Module:Excerpt: Difference between revisions

Revision as of 11:37, 18 May 2018

Latest revision as of 01:57, 28 January 2025

(150 intermediate revisions by 10 users not shown)

Line 1:

-- Module:Excerpt implements the Excerpt template

local p = {}

-- Documentation and master version: https://en.wikipedia.org/wiki/Module:Excerpt

local mRedirect = require('Module:Redirect')

-- Authors: User:Sophivorus, User:Certes, User:Aidan9382 & others

-- License: CC-BY-SA-3.0

local Transcluder = require( 'Module:Transcluder' )

local errors

-- Return blank text, or an error message if requested

local function err(text)

if errors then error(text, 2) end

return ""

end

local yesno = require( 'Module:Yesno' )

-- Check image for suitablity

local function checkimage(image)

local page = mw.ustring.match(image, "([Ff]ile%s*:[^|%]]*)") -- File:(name) ...

or mw.ustring.match(image, "([Ii]mage%s*:[^|%]]*)") -- or Image:(name) ...

if not page then return nil end

local ok, config = pcall( require, 'Module:Excerpt/config' )

local title = mw.title.new(":" .. page) -- Read description page (for :File:Foo rather than File:Foo)

if not titleok then returnconfig nil= {} end

local redirp = mRedirect.getTarget(title){}

if redir then title = mw.title.new(redir) end

-- Helper function to get arguments

local frame = mw.getCurrentFrame()

local args

local desc = frame:preprocess("{{" .. title.prefixedText .. "}}")

local function getArg( key, default )

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

local value = args[ key ]

if value and mw.text.trim( value ) ~= '' then

return value

end

return default

end

-- Helper function to handle errors

-- Attempt to parse [[File:...]] or [[Image:...]], either anywhere or at the start

local function parseimagegetError(text message, startvalue )

if type( message ) == 'string' then

local startre = ""

message = Transcluder.getError( message, value )

if start then startre = "^" end -- a true flag restricts search to start of string

local image = mw.ustring.match(text, startre .. "%[%[%s*[Ff]ile%s*:.*") -- [[File: ...

or mw.ustring.match(text, startre .. "%[%[%s*[Ii]mage%s*:.*") -- or [[Image: ...

if image then

image = mw.ustring.match(image, "%b[]%s*") -- match [[...]] to handle nesting

end

if config.categories and config.categories.errors and mw.title.getCurrentTitle().isContentPage then

return image

message:node( '[[Category:' .. config.categories.errors .. ']]' )

end

return message

end

-- Helper function to get localized messages

-- Attempt to construct a [[File:...]] block from {{infobox ... |image= ...}}

local function argimagegetMessage(text key )

local tokenok, TNT = nilpcall( require, 'Module:TNT' )

if not ok then return key end

if mw.ustring.match(text, "%{%{%s*[Ii]nfobox") then

return TNT.format( 'I18n/Module:Excerpt.tab', key )

local image = mw.ustring.match(text, "|%s*image%s*=%s*([^%}|]*)") -- parse image= argument...

or mw.ustring.match(text, "|%s*Cover%s*=%s*([^%}|]-)") -- or Cover= from Infobox album

if image then -- add in relevant parameters: caption, alt text and image size

token = "[[" -- Add File: unless name already begins File: or Image:

if not (mw.ustring.match(image, "^[Ff]ile%s*:")

or mw.ustring.match(image, "^[Ii]mage%s*:")) then

token = token .. "File:"

end

token = token .. image

local caption = mw.ustring.match(text, "|%s*[Cc]aption%s*=%s*([^%}|]*)")

if caption 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 then token = token .. "|" .. image_size end

token = mw.ustring.gsub(token, "\n","") .. "]]\n"

end

end

return token

end

-- EntryMain entry point for Lua callerstemplates

function p.main( frame )

