", d.unwantedInlineTemplates, "%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 a wanted template has unwanted nested templates, purge them too
noRef = mw.ustring.sub(noRef, 1, 2) .. mw.ustring.gsub(mw.ustring.sub(noRef, 3), "%b{}", stripTemplate)
-- Replace {{audio}} by its text parameter: {{Audio|Foo.ogg|Bar}} → Bar
noRef = mw.ustring.gsub(noRef, "^{{%s*[Aa]udio.-|.-|(.-)%f[|}].*", "%1")
-- Replace {{Nihongo foot}} by its text parameter: {{Nihongo foot|English|英語|eigo}} → English
noRef = mw.ustring.gsub(noRef, "^{{%s*[Nn]ihongo[ _]+foot%s*|(.-)%f[|}].*", "%1")
if noRef ~= t then return noRef end
return nil -- not an unwanted template: keep
end
-- Make sure the requested page exists
-- Get a page's content, following redirects
local page = getArg( 1 )
-- Also returns the page name, or the target page name if a redirect was followed, or false if no page found
if not page or page == '{{{1}}}' then return getError( 'no-page' ) end
-- For file pages, returns the content of the file description page
local function getContent(page)
local title = mw.title.new(page)
if not title then return falsegetError( 'invalid-title', falsepage ) end
if title.isRedirect then title = title.redirectTarget end
if not title.exists then return getError( 'page-not-found', page ) end
page = title.prefixedText
-- Set variables from the template parameters
local target = title.redirectTarget
local section = getArg( 2, mw.ustring.match( getArg( 1 ), '[^#]+#(.+)' ) )
if target then title = target end
local hat = yesno( getArg( 'hat', true ) )
local edit = yesno( getArg( 'edit', true ) )
local this = getArg( 'this' )
local only = getArg( 'only' )
local files = getArg( 'files', getArg( 'file', ( only == 'file' and 1 ) ) )
local lists = getArg( 'lists', getArg( 'list', ( only == 'list' and 1 ) ) )
local tables = getArg( 'tables', getArg( 'table', ( only == 'table' and 1 ) ) )
local templates = getArg( 'templates', getArg( 'template', ( only == 'template' and 1 ) ) )
local paragraphs = getArg( 'paragraphs', getArg( 'paragraph', ( only == 'paragraph' and 1 ) ) )
local references = getArg( 'references' )
local subsections = not yesno( getArg( 'subsections' ) )
local noLinks = not yesno( getArg( 'links', true ) )
local noBold = not yesno( getArg( 'bold' ) )
local onlyFreeFiles = yesno( getArg( 'onlyfreefiles', true ) )
local briefDates = yesno( getArg( 'briefdates', false ) )
local inline = yesno( getArg( 'inline' ) )
local quote = yesno( getArg( 'quote' ) )
local more = yesno( getArg( 'more' ) )
local class = getArg( 'class' )
local displaytitle = getArg( 'displaytitle' ) or page
-- Build the hatnote
return title:getContent(), title.prefixedText
if hat and not inline then
end
if this then
hat = this
-- Get the tables only
elseif quote then
local function getTables(text, options)
hat = getMessage( 'this' )
local tables = {}
elseif only then
for candidate in mw.ustring.gmatch(text, "%b{}") do
hat = getMessage( only )
if mw.ustring.sub(candidate, 1, 2) == '{|' then
table.insert(tables, candidate)
end
end
return table.concat(tables, '\n')
end
-- Get the lists only
local function getLists(text, options)
local lists = {}
for list in mw.ustring.gmatch(text, "\n[*#][^\n]+") do
table.insert(lists, list)
end
return table.concat(lists, '\n')
end
-- Check image for suitability
local function checkImage(image)
local page = matchAny(image, "", d.fileNamespaces, "%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.)
local fileTypes = {"[Gg][Ii][Ff]", "[Jj][Pp][Ee]?[Gg]", "[Pp][Nn][Gg]", "[Ss][Vv][Gg]", "[Tt][Ii][Ff][Ff]", "[Xx][Cc][Ff]"}
if not matchAny(page, "%.", fileTypes, "%s*$") then return false end
-- Check the local wiki
local fileDescription, fileTitle = getContent(page) -- get file description and title after following any redirect
if not fileTitle or fileTitle == "" then return false end -- the image doesn't exist
-- Check Commons
if not fileDescription or fileDescription == "" then
local frame = mw.getCurrentFrame()
fileDescription = frame:preprocess("{{" .. fileTitle .. "}}")
end
-- Filter non-free images
if not fileDescription or fileDescription == "" or mw.ustring.match(fileDescription, "[Nn]on%-free") then return false end
return true
end
-- 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*", d.fileNamespaces, "%s*:.*") -- [[File: or [[Image: ...
