Module:Delink/sandbox
< Module:Delink
Jump to navigation
Jump to search
Revision as of 14:52, 18 April 2014 by blackwiki>Mr. Stradivarius (attempt to simplify the p._delinkWikilink function)
Documentation for this module may be created at Module:Delink/sandbox/doc
-- This module de-links most wikitext.
local getArgs = require('Module:Arguments').getArgs
local yesno = require('Module:Yesno')
-- Often-used functions and variables
local htmlDecode = mw.text.decode
local uriDecode = mw.uri.decode
local isKnownLanguageTag = mw.language.isKnownLanguageTag
local namespaces = mw.site.namespaces
p = {}
local function whitespaceYesno(val)
-- Like yesno, but trims whitespace from vals and removes blank strings.
if type(val) == 'string' then
val = val:match('^%s*(.-)%s*$')
if val == '' then
return nil
end
end
return yesno(val)
end
function p._delinkWikilink(s)
-- s is a string starting with '[[' and ending with ']]'. It does not contain any other ']]' strings.
local linkText = s:sub(3, -3)
-- Deal with nested links
if linkText:find('%[%[') then
return '[[' .. s:sub(3):gsub('%[%[.-%]%]$', p._delinkWikilink)
end
local titleArea, display = linkText:match('^(.-)|(.*)$')
-- Process links with display areas. Pipe tricks aren't processed here, as we need to know more about the link title first.
if display then
display = htmlDecode(display, true) -- decode HTML entities.
if titleArea == '' then
-- We are dealing with a reverse pipe trick.
if display:find('[|\n]') or s == '[[|]]' then
-- The link is invalid.
return s
else
return display
end
elseif display ~= '' then
-- We are dealing with a normal piped link.
return display
end
end
titleArea = titleArea or linkText
-- Decode percent-encoded and HTML-encoded characters.
titleArea = uriDecode(titleArea, 'PATH')
titleArea = htmlDecode(titleArea, true)
-- Find the fragment, if any.
local titleAreaNoFragment, fragment = titleArea:match('^(.-)#(.*)$')
titleAreaNoFragment = titleAreaNoFragment or titleArea
-- Check for bad characters.
if titleAreaNoFragment:find('[%[%]<>{}%%%c\n]') then
return s
end
-- Find the interwiki and the title. Actually, only the prefix before the first
-- comma is counted as the interwiki, so the "title" may contain another interwiki
-- prefix and/or a namespace name, but it's close enough for our purposes.
local interwiki, title = titleAreaNoFragment:match('^(.-):(.*)$') or ''
title = title or titleAreaNoFragment
-- Check for unescaped categories, interwikis, and files. If any are found,
-- return the blank string, as nothing would be displayed.
local ns = namespaces[interwiki]
if isKnownLanguageTag(interwiki)
or ns and (ns.id == 6 or ns.id == 14)
then
return ''
end
-- Remove the colon if the link is using the [[Help:Colon trick]].
titleArea = titleArea:match('^:(.*)$') or titleArea
-- Deal with links using the [[Help:Pipe trick]].
if display == '' then
if fragment then
-- Fragments in a pipe trick are invalid, so return the input string.
return s
end
-- Pipe tricks don't display interwikis, so we only need the title text here.
-- We need to remove parentheses and commas. Parentheses have priority.
local pipeTrickText = title:match('^(.-) ?%(.-%)$')
if pipeTrickText then
return pipeTrickText
else
-- If there are no parentheses, display only the text before the first comma.
pipeTrickText = title:match('(.-),.*$') or title
return pipeTrickText
end
end
-- If we haven't returned any text yet, display the title area.
return titleArea
end
function p._delinkURL(s)
-- Assume we have already delinked internal wikilinks, and that
-- we have been passed some text between two square brackets [foo].
-- If the text contains a line break it is not formatted as a URL, regardless of other content.
if s:find('\n') then
return s
end
-- Check if the text has a valid URL prefix and at least one valid URL character.
local valid_url_prefixes = {'//', 'http://', 'https://', 'ftp://', 'gopher://', 'mailto:', 'news:', 'irc://'}
local url_prefix
for i,v in ipairs(valid_url_prefixes) do
if s:find('^%[' .. v ..'[^"%s].*%]' ) then
url_prefix = v
break
end
end
-- Get display text
if not url_prefix then
-- Deal with nested links or send back original string.
return '[' .. s:sub(2):gsub('%[.-%]', p._delinkURL)
end
s = s:match('^%[' .. url_prefix .. '(.*)%]') -- Grab all of the text after the URL prefix and before the final square bracket.
s = s:match('^.-(["<> [].*)') or '' -- Grab all of the text after the first URL separator character ("<> ).
s = s:match('^%s*(%S.*)$') or '' -- If the separating character was a space, trim it off.
local s_decoded = mw.text.decode(s, true)
if s_decoded:find('%c') then
return s
else
return s_decoded
end
end
function p._delink(args)
local text = args[1] or ''
if whitespaceYesno(args.refs) then
-- Remove any [[Help:Strip markers]] representing ref tags. In most situations
-- this is not a good idea - only use it if you know what you are doing!
text = mw.ustring.gsub(text, 'UNIQ%w*%-ref%-%d*%-QINU', '')
end
if whitespaceYesno(args.comments) ~= false then
text = text:gsub('<!%-%-.-%-%->', '') -- Remove html comments.
end
if whitespaceYesno(args.wikilinks) ~= false then
text = text:gsub('%[%[.-%]%]', p._delinkWikilink) -- De-link wikilinks.
end
if whitespaceYesno(args.urls) ~= false then
text = text:gsub('%[.-%]', p._delinkURL) -- De-link URLs.
end
if whitespaceYesno(args.whitespace) ~= false then
-- Replace single new lines with a single space, but leave double new lines
-- and new lines only containing spaces or tabs before a second new line.
text = text:gsub('([^\n \t][ \t]*)\n([ \t]*[^\n \t])', '%1 %2')
text = text:gsub('[ \t]+', ' ') -- Remove extra tabs and spaces.
end
return text
end
function p.delink(frame)
local args = frame:getParent().args
return p._delink(args)
end
return p