-- Returns a string value: text of the lead of a page

args = Transcluder.parseArgs( frame )

function p._lead(pagenames, options)

errors = options.errors

-- Make sure the requested page exists

if not pagenames or #pagenames < 1 then return err("No page names given") end

local pagenamepage = getArg( 1 )

if not page or page == '{{{1}}}' then return getError( 'no-page' ) end

local text

local pagecounttitle = #pagenamesmw.title.new(page)

if not title then return getError( 'invalid-title', page ) end

local firstpage = pagenames[1] or "(nil)" -- save for error message, as it the name will be deleted

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

-- read the page, or a random one if multiple pages were provided

local section = getArg( 2, mw.ustring.match( getArg( 1 ), '[^#]+#(.+)' ) )

if pagecount > 1 then math.randomseed(os.time()) end

local hat = yesno( getArg( 'hat', true ) )

while not text and pagecount > 0 do

local edit = yesno( getArg( 'edit', true ) )

local pagenum = 1

local this = getArg( 'this' )

if pagecount > 1 then pagenum = math.random(pagecount) end -- pick a random title

local only = getArg( 'only' )

pagename = pagenames[pagenum]

local files = getArg( 'files', getArg( 'file', ( only == 'file' and 1 ) ) )

if pagename and pagename ~= "" then

local lists = getArg( 'lists', getArg( 'list', ( only == 'list' and 1 ) ) )

pagename = mw.ustring.match(pagename, "%[%[%s*(.-)[]|#]") or pagename -- "[[Foo|Bar]]" → "Foo"

local tables = getArg( 'tables', getArg( 'table', ( only == 'table' and 1 ) ) )

pagename = mw.ustring.match(pagename, "%S.*%S") -- strip leading and trailing white space

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

if pagename and pagename ~= "" then

if hat and not inline then

local title = mw.title.new(pagename) -- Find the lead section of the named page

if this then

if not title then return err("No title for page name " .. pagename) end

hat = this

local redir = mRedirect.getTarget(title)

elseif quote then

if redir then title = mw.title.new(redir) end

hat = getMessage( 'this' )

elseif only then

text = title:getContent()

hat = getMessage( only )

end

else

hat = getMessage( 'section' )

end

hat = hat .. ' ' .. getMessage( 'excerpt' ) .. ' '

if not text then table.remove(pagenames, pagenum) end -- this one didn't work; try another

if section then

pagecount = pagecount - 1 -- ensure that we exit the loop eventually

hat = hat .. '[[:' .. page .. '#' .. mw.uri.anchorEncode( section ) .. '|' .. displaytitle

.. ' § ' .. mw.ustring.gsub( section, '%[%[([^]|]+)|?[^]]*%]%]', '%1' ) .. ']].' -- remove nested links

else

hat = hat .. '[[:' .. page .. '|' .. displaytitle .. ']].'

end

if edit then

hat = hat .. '<span class="mw-editsection-like plainlinks"><span class="mw-editsection-bracket">[</span>['

hat = hat .. title:fullUrl( 'action=edit' ) .. ' ' .. mw.message.new( 'editsection' ):plain()

hat = hat .. ']<span class="mw-editsection-bracket">]</span></span>'

end

if config.hat then

hat = config.hat .. hat .. '}}'

hat = frame:preprocess( hat )

else

hat = mw.html.create( 'div' ):addClass( 'dablink excerpt-hat' ):wikitext( hat )

end

else

hat = nil

end

if not text then return err("Cannot read a valid page: first name is " .. firstpage) end