if image then
image = mw.ustring.match(image, "%b[]%s*") -- matching [[...]] to handle wikilinks nested in caption
end
return image
end
-- 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 length = mw.ustring.len(caption)
local position = 1
while position <= length do
local linkStart, linkEnd = mw.ustring.find(caption, "%b[]", position)
linkStart = linkStart or length + 1 -- avoid comparison with nil when no link
local templateStart, templateEnd = mw.ustring.find(caption, "%b{}", position)
templateStart = templateStart or length + 1 -- avoid comparison with nil when no template
local argEnd = mw.ustring.find(caption, "[|}]", position) or length + 1
if linkStart < templateStart and linkStart < argEnd then
position = linkEnd + 1 -- skip wikilink
elseif templateStart < argEnd then
position = templateEnd + 1 -- skip template
else -- argument ends before the next wikilink or template
return mw.ustring.sub(caption, 1, argEnd - 1)
end
end
return caption -- No terminator found: return entire caption
end
-- Attempt to construct a [[File:...]] block from {{infobox ... |image= ...}}
local function argImage(text)
local token = nil
local hasNamedArgs = mw.ustring.find(text, "|") and mw.ustring.find(text, "=")
if not hasNamedArgs then return nil end -- filter out any template that obviously doesn't contain an image
-- ensure image map is captured
text = mw.ustring.gsub(text, '<!%-%-imagemap%-%->', '|imagemap=')
-- find all images
local hasImages = false
local images = {}
local captureFrom = 1
while captureFrom < mw.ustring.len(text) do
local argname, position, image = mw.ustring.match(text, "|%s*([^=|]-[Ii][Mm][Aa][Gg][Ee][^=|]-)%s*=%s*()(.*)", captureFrom)
if image then -- ImageCaption=, image_size=, image_upright=, etc. do not introduce an image
local lcArgName = mw.ustring.lower(argname)
if mw.ustring.find(lcArgName, "caption")
or mw.ustring.find(lcArgName, "size")
or mw.ustring.find(lcArgName, "upright") then
image = nil
end
end
if image then
hasImages = true
images[position] = image
captureFrom = position
else
hat = getMessage( 'section' )
captureFrom = mw.ustring.len(text)
end
hat = hat .. ' ' .. getMessage( 'excerpt' ) .. ' '
end
if section then
captureFrom = 1
hat = hat .. '[[:' .. page .. '#' .. mw.uri.anchorEncode( section ) .. '|' .. displaytitle
while captureFrom < mw.ustring.len(text) do
local .. position,' image§ =' .. mw.ustring.matchgsub(text section, "|'%s*[^=|]-%[Pp]([Hh^][Oo|]+)|?[Tt^][Oo][^=|*%]-%s*=]', '%s*(1' )( ..*)", captureFrom)']].' -- remove nested links
if image then
hasImages = true
images[position] = image
captureFrom = position
else
hat = hat .. '[[:' .. page .. '|' .. displaytitle .. ']].'
captureFrom = mw.ustring.len(text)
end
if edit then
end
hat = hat .. '<span class="mw-editsection-like plainlinks"><span class="mw-editsection-bracket">[</span>['
captureFrom = 1
hat = hat .. title:fullUrl( 'action=edit' ) .. ' ' .. mw.message.new( 'editsection' ):plain()
while captureFrom < mw.ustring.len(text) do
hat = hat .. ']<span class="mw-editsection-bracket">]</span></span>'
local position, image = mw.ustring.match(text, "|%s*[^=|{}]-%s*=%s*()%[?%[?([^|{}]*%.%a%a%a%a?)%s*%f[|}]", captureFrom)
if image then
hasImages = true
if not images[position] then
images[position] = image
end
captureFrom = position
else
captureFrom = mw.ustring.len(text)
end
if config.hat then
end
hat = config.hat .. hat .. '}}'
hat = frame:preprocess( hat )
if not hasImages then return nil end
-- find all captions
local captions = {}
captureFrom = 1
while captureFrom < mw.ustring.len(text) do
local position, caption = matchAny(text, "|%s*", d.captionParams, "%s*=%s*()([^\n]+)", captureFrom)
if caption then
-- extend caption to parse "| caption = Foo {{Template\n on\n multiple lines}} Bar\n"
local bracedCaption = mw.ustring.match(text, "^[^\n]-%b{}[^\n]+", position)
if bracedCaption and bracedCaption ~= "" then caption = bracedCaption end
caption = mw.text.trim(caption)
local captionStart = mw.ustring.sub(caption, 1, 1)
