Difference between revisions of "Module:URL"

From annadreambrush.com/wiki
Jump to navigation Jump to search
imported>Toohool
(Lua implementation of {{URL}})
 
m (1 revision imported)
 
(27 intermediate revisions by 8 users not shown)
Line 1: Line 1:
 +
--
 +
-- This module implements {{URL}}
 +
--
 +
-- See unit tests at [[Module:URL/testcases]]
 +
 
local p = {}
 
local p = {}
 
   
 
   
function trim(s)
+
local function safeUri(s)
    if s == nil then return '' end
+
local success, uri = pcall(function()
    return (s:gsub("^%s*(.-)%s*$", "%1"))
+
return mw.uri.new(s)
 +
end)
 +
if success then
 +
return uri
 +
end
 
end
 
end
  
function startsWith(s, prefix)
+
local function extractUrl(args)
    return (s:find(prefix, 1, true)) == 1
+
for name, val in pairs(args) do
 +
if name ~= 2 and name ~= "msg" then
 +
local url = name .. "=" .. val;
 +
url = mw.ustring.gsub(url, '^[Hh][Tt][Tt][Pp]([Ss]?):(/?)([^/])', 'http%1://%3')
 +
local uri = safeUri(url);
 +
if uri and uri.host then
 +
return url
 +
end
 +
end
 +
end
 
end
 
end
  
function split(s, sep) -- string split function for a single-character separator
+
function p._url(url, text, msg)
    local ret = {}
+
url = mw.text.trim(url or '')
    local n = 1
+
text = mw.text.trim(text or '')
    for w in s:gmatch("([^" .. sep .. "]*)") do
+
local nomsg = (msg or ''):sub(1,1):lower() == "n" or msg == 'false' -- boolean: true if msg is "false" or starts with n or N
        ret[n] = ret[n] or w  -- only set once (so the blank after a string is ignored)
+
 
        if w == '' then n = n + 1 end -- step forwards on a blank but not a string
+
if url == '' then
    end
+
if text == '' then
    return ret
+
if nomsg then
end
+
return nil
 +
else
 +
return mw.getCurrentFrame():expandTemplate{ title = 'tlx', args = { 'URL', "''example.com''", "''optional display text''" } }
 +
end
 +
else
 +
return text
 +
end
 +
end
 +
 
 +
-- If the URL contains any unencoded spaces, encode them, because MediaWiki will otherwise interpret a space as the end of the URL.
 +
url = mw.ustring.gsub(url, '%s', function(s) return mw.uri.encode(s, 'PATH') end)
 +
 
 +
-- If there is an empty query string or fragment id, remove it as it will cause mw.uri.new to throw an error
 +
url = mw.ustring.gsub(url, '#$', '')
 +
url = mw.ustring.gsub(url, '%?$', '')
 +
 
 +
-- If it's an HTTP[S] URL without the double slash, fix it.
 +
url = mw.ustring.gsub(url, '^[Hh][Tt][Tt][Pp]([Ss]?):(/?)([^/])', 'http%1://%3')
 +
 
 +
-- Handle URLs from Wikidata of the format http://
 +
url = mw.ustring.gsub(url, '^[Hh][Tt][Tt][Pp]([Ss]?)://', 'http%1://')
 +
 
 +
local uri = safeUri(url)
 +
 
 +
-- Handle URL's without a protocol and URL's that are protocol-relative,
 +
-- e.g. www.example.com/foo or www.example.com:8080/foo, and //www.example.com/foo
 +
if uri and (not uri.protocol or (uri.protocol and not uri.host)) and url:sub(1, 2) ~= '//' then
 +
url = 'http://' .. url
 +
uri = safeUri(url)
 +
end
 +
 
 +
if text == '' then
 +
if uri then
 +
if uri.path == '/' then uri.path = '' end
 +
 