-- Build the "Read more" link

text = mw.ustring.gsub(text, "<!%-%-.-%-%->","") -- remove HTML comments

if more and not inline then

text = mw.ustring.gsub(text, "%c%s*==.*","") -- remove first heading and everything after it

more = "'''[[" .. page .. '#' .. ( section or '' ) .. "|" .. getMessage( 'more' ) .. "]]'''"

text = mw.ustring.gsub(text, "<noinclude>.-</noinclude>", "") -- remove noinclude bits

more = mw.html.create( 'div' ):addClass( 'noprint excerpt-more' ):wikitext( more )

text = mw.ustring.gsub(text, "<%s*ref[^>]-/%s*>", "") -- remove refs cited elsewhere

else

text = mw.ustring.gsub(text, "<%s*ref.->.-<%s*/%s*ref%s*>", "") -- remove refs

more = nil

text = mw.ustring.gsub(text, "<%s*imagemap.->.-<%s*/%s*imagemap%s*>", "") -- remove imagemaps

for _, t in pairs {"[Ee]fn", "[Ee]fn-la", "[Ee]l[mn]", "[Rr]p?", "[Ss]fn[bp]", "[Ss]f[bn]", "NoteTag", "#[Tt]ag:%s*[Rr]ef", "[Rr]efn",

"[CcDd]n", "[Cc]itation needed", "[Dd]isambiguation needed"} do

text = mw.ustring.gsub(text, "{{%s*" .. t .. "%s*|.-}}", "") -- remove ref and footnote templates

end

text = mw.ustring.gsub(text, "\n%s*{{%s*[Tt][Oo][Cc].-}}", "\n") -- remove most common tables of contents

-- Build the options for Module:Transcluder out of the template parameters and the desired defaults

local allparas = true -- keep all paragraphs?

iflocal options.paraflags then= {

files = files,

for _, v in pairs(options.paraflags) do

lists = lists,

if v then allparas = false end -- if any para specifically requested, don't keep all

tables = tables,

end

paragraphs = paragraphs,

sections = subsections,

categories = 0,

references = references,

only = only and mw.text.trim( only, 's' ) .. 's',

noLinks = noLinks,

noBold = noBold,

noSelfLinks = true,

noNonFreeFiles = onlyFreeFiles,

noBehaviorSwitches = true,

fixReferences = true,

linkBold = true,

}

-- Get the excerpt itself

local title = page .. '#' .. ( section or '' )

local ok, excerpt = pcall( Transcluder.get, title, options )

if not ok then return getError( excerpt ) end

if mw.text.trim( excerpt ) == '' and not only then

if section then return getError( 'section-empty', section ) else return getError( 'lead-empty' ) end

end

-- Fix birth and death dates, but only in the first paragraph

local maxfile = 0 -- for efficiency, stop checking images after this many have been found

if options.fileflagsbriefDates then

local startpos = 1 -- skip initial templates

for k, v in pairs(options.fileflags) do

local s

if v and k > maxfile then maxfile = k end

local e = 0

repeat

startpos = e + 1

s, e = mw.ustring.find( excerpt, "%s*%b{}%s*", startpos )

until not s or s > startpos

s, e = mw.ustring.find( excerpt, "%b()", startpos ) -- get (...), which may be (year–year)

if s and s < startpos + 100 then -- look only near the start

local year1, conjunction, year2 = mw.ustring.match( mw.ustring.sub( excerpt, s, e ), '(%d%d%d+)(.-)(%d%d%d+)' )

if year1 and year2 and (mw.ustring.match( conjunction, '[%-–—]' ) or mw.ustring.match( conjunction, '{{%s*[sS]nd%s*}}' )) then

local y1 = tonumber(year1)

local y2 = tonumber(year2)

if y2 > y1 and y2 < y1 + 125 and y1 <= tonumber( os.date( "%Y" )) then

excerpt = mw.ustring.sub( excerpt, 1, s ) .. year1 .. "–" .. year2 .. mw.ustring.sub( excerpt, e )

end

end

end

end

-- aIf basicno parserfile was found, try to trimget one downfrom the leadinfobox

local fileNamespaces = Transcluder.getNamespaces( 'File' )

local inlead = false -- have we found some text yet?