if captionStart == '|' or captionStart == '}' then caption = nil end
end
if caption then
-- find nearest image, and use same index for captions table
local i = position
while i > 0 and not images[i] do
i = i - 1
if images[i] then
if not captions[i] then
captions[i] = parseCaption(caption)
end
end
end
captureFrom = position
else
hat = mw.html.create( 'div' ):addClass( 'dablink excerpt-hat' ):wikitext( hat )
captureFrom = mw.ustring.len(text)
end
else
hat = nil
end
-- findBuild allthe alt"Read textmore" link
if more and not inline then
local altTexts = {}
more = "'''[[" .. page .. '#' .. ( section or '' ) .. "|" .. getMessage( 'more' ) .. "]]'''"
for position, altText in mw.ustring.gmatch(text, "|%s*[Aa][Ll][Tt]%s*=%s*()([^\n]*)") do
more = mw.html.create( 'div' ):addClass( 'noprint excerpt-more' ):wikitext( more )
if altText then
else
more = nil
-- altText is terminated by }} or |, but first skip any matched [[...]] and {{...}}
local lookFrom = math.max( -- find position after whichever comes last: start of string, end of last ]] or end of last }}
mw.ustring.match(altText, ".*{%b{}}()") or 1, -- if multiple {{...}}, .* consumes all but one, leaving the last for %b
mw.ustring.match(altText, ".*%[%b[]%]()") or 1)
local length = mw.ustring.len(altText)
local afterText = math.min( -- find position after whichever comes first: end of string, }} or |
mw.ustring.match(altText, "()}}", lookFrom) or length+1,
mw.ustring.match(altText, "()|", lookFrom) or length+1)
altText = mw.ustring.sub(altText, 1, afterText-1) -- chop off |... or }}... which is not part of [[...]] or {{...}}
altText = mw.text.trim(altText)
local altTextStart = mw.ustring.sub(altText, 1, 1)
if altTextStart == '|' or altTextStart == '}' then altText = nil end
end
if altText then
-- find nearest image, and use same index for altTexts table
local i = position
while i > 0 and not images[i] do
i = i - 1
if images[i] then
if not altTexts[i] then
altTexts[i] = altText
end
end
end
end
end
-- Build the options for Module:Transcluder out of the template parameters and the desired defaults
-- find all image sizes
local imageSizesoptions = {}
files = files,
for position, imageSizeMatch in mw.ustring.gmatch(text, "|%s*[Ii][Mm][Aa][Gg][Ee][ _]?[Ss][Ii][Zz][Ee]%s*=%s*()([^}|\n]*)") do
lists = lists,
local imageSize = mw.ustring.match(imageSizeMatch, "=%s*([^}|\n]*)")
tables = tables,
if imageSize then
paragraphs = paragraphs,
imageSize = mw.text.trim(imageSize )
sections = subsections,
local imageSizeStart = mw.ustring.sub(imageSize, 1, 1)
categories = 0,
if imageSizeStart == '|' or imageSizeStart == '}' then imageSize = nil end
references = references,
end
only = only and mw.text.trim( only, 's' ) .. 's',
if imageSize then
noLinks = noLinks,
-- find nearest image, and use same index for imageSizes table
local inoBold = positionnoBold,
noSelfLinks = true,
while i > 0 and not images[i] do
noNonFreeFiles = onlyFreeFiles,
i = i - 1
noBehaviorSwitches = true,
if images[i] then
fixReferences = true,
if not imageSizes[i] then
linkBold = true,
imageSizes[i] = imageSize
}
end
end
end
end
end
-- Get the excerpt itself
-- sort the keys of the images table (in a table sequence), so that images can be iterated over in order
local keystitle = {}page .. '#' .. ( section or '' )
local ok, excerpt = pcall( Transcluder.get, title, options )
for key, val in pairs(images) do
if not ok then return getError( excerpt ) end
table.insert(keys, key)
if mw.text.trim( excerpt ) == '' and not only then
if section then return getError( 'section-empty', section ) else return getError( 'lead-empty' ) end
end
table.sort(keys)
-- Fix birth and death dates, but only in the first paragraph
-- add in relevant optional parameters for each image: caption, alt text and image size
if briefDates then
local imageTokens = {}
local startpos = 1 -- skip initial templates
for _, index in ipairs(keys) do
local image = images[index]s
local e = 0
local token = parseImage(image, true) -- look for image=[[File:...]] etc.
