Difference between revisions of "Module:Delink/sandbox"

From blackwiki
Jump to navigation Jump to search
blackwiki>Darklama
(try another way, most cases covered)
m (24 revisions imported)
 
(22 intermediate revisions by 7 users not shown)
Line 1: Line 1:
 
-- This module de-links most wikitext.
 
-- This module de-links most wikitext.
  
p = {};
+
require('Module:No globals')
  
local function internalLink(s)
+
local p = {}
     return (mw.ustring.gsub( s, "^([^|]*)(|?)(.*)$", function(l, p, t)
+
 
        -- check for malformed links and text
+
local getArgs
         if mw.ustring.match( l, "[<>{}%c\n]" ) or mw.ustring.match( t, "\n" ) then
+
 
            return s;
+
local function delinkReversePipeTrick(s)
        end
+
     if s:match("^%[%[|.*[|\n]") then -- Check for newlines or multiple pipes.
       
+
        return s
        -- remove whitespace from begining and end of link and text
+
    else
         l = mw.text.trim(l);
+
        return s:match("%[%[|(.*)%]%]")
         t = mw.text.trim(t);
+
    end
       
+
end
        local file = false;
+
 
       
+
local function delinkPipeTrick(s)
        if l == "" then
+
    local linkarea, display = "", ""
            return s;
+
    -- We need to deal with colons, brackets, and commas, per [[Help:Pipe trick]].
        elseif mw.ustring.match( l, "^:" ) then
+
   
            l = mw.ustring.gsub( l, "^:+", "" );
+
    -- First, remove the text before the first colon, if any.
         elseif mw.ustring.match( l, "^[Cc][Aa][Tt][Ee][Gg][Oo][Rr][Yy]:" ) then
+
    if s:match(":") then
            return "";
+
         s = s:match("%[%[.-:(.*)|%]%]")
         elseif mw.ustring.match( l, "^[Ff][Ii][Ll][Ee]:" ) then
+
    -- If there are no colons, grab all of the text apart from the square brackets and the pipe.
            file = true;
+
    else
        elseif mw.ustring.match( l, "^[Ii][Mm][Aa][Gg][Ee]:" ) then
+
        s = s:match("%[%[(.*)|%]%]")
            file = true;
+
    end
        end
+
   
       
+
    -- Next up, brackets and commas.
        if file == true then
+
    if s:match("%(.-%)$") then -- Brackets trump commas.
            return "";
+
         s = s:match("(.-) ?%(.-%)$")
        elseif p == "|" and t ~= "" then
+
    elseif s:match(",") then -- If there are no brackets, display only the text before the first comma.
            return t;
+
         s = s:match("(.-),.*$")
        else
+
    end
            l = mw.ustring.gsub( l, "(#.+)$", function(h)
+
    return s
                return mw.ustring.gsub( h, "%.([0-9A-Fa-f][0-9A-Fa-f])", "%%%1" );
+
end
            end);
+
 
           
+
local function delinkWikilink(s)
            if p == "" and t == "" then
+
    local result = s
                return mw.uri.decode(l, "PATH");
+
    -- Deal with the reverse pipe trick.
            end
+
    if result:match("%[%[|") then
           
+
        return delinkReversePipeTrick(result)
            -- Need to transform text based on link per [[Help:Pipe trick]].
+
    end
            t = mw.ustring.match( l, "#(.+)$" );
+
   
            if t ~= nil then
+
    result = mw.uri.decode(result, "PATH") -- decode percent-encoded entities. Leave underscores and plus signs.
                return mw.uri.decode(t, "PATH");
+
    result = mw.text.decode(result, true) -- decode HTML entities.
            end
+
   