if ( ( only == 'file' or only == 'files' ) or ( not only and ( files ~= '0' or not files ) ) ) and -- caller asked for files

local t = "" -- the stripped down output text

not Transcluder.matchAny( excerpt, '%[%[', fileNamespaces, ':' ) and -- and there are no files in Transcluder's output

local files = 0 -- how many [[Image: or [[File: so far

config.captions -- and we have the config option required to try finding files in templates

local paras = 0 -- how many paragraphs so far

then

-- We cannot distinguish the infobox from the other templates so we search them all

text = mw.ustring.gsub(text,"^%s*","") -- remove initial white space

local infobox = Transcluder.getTemplates( excerpt );

repeat -- loop around parsing a template, image or paragraph

infobox = table.concat( infobox )

local token = mw.ustring.match(text, "^%b{}%s*") or false -- {{Template}}

local parameters = Transcluder.getParameters( infobox )

if token then

local file, captions, caption, cssclasses, cssclass

if inlead then -- keep comments and templates only within text body

for _, pair in pairs( config.captions ) do

t = t .. token

file = pair[1]

elseif files < maxfile then -- look for [[File:... embedded in an infobox etc. in the preamble

file = parameters[file]

local image = parseimage(token, false) or argimage(token)

if file and Transcluder.matchAny( file, '^.*%.', { '[Jj][Pp][Ee]?[Gg]', '[Pp][Nn][Gg]', '[Gg][Ii][Ff]', '[Ss][Vv][Gg]' }, '.*' ) then

if image and checkimage(image) then -- keep comments and templates only within text body

file = mw.ustring.match( file, '%[?%[?.-:([^{|]+)%]?%]?' ) or file -- [[File:Example.jpg{{!}}upright=1.5]] to Example.jpg

files = files + 1

captions = pair[2]

if options.fileflags and options.fileflags[files] then

for _, p in pairs( captions ) do

image = mw.ustring.gsub(image, "|%s*frameless%s*%f[|%]]", "") -- make image a thumbnail, not frameless etc.

if parameters[ p ] then caption = parameters[ p ] break end

image = mw.ustring.gsub(image, "|%s*framed?%s*%f[|%]]", "")

end

if not mw.ustring.match(image, "|%s*thumb%s*%f[|%]]")

and not mw.ustring.match(image, "|%s*thumbnail%s*%f[|%]]") then

-- Check for CSS classes

image = mw.ustring.gsub(image, "(%]%]%s*)$", "|thumb%1")

-- We opt to use skin-invert-image instead of skin-invert

-- in all other cases, the CSS provided in the infobox is used

if pair[3] then

cssclasses = pair[3]

for _, p in pairs(cssclasses) do

if parameters[p] then

cssclass = ((parameters[p] == 'skin-invert') and 'skin-invert-image' or parameters[p])

break

end

if options.fileargs then image = mw.ustring.gsub(image, "(%]%]%s*)$", "|" .. options.fileargs .. "%1") end

t = t .. image

end

end

end

excerpt = '[[File:' .. file ..

else

(cssclass and ('|class=' .. cssclass) or '') ..

token = parseimage(text, true)

'|thumb|' .. (caption or '') .. ']]' .. excerpt

if token then

if files( <onlyFreeFiles maxfile and checkimage(token) then

excerpt = Transcluder.removeNonFreeFiles( excerpt )

files = files + 1

if options.fileflags and options.fileflags[files] then

local image = token

if options.fileargs then image = mw.ustring.gsub(image, "(%]%]%s*)$", "|" .. options.fileargs .. "%1") end

t = t .. 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 blankpos = mw.ustring.find(text, "\n%s*\n") or afterend

local endpos = math.min(

mw.ustring.find(text, "%[%[%s*[Ff]ile%s*:") or afterend,

mw.ustring.find(text, "%[%[%s*[Ff]ile%s*:") or afterend,

blankpos)