repeat
if not token then
startpos = e + 1
image = mw.ustring.match(image, "^[^}|\n]*") -- remove later arguments
s, e = mw.ustring.find( excerpt, "%s*%b{}%s*", startpos )
token = "[[" -- Add File: unless name already begins File: or Image:
until not s or s > startpos
if not matchAny(image, "^", d.fileNamespaces, "%s*:") then
s, e = mw.ustring.find( excerpt, "%b()", startpos ) -- get (...), which may be (year–year)
token = token .. "File:"
if s and s < startpos + 100 then -- look only near the start
end
local year1, conjunction, year2 = mw.ustring.match( mw.ustring.sub( excerpt, s, e ), '(%d%d%d+)(.-)(%d%d%d+)' )
token = token .. image
if year1 and year2 and (mw.ustring.match( conjunction, '[%-–—]' ) or mw.ustring.match( conjunction, '{{%s*[sS]nd%s*}}' )) then
local caption = captions[index]
local y1 = tonumber(year1)
if caption and mw.ustring.match(caption, "%S") then token = token .. "|" .. caption end
local alty2 = altTexts[index]tonumber(year2)
if alty2 then> tokeny1 and y2 < y1 + 125 and y1 <= tokentonumber( os..date( "|alt=%Y" ..)) alt endthen
excerpt = mw.ustring.sub( excerpt, 1, s ) .. year1 .. "–" .. year2 .. mw.ustring.sub( excerpt, e )
local image_size = imageSizes[index]
if image_size and mw.ustring.match(image_size, "%S") then token = token .. "|" .. image_size end
token = token .. "]]"
end
token = mw.ustring.gsub(token, "\n","") .. "\n"
table.insert(imageTokens, token)
end
return imageTokens
end
local function modifyImage(image, fileArgs)
if fileArgs then
for _, filearg in pairs(mw.text.split(fileArgs, "|")) do -- handle fileArgs=left|border etc.
local fa = mw.ustring.gsub(filearg, "=.*", "") -- "upright=0.75" → "upright"
local group = {fa} -- group of "border" is ["border"]...
for _, g in pairs(d.imageParams) do
for _, a in pairs(g) do
if fa == a then group = g end -- ...but group of "left" is ["right", "left", "center", "none"]
end
end
for _, a in pairs(group) do
image = mw.ustring.gsub(image, "|%s*" .. a .. "%f[%A]%s*=[^|%]]*", "") -- remove "|upright=0.75" etc.
image = mw.ustring.gsub(image, "|%s*" .. a .. "%s*([|%]])", "%1") -- replace "|left|" by "|" etc.
end
image = mw.ustring.gsub(image, "([|%]])", "|" .. filearg .. "%1", 1) -- replace "|" by "|left|" etc.
end
end
return image
end
-- 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'
-- options.filesOnly : only return the files and not the prose
local function parse(text, options)
local allParagraphs = 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 allParagraphs = false end -- if any para specifically requested, don't keep all
end
end
if is(options.filesOnly) then
allParagraphs = false
options.paraflags = {}
end
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
end
end
local fileArgs = options.fileargs and mw.text.trim(options.fileargs)
if fileArgs == '' then fileArgs = nil end
local leadStart = nil -- have we found some text yet?
local t = "" -- the stripped down output text
local fileText = "" -- output text with concatenated [[File:Foo|...]]\n entries
local files = 0 -- how many images so far
local paras = 0 -- how many paragraphs so far
local startLine = true -- at the start of a line (no non-spaces found since last \n)?
text = mw.ustring.gsub(text,"^%s*","") -- remove initial white space
-- Add named files
local f = options.files
if f and mw.ustring.match(f, "[^%d%s%-,]") then -- filename rather than number list
f = mw.ustring.gsub(f, "^%s*File%s*:%s*", "", 1)
f = mw.ustring.gsub(f, "^%s*Image%s*:%s*", "", 1)
f = "[[File:" .. f .. "]]"
f = modifyImage(f, "thumb")
f = modifyImage(f, fileArgs)
if checkImage(f) then fileText = fileText .. f .. "\n" end
end
repeat -- loop around parsing a template, image or paragraph
local token = mw.ustring.match(text, "^%b{}%s*") or false -- {{Template}} or {| Table |}
if not leadStart and not token then token = mw.ustring.match(text, "^%b<>%s*%b{}%s*") end -- allow <tag>{{template}} before lead has started
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
line = mw.ustring.gsub(line, "%b<>", "") -- remove all HTML tags 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
end
end
if token then -- found a template which is not the prefix to a line of text
if is(options.keepTables) and mw.ustring.sub(token, 1, 2) == '{|' then
t = t .. token -- keep tables
elseif mw.ustring.sub(token, 1, 3) == '{{#' then
t = t .. token -- keep parser functions
elseif leadStart then -- lead has already started, so keep the template within the text, unless it's a whole line (navbox etc.)
if not is(options.filesOnly) and not startLine then t = t .. token end
elseif matchAny(token, "^{{%s*", d.wantedBlockTemplates, "%s*%f[|}]") then
t = t .. token -- keep wanted block templates
-- If no file was found, try to get one from the infobox
elseif files < maxfile then -- discard template, but if we are still collecting images...
local fileNamespaces = Transcluder.getNamespaces( 'File' )
local images = argImage(token) or {}
if ( ( only == 'file' or only == 'files' ) or ( not only and ( files ~= '0' or not files ) ) ) and -- caller asked for files
if not images then
not Transcluder.matchAny( excerpt, '%[%[', fileNamespaces, ':' ) and -- and there are no files in Transcluder's output
local image = parseImage(token, false) -- look for embedded [[File:...]], |image=, etc.