           
+
    -- Check for bad titles. To do this we need to find the
            t = mw.ustring.gsub( l, "^([^#]*)#$", "%1", 1 );
+
    -- title area of the link, i.e. the part before any pipes.
            t = mw.ustring.gsub( t, "(.-) ?%b()$", "%1", 1 );
+
    local titlearea
            t = mw.ustring.gsub( t, "^(.-),.*$", "%1", 1 );
+
    if result:match("|") then -- Find if we're dealing with a piped link.
            t = mw.uri.decode(t, "PATH");
+
         titlearea = result:match("^%[%[(.-)|.*%]%]")
            return t;
+
    else
        end
+
         titlearea = result:match("^%[%[(.-)%]%]")
     end, 1));
+
    end
 +
    -- Check for bad characters.
 +
    if mw.ustring.match(titlearea, "[%[%]<>{}%%%c\n]") then
 +
        return s
 +
    end
 +
   
 +
    -- Check for categories, interwikis, and files.
 +
    local colonprefix = result:match("%[%[(.-):.*%]%]") or "" -- Get the text before the first colon.
 +
    local ns = mw.site.namespaces[colonprefix] -- see if this is a known namespace
 +
    if mw.language.isKnownLanguageTag(colonprefix)
 +
    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]].
 +
    if result:match("%[%[:") then
 +
        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", "")
 +
    else
 +
        result = result:match("^%[%[(.-)%]%]")
 +
    end
 +
 
 +
     return result
 
end
 
end
  
local function externalLink(s)
+
local function 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 the text contains a line break it is not formatted as a URL, regardless of other content.
     if mw.ustring.match(s, "\n") then
+
     if s:match("\n") then
         return s;
+
         return s
 
     end
 
     end
 
      
 
      
     return (mw.ustring.gsub(mw.text.trim(s), "^([^ \t]*)[ \t]*(.*)$", function(l, t)
+
     -- Check if the text has a valid URL prefix and at least one valid URL character.
        t = mw.text.trim(t);
+
    local valid_url_prefixes = {"//", "http://", "https://", "ftp://", "gopher://", "mailto:", "news:", "irc://"}
       
+
    local url_prefix
         if t == "" then
+
    for i,v in ipairs(valid_url_prefixes) do
            return s;
+
         if mw.ustring.match(s, '^%[' .. v ..'[^"%s].*%]' ) then
        elseif mw.ustring.sub( l, 1, 2 ) == "//" then
+
             url_prefix = v
             l = "http:" .. l;
+
            break
 
         end
 
         end
          
+
    end
        l = mw.uri.new( l );
+
   
       
+
    -- Get display text
        if l.protocol == nil or l.host == nil or mw.ustring.match( l.host, "%." ) == nil then
+
    if not url_prefix then
            return s;
+
         return s
        else
+
    end
            return t;
+
    s = s:match("^%[" .. url_prefix .. "(.*)%]") -- Grab all of the text after the URL prefix and before the final square bracket.
        end
+
    s = s:match('^.-(["<> ].*)') or "" -- Grab all of the text after the first URL separator character ("<> ).
     end, 1));
+
    s = mw.ustring.match(s, "^%s*(%S.*)$") or "" -- If the separating character was a space, trim it off.
 +
   
 +
    local s_decoded = mw.text.decode(s, true)
 +
    if mw.ustring.match(s_decoded, "%c") then
 +
        return s
 +
    else  
 +
        return s_decoded
 +
     end
 
end
 
end
  
local function delinkLinkClass(s, pattern, delinker)
+
local function delinkLinkClass(s, pattern, delinkFunction)
     if type(s) ~= "string" then
+
     if not type(s) == "string" then
         error("Attempt to de-link non-string input.", 2);
+
         error("Attempt to de-link non-string input.", 2)
     elseif type(pattern) ~= "string" then
+
     end
         error('Patterns must begin a string.', 2);
+
    if not ( type(pattern) == "string" and mw.ustring.sub(pattern, 1, 1) == "^" ) then
 +
         error('Invalid pattern detected. Patterns must begin with "^".', 2)
 
     end
 
     end
 
     -- Iterate over the text string, and replace any matched text. using the  
 
     -- Iterate over the text string, and replace any matched text. using the  
     -- delink function. We need to iterate one by one otherwise  nested links
+
     -- delink function. We need to iterate character by character rather
     -- aren't detected properly.
+
     -- than just use gsub, otherwise nested links aren't detected properly.
     local result, f, e, b, a = {};
+
     local result = ""
      