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)

end

inlead = truebreak

paras = paras + 1

if allparas or (options.paraflags and options.paraflags[paras]) then t = t .. token end

end

end

end

-- Unlike other elements, templates are filtered here

if token then text = mw.ustring.sub(text, mw.ustring.len(token)+1) end

-- because we had to search the infoboxes for files

until not text or text == "" or not token or token == ""

local trash

if only and ( only == 'template' or only == 'templates' ) then

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

trash, excerpt = Transcluder.getTemplates( excerpt, templates );

else -- Remove blacklisted templates

if options.more then text = text .. " '''[[" .. pagename .. "|" .. options.more .. "]]'''" end

local blacklist = config.blacklist and table.concat( config.blacklist, ',' ) or ''

return text

if templates then

end

if string.sub( templates, 1, 1 ) == '-' then --Unwanted templates. Append to blacklist

blacklist = templates .. ',' .. blacklist

-- 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}

else --Wanted templates. Replaces blacklist and acts as whitelist

local function numberflags(str)

blacklist = templates

local ranges = mw.text.split(str, ",") -- parse ranges, e.g. "1,3-5" → {"1","3-5"}

end

local flags = {}

else

for _, r in pairs(ranges) do

blacklist = '-' .. blacklist

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

end

trash, excerpt = Transcluder.getTemplates( excerpt, blacklist );

end

return flags

end

-- Remove extra line breaks but leave one before and after so the parser interprets lists, tables, etc. correctly

-- Shared template invocation code for lead and random functions

excerpt = mw.text.trim( excerpt )

local function leadrandom(frame, israndom)

excerpt = string.gsub( excerpt, '\n\n\n+', '\n\n' )

-- args = { 1,2,... = page names, paragraphs = list e.g. "1,3-5", files = list, more = text}

excerpt = '\n' .. excerpt .. '\n'

local args = frame.args -- from calling module

local pargs = frame:getParent().args -- from template

-- Remove nested categories

local pagenames = { args[1] or pargs[1] } -- For lead, ignore all but the first unnamed argument

excerpt = frame:preprocess( excerpt )

if israndom then

local categories, excerpt = Transcluder.getCategories( excerpt, options.categories )

-- For random, accept any number of page names. If more than one, we'll pick one randomly

for i, p in pairs(args) do

-- Add tracking categories

if p and type(i) == 'number' and i > 1 then table.insert(pagenames, p) end

if config.categories then

local contentCategory = config.categories.content

if contentCategory and mw.title.getCurrentTitle().isContentPage then

excerpt = excerpt .. '[[Category:' .. contentCategory .. ']]'

end

local namespaceCategory = config.categories[ mw.title.getCurrentTitle().namespace ]

for i, p in pairs(pargs) do

if namespaceCategory then

if p and type(i) == 'number' and i > 1 and not args[i] then table.insert(pagenames, p) end

excerpt = excerpt .. '[[Category:' .. namespaceCategory .. ']]'

end

end

-- Load the styles

local options = {}

local styles

options.paraflags = numberflags(args["paragraphs"] or pargs["paragraphs"] or "") -- parse paragraphs, e.g. "1,3-5" → {"1","3-5"}

if config.styles then

options.fileflags = numberflags(args["files"] or pargs["files"] or "") -- parse file numbers

styles = frame:extensionTag( 'templatestyles', '', { src = config.styles } )

options.fileargs = args["fileargs"] or pargs["fileargs"]

end

options.more = args["more"] or pargs["more"]

if options.more and options.more == "" then options.more = "Read more..." end -- more= is short for this default text

options.errors = args["errors"] or pargs["errors"]

-- Combine and return the elements

local text = p._lead(pagenames, options)

if inline then

return frame:preprocess(text)

return mw.text.trim( excerpt )

end

local tag = 'div'

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 templatebackwards callers using #invoke:compatibility

function p.lead( frame ) return leadrandomp.main( frame, false) end

function p.randomexcerpt( frame ) return leadrandomp.main( frame, true) end

return p