config.captions -- and we have the config option required to try finding files in templates
if image then table.insert(images, image) end
then
-- We cannot distinguish the infobox from the other templates so we search them all
local infobox = Transcluder.getTemplates( excerpt );
infobox = table.concat( infobox )
local parameters = Transcluder.getParameters( infobox )
local file, captions, caption, cssclasses, cssclass
for _, pair in pairs( config.captions ) do
file = pair[1]
file = parameters[file]
if file and Transcluder.matchAny( file, '^.*%.', { '[Jj][Pp][Ee]?[Gg]', '[Pp][Nn][Gg]', '[Gg][Ii][Ff]', '[Ss][Vv][Gg]' }, '.*' ) then
file = mw.ustring.match( file, '%[?%[?.-:([^{|]+)%]?%]?' ) or file -- [[File:Example.jpg{{!}}upright=1.5]] to Example.jpg
captions = pair[2]
for _, p in pairs( captions ) do
if parameters[ p ] then caption = parameters[ p ] break end
end
for _, image in ipairs(images) do
-- Check for CSS classes
if files < maxfile and checkImage(image) then -- if image is found and qualifies (not a sound file, non-free, etc.)
-- We opt to use skin-invert-image instead of skin-invert
files = files + 1 -- count the file, whether displaying it or not
-- in all other cases, the CSS provided in the infobox is used
if options.fileflags and options.fileflags[files] then -- if displaying this image
if pair[3] then
image = modifyImage(image, "thumb")
cssclasses = pair[3]
image = modifyImage(image, fileArgs)
for _, p in pairs(cssclasses) do
fileText = fileText .. image
if parameters[p] then
cssclass = ((parameters[p] == 'skin-invert') and 'skin-invert-image' or parameters[p])
break
end
end
end
excerpt = '[[File:' .. file ..
(cssclass and ('|class=' .. cssclass) or '') ..
'|thumb|' .. (caption or '') .. ']]' .. excerpt
if ( onlyFreeFiles ) then
excerpt = Transcluder.removeNonFreeFiles( excerpt )
end
break
end
end
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
image = modifyImage(image, fileArgs)
fileText = fileText .. image
end
end
else -- got a paragraph, which ends at a file, image, blank line or end of text
local afterEnd = mw.ustring.len(text) + 1
local blankPosition = mw.ustring.find(text, "\n%s*\n") or afterEnd -- position of next paragraph delimiter (or end of text)
local endPosition = 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,
blankPosition)
token = mw.ustring.sub(text, 1, endPosition-1)
if blankPosition < afterEnd and blankPosition == endPosition then -- paragraph ends with a blank line
token = token .. mw.ustring.match(text, "\n%s*\n", blankPosition)
end
local isHatnote = not(leadStart) and mw.ustring.sub(token, 1, 1) == ':'
if not isHatnote then
leadStart = leadStart or mw.ustring.len(t) + 1 -- we got a paragraph, so mark the start of the lead section
paras = paras + 1
if allParagraphs or (options.paraflags and options.paraflags[paras]) then t = t .. token end -- add if this paragraph wanted
end
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
startLine = mw.ustring.find(token, "\n%s*$") -- will the next token be the first non-space on a line?
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 fileText .. text
end
local function cleanupText(text, options)
text = mw.ustring.gsub(text, "<!%-%-.-%-%->","") -- remove HTML comments
text = mw.ustring.gsub(text, "<[Nn][Oo][Ii][Nn][Cc][Ll][Uu][Dd][Ee]>.-</[Nn][Oo][Ii][Nn][Cc][Ll][Uu][Dd][Ee]>", "") -- remove noinclude bits
if mw.ustring.find(text, "[Oo][Nn][Ll][Yy][Ii][Nn][Cc][Ll][Uu][Dd][Ee]") then -- avoid expensive search if possible
text = mw.ustring.gsub(text, "</[Oo][Nn][Ll][Yy][Ii][Nn][Cc][Ll][Uu][Dd][Ee]>.-<[Oo][Nn][Ll][Yy][Ii][Nn][Cc][Ll][Uu][Dd][Ee]>", "") -- remove text between onlyinclude sections
text = mw.ustring.gsub(text, "^.-<[Oo][Nn][Ll][Yy][Ii][Nn][Cc][Ll][Uu][Dd][Ee]>", "") -- remove text before first onlyinclude section
text = mw.ustring.gsub(text, "</[Oo][Nn][Ll][Yy][Ii][Nn][Cc][Ll][Uu][Dd][Ee]>.*", "") -- remove text after last onlyinclude section
end
if not is(options.keepSubsections) then
text = mw.ustring.gsub(text, "\n==.*","") -- remove first ==Heading== and everything after it
text = mw.ustring.gsub(text, "^==.*","") -- ...even if the lead is empty
end
if not is(options.keepRefs) then
text = mw.ustring.gsub(text, "<%s*[Rr][Ee][Ff][^>]-/%s*>", "") -- remove refs cited elsewhere