 +
local port = ''
 +
if uri.port then port = ':' .. uri.port end
  
function rest(tbl, start) -- get the elements of the table beginning at the specified starting index
+
text = mw.ustring.lower(uri.host or '') .. port .. (uri.relativePath or '')
    local ret = {}
 
    if start > #tbl then
 
        return ret
 
    end
 
    for i = start, #tbl do
 
        ret[i - start + 1] = tbl[i]
 
    end
 
    return ret
 
end
 
  
function p._url(url, text)
+
-- Add <wbr> before _/.-# sequences
    local lcUrl = url:lower()
+
text = mw.ustring.gsub(text,"(/+)","<wbr/>%1")      -- This entry MUST be the first. "<wbr/>" has a "/" in it, you know.
    if not startsWith(lcUrl, 'http://') and not startsWith(lcUrl, 'https://') and not startsWith(lcUrl, 'ftp://') then
+
text = mw.ustring.gsub(text,"(%.+)","<wbr/>%1")
        url = 'http://' .. url
+
-- text = mw.ustring.gsub(text,"(%-+)","<wbr/>%1") -- DISABLED for now
    end
+
text = mw.ustring.gsub(text,"(%#+)","<wbr/>%1")
   
+
text = mw.ustring.gsub(text,"(_+)","<wbr/>%1")
    if text == '' then
+
else -- URL is badly-formed, so just display whatever was passed in
        local comps = split(url, '/')
+
text = url
        local host = comps[3]:lower()
+
end
        local path = table.concat(rest(comps, 4), '/')
+
end
  
        text = host
+
return mw.ustring.format('<span class="url">[%s %s]</span>', url, text)
        if path ~= '' then
 
            text = text .. '/' .. path
 
        end
 
    end
 
   
 
    return string.format('<span class="url">[%s %s]</span>', url, text)
 
 
end
 
end
  
 +
--[[
 +
The main entry point for calling from Template:URL.
 +
--]]
 
function p.url(frame)
 
function p.url(frame)
    local url = trim(frame.args[1])
+
local templateArgs = frame.args
    local text = trim(frame.args[2])
+
local parentArgs = frame:getParent().args
 +
local url = templateArgs[1] or parentArgs[1]
 +
local text = templateArgs[2] or parentArgs[2] or ''
 +
local msg = templateArgs.msg or parentArgs.msg or ''
 +
url = url or extractUrl(templateArgs) or extractUrl(parentArgs) or ''
 +
return p._url(url, text, msg)
 +
end
  
    if url == '' then
+
--[[
        return frame:expandTemplate{ title = 'tlx', args = { 'URL', "''example.com''", "''optional display text''" } }
+
The entry point for calling from the forked Template:URL2.
    end
+
This function returns no message by default.
   
+
It strips out wiki-link markup, html tags, and everything after a space.
    return p._url(url, text)
+
--]]
 +
function p.url2(frame)
 +
local templateArgs = frame.args
 +
local parentArgs = frame:getParent().args
 +
local url = templateArgs[1] or parentArgs[1]
 +
local text = templateArgs[2] or parentArgs[2] or ''
 +
-- default to no message
 +
local msg = templateArgs.msg or parentArgs.msg or 'no'
 +
url = url or extractUrl(templateArgs) or extractUrl(parentArgs) or ''
 +
-- strip out html tags and [ ] from url
 +
url = (url or ''):gsub("<[^>]*>", ""):gsub("[%[%]]", "")
 +
-- truncate anything after a space
 +
url = url:gsub("%%20", " "):gsub(" .*", "")
 +
return p._url(url, text, msg)
 
end
 
end
  
 
return p
 
return p

Latest revision as of 18:56, 9 March 2020

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

--
-- This module implements {{URL}}
--
-- See unit tests at [[Module:URL/testcases]]

local p = {}
 
local function safeUri(s)
	local success, uri = pcall(function()
		return mw.uri.new(s)
	end)
	if success then
		return uri
	end
end

local function extractUrl(args)
	for name, val in pairs(args) do
		if name ~= 2 and name ~= "msg" then
			local url = name .. "=" .. val;
			url = mw.ustring.gsub(url, '^[Hh][Tt][Tt][Pp]([Ss]?):(/?)([^/])', 'http%1://%3')
			local uri = safeUri(url);
			if uri and uri.host then
				return url
			end
		end
	end
end