+
     while s ~= '' do
    repeat
+
         -- Replace text using one iteration of gsub.
         f, e = mw.ustring.find(s, pattern);
+
         s = mw.ustring.gsub(s, pattern, delinkFunction, 1)
          
+
         -- Append the left-most character to the result string.
        if f == nil then
+
         result = result .. mw.ustring.sub(s, 1, 1)
            table.insert( result, s );
+
         s = mw.ustring.sub(s, 2, -1)
            break;
+
     end
        end
+
     return result
       
 
        if f > 1 then
 
            b = mw.ustring.gsub( mw.ustring.sub(s, 1, f-1), "(%[+)$", mw.text.nowiki, 1 );
 
         end
 
          
 
        if e ~= nil then
 
            a = mw.ustring.gsub( mw.ustring.sub(s, e+1), "^(%]+)", mw.text.nowiki, 1 );
 
        else
 
            a = "";
 
        end
 
       
 
        -- Replace text using one iteration.
 
         s = mw.ustring.gsub( mw.ustring.sub(s, f, e), pattern, delinker, 1 );
 
       
 
        -- Append to result and start after last result.
 
        table.insert( result, b );
 
        table.insert( result, s );
 
        s = a;
 
    until f == nil or s == "";
 
      
 
     return table.concat( result );
 
 
end
 
end
  
local function _delink(args)
+
function p._delink(args)
 
     local text = args[1] or ""
 
     local text = args[1] or ""
 
     if args.refs == "yes" then
 
     if args.refs == "yes" then
Line 129: Line 153:
 
         text = mw.ustring.gsub(text, "UNIQ%w*%-ref%-%d*%-QINU", "")
 
         text = mw.ustring.gsub(text, "UNIQ%w*%-ref%-%d*%-QINU", "")
 
     end
 
     end
     if args.comments ~= "no" then
+
     if not (args.comments == "no") then
         text = mw.ustring.gsub(text, "<!%-%-.-%-%->", "") -- Remove html comments.
+
         text = text:gsub("<!%-%-.-%-%->", "") -- Remove html comments.
 
     end
 
     end
     if args.wikilinks ~= "no" then
+
     if not (args.wikilinks == "no") then
         text = delinkLinkClass(text, "%[%[([^%[%]]*)%]%]", internalLink) -- De-link wikilinks.
+
         text = delinkLinkClass(text, "^%[%[.-%]%]", delinkWikilink) -- De-link wikilinks.
 
     end
 
     end
     if args.urls ~= "no" then
+
     if not (args.urls == "no") then
         text = delinkLinkClass(text, "%[([^%[%]]*)%]", externalLink) -- De-link URLs.
+
         text = delinkLinkClass(text, "^%[.-%]", delinkURL) -- De-link URLs.
 
     end
 
     end
     if args.whitespace ~= "no" then
+
     if not (args.whitespace == "no") then
 +
        text = text:gsub("<br[ \t]*/?>", "\n") -- Replace HTML line breaks with newlines.
 
         -- Replace single new lines with a single space, but leave double new lines
 
         -- 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.
 
         -- and new lines only containing spaces or tabs before a second new line.
 
         text = mw.ustring.gsub(text, "([^\n \t][ \t]*)\n([ \t]*[^\n \t])", "%1 %2")
 
         text = mw.ustring.gsub(text, "([^\n \t][ \t]*)\n([ \t]*[^\n \t])", "%1 %2")
         text = mw.ustring.gsub(text, "[ \t]+", " ") -- Remove extra tabs and spaces.
+
         text = text:gsub("[ \t]+", " ") -- Remove extra tabs and spaces.
 +
    end
 +
    if args.nowiki == "yes" then
 +
    text = '<nowiki />' .. text .. 'foo'
 
     end
 
     end
 
     return text
 
     return text
Line 148: Line 176:
  
 
function p.delink(frame)
 
function p.delink(frame)
     local args
+
     if not getArgs then
    if frame == mw.getCurrentFrame() then
+
    getArgs = require('Module:Arguments').getArgs
        -- We're being called via #invoke. If the invoking template passed any args, use
 
        -- them. Otherwise, use the args that were passed into the template.
 