text = mw.ustring.gsub(text, "<%s*[Rr][Ee][Ff].->.-<%s*/%s*[Rr][Ee][Ff]%s*>", "") -- remove refs
text = mw.ustring.gsub(text, "%b{}", stripTemplate) -- remove unwanted templates such as references
end
text = mw.ustring.gsub(text, "<%s*[Ss][Cc][Oo][Rr][Ee].->.-<%s*/%s*[Ss][Cc][Oo][Rr][Ee]%s*>", "") -- remove musical scores
text = mw.ustring.gsub(text, "<%s*[Ii][Mm][Aa][Gg][Ee][Mm][Aa][Pp].->.-<%s*/%s*[Ii][Mm][Aa][Gg][Ee][Mm][Aa][Pp]%s*>", convertImageMap) -- convert imagemaps into standard images
text = mw.ustring.gsub(text, "%s*{{%s*[Tt][Oo][Cc].-}}", "") -- remove most common tables of contents
text = mw.ustring.gsub(text, "%s*__[A-Z]*TOC__", "") -- remove TOC behavior switches
text = mw.ustring.gsub(text, "\n%s*{{%s*[Pp]p%-.-}}", "\n") -- remove protection templates
text = mw.ustring.gsub(text, "%s*{{[^{|}]*[Ss]idebar%s*}}", "") -- remove most sidebars
text = mw.ustring.gsub(text, "%s*{{[^{|}]*%-[Ss]tub%s*}}", "") -- remove most stub templates
text = mw.ustring.gsub(text, "%s*%[%[%s*:?[Cc]ategory:.-%]%]", "") -- remove categories
text = mw.ustring.gsub(text, "^:[^\n]+\n","") -- remove DIY hatnote indented with a colon
return text
end
-- Unlike other elements, templates are filtered here
-- Parse a ==Section== from a page
-- because we had to search the infoboxes for files
local function getSection(text, section, mainOnly)
local trash
local escapedSection = mw.ustring.gsub(mw.uri.decode(section), "([%^%$%(%)%%%.%[%]%*%+%-%?])", "%%%1") -- %26 → & etc, then ^ → %^ etc.
if only and ( only == 'template' or only == 'templates' ) then
local level, content = mw.ustring.match(text .. "\n", "\n(==+)%s*" .. escapedSection .. "%s*==.-\n(.*)")
trash, excerpt = Transcluder.getTemplates( excerpt, templates );
if not content then return luaError("sectionNotFound", section) end
else -- Remove blacklisted templates
local nextSection
local blacklist = config.blacklist and table.concat( config.blacklist, ',' ) or ''
if mainOnly then
if templates then
nextSection = "\n==.*" -- Main part of section terminates at any level of header
if string.sub( templates, 1, 1 ) == '-' then --Unwanted templates. Append to blacklist
else
blacklist = templates .. ',' .. blacklist
nextSection = "\n==" .. mw.ustring.rep("=?", #level - 2) .. "[^=].*" -- "===" → "\n===?[^=].*", matching "==" or "===" but not "===="
else --Wanted templates. Replaces blacklist and acts as whitelist
end
blacklist = templates
content = mw.ustring.gsub(content, nextSection, "") -- remove later sections with headings at this level or higher
if mw.ustring.match(content, "^%s*$") then return luaError("sectionEmpty", section) end
return content
end
-- Parse a <section begin="Name of the fragment">
-- @todo Implement custom parsing of fragments rather than relying on #lst
local function getFragment(page, fragment)
local frame = mw.getCurrentFrame()
local text = frame:callParserFunction('#lst', page, fragment)
if mw.ustring.match(text, "^%s*$") then return luaError("fragmentEmpty", fragment) end
return text
end
-- Remove unmatched <tag> or </tag> tags
local function fixTags(text, tag)
local startCount = 0
for i in mw.ustring.gmatch(text, "<%s*" .. tag .. "%f[^%w_].->") do startCount = startCount + 1 end
local endCount = 0
for i in mw.ustring.gmatch(text, "<%s*/" .. tag .. "%f[^%w_].->") do endCount = endCount + 1 end
if startCount > endCount then -- more <tag> than </tag>: remove the last few <tag>s
local i = 0
text = mw.ustring.gsub(text, "<%s*" .. tag .. "%f[^%w_].->", function(t)
i = i + 1
if i > endCount then return "" else return nil end
end) -- "end" here terminates the anonymous replacement function(t) passed to gsub
elseif endCount > startCount then -- more </tag> than <tag>: remove the first few </tag>s
text = mw.ustring.gsub(text, "<%s*/" .. tag .. "%f[^%w_].->", "", endCount - startCount)
end
return text
end
local function fixTemplates(text)
repeat -- hide matched {{template}}s including nested templates
local t = text
text = mw.ustring.gsub(text, "{(%b{})}", "\27{\27%1\27}\27") -- {{sometemplate}} → E{Esometemplate}E}E where E represents escape
text = mw.ustring.gsub(text, "(< *math[^>]*>[^<]-)}}(.-< */math *>)", "%1}\27}\27%2") -- <math>\{sqrt\{hat{x}}</math> → <math>\{sqrt\{hat{x}E}E</math>
until text == t
text = text.gsub(text, "([{}])%1[^\27].*", "") -- remove unmatched {{, }} and everything thereafter, avoiding }E}E etc.
