Content deleted Content added
attempt to simplify the p._delinkWikilink function |
Synced (rip one testcase but whatever), dramatically improved efficiency of one of the regex patterns |
||
| (24 intermediate revisions by 11 users not shown) | |||
Line 1:
-- This module de-links most wikitext.
require("strict")
local p = {}
local getArgs
local function
if s:match("^%[%[|.*[|\n]") then -- Check for newlines or multiple pipes.
return s
end
return
end
local function
-- We need to deal with colons, brackets, and commas, per [[Help:Pipe trick]].
-- First, remove the text before the first colon, if any.
if s:match(":") then
s = s:match("%[%[.-:(.*)|%]%]")
-- If there are no colons, grab all of the text apart from the square brackets and the pipe.
else
s = s:match("%[%[(.*)|%]%]")
end
-- Next up, brackets and commas.
if s:match("%(.-%)$") then -- Brackets trump commas.
s = s:match("(.-) ?%(.-%)$")
elseif s:match(",") then -- If there are no brackets, display only the text before the first comma.
s = s:match("[^,]*")
end
return s
end
-- Return wikilink target |wikilinks=target
local function getDelinkedTarget(s)
local result = s
-- Deal with the reverse pipe trick.
if result:match("%[%[|") then
return delinkReversePipeTrick(result)
end
result = mw.uri.decode(result, "PATH") -- decode percent-encoded entities. Leave underscores and plus signs.
result = mw.text.decode(result, true) -- decode HTML entities.
-- Check for bad titles. To do this we need to find the
-- title area of the link, i.e. the part before any pipes.
local target_area
if result:match("|") then -- Find if we're dealing with a piped link.
target_area = result:match("^%[%[(.-)|.*%]%]")
else
target_area = result:match("^%[%[(.-)%]%]")
end
-- Check for bad characters.
if mw.ustring.match(target_area, "[%[%]<>{}%%%c\n]") and mw.ustring.match(target_area, "[%[%]<>{}%%%c\n]") ~= "?" then
return s
end
return target_area
end
local function getDelinkedLabel(s)
local result = s
-- Deal with the reverse pipe trick.
return delinkReversePipeTrick(result)
end
result = mw.uri.decode(result, "PATH") -- decode percent-encoded entities. Leave underscores and plus signs.
result = mw.text.decode(result, true) -- decode HTML entities.
-- Check for bad titles. To do this we need to find the
-- title area of the link, i.e. the part before any pipes.
local target_area
if result:match("|") then -- Find if we're dealing with a piped link.
target_area = result:match("^%[%[(.-)|.*%]%]")
else
target_area = result:match("^%[%[(.-)%]%]")
end
-- Check for bad characters.
if
return s
end
-- Check for categories, interwikis, and files.
local colon_prefix = result:match("%[%[(.-):.*%]%]") or "" -- Get the text before the first colon.
local ns = mw.site.namespaces[colon_prefix] -- see if this is a known namespace
if mw.language.isKnownLanguageTag(colon_prefix) or (ns and (ns.canonicalName == "File" or ns.canonicalName == "Category")) then
return ""
end
-- Remove the colon if the link is using the [[Help:Colon trick]].
result = "[[" .. result:match("%[%[:(.*%]%])")
end
-- Deal with links using the [[Help:Pipe trick]].
if mw.ustring.match(result, "^%[%[[^|]*|%]%]") then
return delinkPipeTrick(result)
end
-- Find the display area of the wikilink
if result:match("|") then -- Find if we're dealing with a piped link.
result = result:match("^%[%[.-|(.+)%]%]")
-- Remove new lines from the display of multiline piped links,
-- where the pipe is before the first new line.
result = result:gsub("\n", "")
result = result:match("^%[%[(.-)%]%]")
end
return result
end
local function
-- 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:
return s
end
-- Check if the text has a valid URL prefix and at least one valid URL character.
local valid_url_prefixes = {
local url_prefix
for
if
url_prefix = v
break
end
end
-- Get display text
if not url_prefix then
return s
end
s = s:match(
s = s:match('^.-(["<>
s =
local s_decoded = mw.text.decode(s, true)
if
return s
end
return s_decoded
end
local function delinkLinkClass(text, pattern, delinkFunction)
if type(text) ~= "string" then
error("Attempt to de-link non-string input.", 2)
end
if type(pattern) ~= "string" or mw.ustring.sub(pattern, 1, 1) ~= "^" then
error('Invalid pattern detected. Patterns must begin with "^".', 2)
end
-- Iterate over the text string, and replace any matched text. using the
-- delink function. We need to iterate character by character rather
-- than just use gsub, otherwise nested links aren't detected properly.
local result = ""
while text ~= "" do
-- Replace text using one iteration of gsub.
text = mw.ustring.gsub(text, pattern, delinkFunction, 1)
-- Append the left-most character to the result string.
result = result .. mw.ustring.sub(text, 1, 1)
text = mw.ustring.sub(text, 2, -1)
end
return result
end
function p._delink(args)
local text = args[1] or
if
-- 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,
end
if
text = text:gsub(
end
if args.wikilinks ~= "no" and args.wikilinks ~= "target" then
-- De-link wikilinks and return the label portion of the wikilink.
text = delinkLinkClass(text, "^%[%[.-%]%]", getDelinkedLabel)
elseif args.wikilinks == "target" then
-- De-link wikilinks and return the target portions of the wikilink.
text = delinkLinkClass(text, "^%[%[.-%]%]", getDelinkedTarget)
end
if
text =
end
if
-- 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 = text:gsub(
end
return text
Line 169 ⟶ 204:
function p.delink(frame)
if not getArgs then
getArgs = require('Module:Arguments').getArgs
end
return p._delink(getArgs(frame, {wrappers = 'Template:Delink'}))
end
| |||