        for k, v in pairs(frame.args) do
 
            args = frame.args;
 
            break;
 
        end
 
        if args == nil then args = frame:getParent().args; end
 
    else
 
        -- We're being called from another module or from the debug console, so assume
 
        -- the args are passed in directly.
 
        args = frame;
 
 
     end
 
     end
   
+
     return p._delink(getArgs(frame, {wrappers = 'Template:Delink'}))
     return _delink(args);
 
 
end
 
end
  
 
return p
 
return p

Latest revision as of 16:03, 26 September 2020

Documentation for this module may be created at Module:Delink/sandbox/doc

-- This module de-links most wikitext.

require('Module:No globals')

local p = {}

local getArgs

local function delinkReversePipeTrick(s)
    if s:match("^%[%[|.*[|\n]") then -- Check for newlines or multiple pipes.
        return s
    else
        return s:match("%[%[|(.*)%]%]")
    end
end

local function delinkPipeTrick(s)
    local linkarea, display = "", ""
    -- 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

local function delinkWikilink(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 titlearea
    if result:match("|") then -- Find if we're dealing with a piped link.
        titlearea = result:match("^%[%[(.-)|.*%]%]")
    else
        titlearea = result:match("^%[%[(.-)%]%]")
    end
    -- Check for bad characters.
    if mw.ustring.match(titlearea, "[%[%]<>{}%%%c\n]") then
        return s
    end
    
    -- Check for categories, interwikis, and files.
    local colonprefix = result:match("%[%[(.-):.*%]%]") or "" -- Get the text before the first colon.
    local ns = mw.site.namespaces[colonprefix] -- see if this is a known namespace
    if mw.language.isKnownLanguageTag(colonprefix)
    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]].
    if result:match("%[%[:") then
        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", "")
    else
        result = result:match("^%[%[(.-)%]%]")
    end

    return result
end

local function 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:match("\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 mw.ustring.match(s, '^%[' .. v ..'[^"%s].*%]' ) then
            url_prefix = v
            break
        end
    end
    
    -- Get display text
    if not url_prefix then
        return s
    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 = mw.ustring.match(s, "^%s*(%S.*)$") or "" -- If the separating character was a space, trim it off.
    
    local s_decoded = mw.text.decode(s, true)
    if mw.ustring.match(s_decoded, "%c") then
        return s
    else    
        return s_decoded
    end
end

local function delinkLinkClass(s, pattern, delinkFunction)
    if not type(s) == "string" then
        error("Attempt to de-link non-string input.", 2)
    end
    if not ( type(pattern) == "string" and 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 s ~= '' do
        -- Replace text using one iteration of gsub.
        s = mw.ustring.gsub(s, pattern, delinkFunction, 1)
        -- Append the left-most character to the result string.
        result = result .. mw.ustring.sub(s, 1, 1)
        s = mw.ustring.sub(s, 2, -1)
    end
    return result
end

function p._delink(args)
    local text = args[1] or ""
    if args.refs == "yes" 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 not (args.comments == "no") then
        text = text:gsub("<!%-%-.-%-%->", "") -- Remove html comments.
    end
    if not (args.wikilinks == "no") then
        text = delinkLinkClass(text, "^%[%[.-%]%]", delinkWikilink) -- De-link wikilinks.
    end
    if not (args.urls == "no") then
        text = delinkLinkClass(text, "^%[.-%]", delinkURL) -- De-link URLs.
    end
    if not (args.whitespace == "no") then
        text = text:gsub("<br[ \t]*/?>", "\n") -- Replace HTML line breaks with newlines.
        -- 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 = mw.ustring.gsub(text, "([^\n \t][ \t]*)\n([ \t]*[^\n \t])", "%1 %2")
        text = text:gsub("[ \t]+", " ") -- Remove extra tabs and spaces.
    end
    if args.nowiki == "yes" then
    	text = '<nowiki />' .. text .. 'foo'
    end
    return text
end

function p.delink(frame)
    if not getArgs then
    	getArgs = require('Module:Arguments').getArgs
    end
    return p._delink(getArgs(frame, {wrappers = 'Template:Delink'}))
end

return p