text = text.gsub(text, "([{}])%1$", "") -- remove unmatched {{, }} at end of text
text = mw.ustring.gsub(text, "\27", "") -- unhide matched pairs: E{E{ → {{, etc.
return text
end
local function fixLinks(text)
repeat -- hide matched [[wikilink]]s including nested links like [[File:Example.jpg|Some [[nested]] link.]]
local t = text
text = mw.ustring.gsub(text, "%[(%b[])%]", "\27[\27%1\27]\27")
until text == t
text = text.gsub(text, "([%[%]])%1[^\27].*", "") -- remove unmatched [[ or ]] and everything thereafter, avoiding ]E]E etc.
text = text.gsub(text, "([%[%]])%1$", "") -- remove unmatched [[ or ]] at end of text
text = mw.ustring.gsub(text, "\27", "") -- unhide matched pairs: ]E]E → ]], etc.
return text
end
-- Replace the first call to each reference defined outside of the text for the full reference, to prevent undefined references
-- Then prefix the page title to the reference names to prevent conflicts
-- that is, replace <ref name="Foo"> for <ref name="Title of the article Foo">
-- and also <ref name="Foo" /> for <ref name="Title of the article Foo" />
-- also remove reference groups: <ref name="Foo" group="Bar"> for <ref name="Title of the article Foo">
-- and <ref group="Bar"> for <ref>
-- @todo The current regex may fail in cases with both kinds of quotes, like <ref name="Darwin's book">
local function fixRefs(text, page, full)
if not full then full = getContent(page) end
local refNames = {}
local refName
local refBody
local position = 1
while position < mw.ustring.len(text) do
refName, position = mw.ustring.match(text, "<%s*[Rr][Ee][Ff][^>]*name%s*=%s*[\"']?([^\"'>]+)[\"']?[^>]*/%s*>()", position)
if refName then
refName = mw.text.trim(refName)
if not refNames[refName] then -- make sure we process each ref name only once
table.insert(refNames, refName)
refName = mw.ustring.gsub(refName, "[%^%$%(%)%.%[%]%*%+%-%?%%]", "%%%0") -- escape special characters
refBody = mw.ustring.match(text, "<%s*[Rr][Ee][Ff][^>]*name%s*=%s*[\"']?%s*" .. refName .. "%s*[\"']?[^>/]*>.-<%s*/%s*[Rr][Ee][Ff]%s*>")
if not refBody then -- the ref body is not in the excerpt
refBody = mw.ustring.match(full, "<%s*[Rr][Ee][Ff][^>]*name%s*=%s*[\"']?%s*" .. refName .. "%s*[\"']?[^/>]*>.-<%s*/%s*[Rr][Ee][Ff]%s*>")
if refBody then -- the ref body was found elsewhere
text = mw.ustring.gsub(text, "<%s*[Rr][Ee][Ff][^>]*name%s*=%s*[\"']?%s*" .. refName .. "%s*[\"']?[^>]*/?%s*>", refBody, 1)
end
end
end
else
positionblacklist = mw'-' .ustring.len(text) blacklist
end
trash, excerpt = Transcluder.getTemplates( excerpt, blacklist );
end
text = mw.ustring.gsub(text, "<%s*[Rr][Ee][Ff][^>]*name%s*=%s*[\"']?([^\"'>/]+)[\"']?[^>/]*(/?)%s*>", '<ref name="' .. page .. ' %1" %2>')
text = mw.ustring.gsub(text, "<%s*[Rr][Ee][Ff][^>]*group%s*=%s*[\"']?[^\"'>/]+[\"']%s*>", '<ref>')
return text
end
-- Remove extra line breaks but leave one before and after so the parser interprets lists, tables, etc. correctly
-- Replace the bold title or synonym near the start of the article by a wikilink to the article
excerpt = mw.text.trim( excerpt )
function linkBold(text, page)
excerpt = string.gsub( excerpt, '\n\n\n+', '\n\n' )
local lang = mw.language.getContentLanguage()
excerpt = '\n' .. excerpt .. '\n'
local position = mw.ustring.find(text, "'''" .. lang:ucfirst(page) .. "'''", 1, true) -- look for "'''Foo''' is..." (uc) or "A '''foo''' is..." (lc)
or mw.ustring.find(text, "'''" .. lang:lcfirst(page) .. "'''", 1, true) -- plain search: special characters in page represent themselves
if position then
local length = mw.ustring.len(page)
text = mw.ustring.sub(text, 1, position + 2) .. "[[" .. mw.ustring.sub(text, position + 3, position + length + 2) .. "]]" .. mw.ustring.sub(text, position + length + 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 not mw.ustring.find(b, "%[") then -- if not wikilinked
return "'''[[" .. page .. "|" .. b .. "]]'''" -- replace '''Foo''' by '''[[page|Foo]]'''
else
return nil -- instruct gsub to make no change
end
end, 1) -- "end" here terminates the anonymous replacement function(a, b) passed to gsub
end
return text
end
-- Remove nested categories
-- Main function for modules
excerpt = frame:preprocess( excerpt )
local function get(page, options)
local categories, excerpt = Transcluder.getCategories( excerpt, options.categories )
if options.errors then errors = options.errors end
-- Add tracking categories
if not page or page == "" then return luaError("noPage") end
if config.categories then
local contentCategory = config.categories.content
local text
if contentCategory and mw.title.getCurrentTitle().isContentPage then
page, section = mw.ustring.match(page, "([^#]+)#?([^#]*)")
excerpt = excerpt .. '[[Category:' .. contentCategory .. ']]'
text, page = getContent(page)
end
if not page then return luaError("noPage") end
local namespaceCategory = config.categories[ mw.title.getCurrentTitle().namespace ]
if not text then return luaError("pageNotFound", page) end
if namespaceCategory then
local full = text -- save the full text for later
excerpt = excerpt .. '[[Category:' .. namespaceCategory .. ']]'
end
if is(options.fragment) then
text = getFragment(page, options.fragment)
end
-- Load the styles
if is(section) then
local styles
text = getSection(text, section)
if config.styles then
styles = frame:extensionTag( 'templatestyles', '', { src = config.styles } )
end
-- Combine and return the elements
-- Strip text of all undersirables
if inline then
text = cleanupText(text, options)
return mw.text.trim( excerpt )
text = parse(text, options)
-- Replace the bold title or synonym near the start of the article by a wikilink to the article
text = linkBold(text, page)
-- Remove '''bold text''' if requested
if is(options.nobold) then text = mw.ustring.gsub(text, "'''", "") end
-- Keep only tables if requested
if is(options.tablesOnly) then text = getTables(text) end
-- Keep only lists if requested
if is(options.listsOnly) then text = getLists(text) end
-- Seek and destroy unterminated templates, links and tags
text = fixTemplates(text)
text = fixLinks(text)
text = fixTags(text, "div")
-- Fix broken references
if is(options.keepRefs) then text = fixRefs(text, page, full) end
return text
end
-- Main invocation function for templates
local function main(frame)
local args = parseArgs(frame)
local page = args[1]
local ok, text = pcall(get, page, args)
if not ok then
text = d.errors.prefix .. text
if d.errorsCategory and d.errorsCategory ~= "" and mw.title.getCurrentTitle().isContentPage then
text = text .. '[[' .. d.errorsCategory .. ']]'
end
return mw.html.create('div'):addClass('error'):wikitext(text)
end
local tag = 'div'
return frame:preprocess(text)
if quote then
tag = 'blockquote'
end
excerpt = mw.html.create( 'div' ):addClass( 'excerpt' ):wikitext( excerpt )
local block = mw.html.create( tag ):addClass( 'excerpt-block' ):addClass( class )
return block:node( styles ):node( hat ):node( excerpt ):node( more )
end
-- Entry points for templates
function p.main(frame) return main(frame) end
function p.wikiError(message, value) return wikiError(message, value) end
-- Entry points for other Lua modules
function p.get(page, options) return get(page, options) end
function p.getContent(page) return getContent(page) end
function p.getSection(text, section) return getSection(text, section) end
function p.getTables(text, options) return getTables(text, options) end
function p.getLists(text, options) return getLists(text, options) end
function p.parse(text, options) return parse(text, options) end
function p.parseImage(text, start) return parseImage(text, start) end
function p.parseArgs(frame) return parseArgs(frame) end
function p.argImage(text) return argImage(text) end
function p.checkImage(image) return checkImage(image) end
function p.cleanupText(text, options) return cleanupText(text, options) end
function p.luaError(message, value) return luaError(message, value) end
function p.is(value) return is(value) end
function p.numberFlags(str) return numberFlags(str) end
-- Entry points for backwards compatibility
function p.getsectionlead(text, sectionframe ) return getSectionp.main(text, sectionframe ) end
function p.parseimageexcerpt(text, startframe ) return parseImagep.main(text, startframe ) end
function p.checkimage(image) return checkImage(image) end
function p.argimage(text) return argImage(text) end
function p.numberflags(str) return numberFlags(str) end
return p
| 