function p._url(url, text, msg)
	url = mw.text.trim(url or '')
	text = mw.text.trim(text or '')
	local nomsg = (msg or ''):sub(1,1):lower() == "n" or msg == 'false' -- boolean: true if msg is "false" or starts with n or N

	if url == '' then
		if text == '' then
			if nomsg then
				return nil
			else
				return mw.getCurrentFrame():expandTemplate{ title = 'tlx', args = { 'URL', "''example.com''", "''optional display text''" } }
			end
		else
			return text
		end
	end

	-- If the URL contains any unencoded spaces, encode them, because MediaWiki will otherwise interpret a space as the end of the URL.
	url = mw.ustring.gsub(url, '%s', function(s) return mw.uri.encode(s, 'PATH') end)

	-- If there is an empty query string or fragment id, remove it as it will cause mw.uri.new to throw an error
	url = mw.ustring.gsub(url, '#$', '')
	url = mw.ustring.gsub(url, '%?$', '')

	-- If it's an HTTP[S] URL without the double slash, fix it.
	url = mw.ustring.gsub(url, '^[Hh][Tt][Tt][Pp]([Ss]?):(/?)([^/])', 'http%1://%3')

	-- Handle URLs from Wikidata of the format http&#58;//
	url = mw.ustring.gsub(url, '^[Hh][Tt][Tt][Pp]([Ss]?)&#58;//', 'http%1://')

	local uri = safeUri(url)

	-- Handle URL's without a protocol and URL's that are protocol-relative, 
	-- e.g. www.example.com/foo or www.example.com:8080/foo, and //www.example.com/foo
	if uri and (not uri.protocol or (uri.protocol and not uri.host)) and url:sub(1, 2) ~= '//' then
		url = 'http://' .. url
		uri = safeUri(url)
	end

	if text == '' then
		if uri then
			if uri.path == '/' then uri.path = '' end

			local port = ''
			if uri.port then port = ':' .. uri.port end

			text = mw.ustring.lower(uri.host or '') .. port .. (uri.relativePath or '')

			-- Add <wbr> before _/.-# sequences
			text = mw.ustring.gsub(text,"(/+)","<wbr/>%1")      -- This entry MUST be the first. "<wbr/>" has a "/" in it, you know.
			text = mw.ustring.gsub(text,"(%.+)","<wbr/>%1")
			-- text = mw.ustring.gsub(text,"(%-+)","<wbr/>%1") 	-- DISABLED for now
			text = mw.ustring.gsub(text,"(%#+)","<wbr/>%1")
			text = mw.ustring.gsub(text,"(_+)","<wbr/>%1")
		else -- URL is badly-formed, so just display whatever was passed in
			text = url
		end
	end

	return mw.ustring.format('<span class="url">[%s %s]</span>', url, text)
end

--[[
The main entry point for calling from Template:URL.
--]]
function p.url(frame)
	local templateArgs = frame.args
	local parentArgs = frame:getParent().args
	local url = templateArgs[1] or parentArgs[1]
	local text = templateArgs[2] or parentArgs[2] or ''
	local msg = templateArgs.msg or parentArgs.msg or ''
	url = url or extractUrl(templateArgs) or extractUrl(parentArgs) or ''
	return p._url(url, text, msg)
end

--[[
The entry point for calling from the forked Template:URL2.
This function returns no message by default.
It strips out wiki-link markup, html tags, and everything after a space.
--]]
function p.url2(frame)
	local templateArgs = frame.args
	local parentArgs = frame:getParent().args
	local url = templateArgs[1] or parentArgs[1]
	local text = templateArgs[2] or parentArgs[2] or ''
	-- default to no message
	local msg = templateArgs.msg or parentArgs.msg or 'no'
	url = url or extractUrl(templateArgs) or extractUrl(parentArgs) or ''
	-- strip out html tags and [ ] from url
	url = (url or ''):gsub("<[^>]*>", ""):gsub("[%[%]]", "")
	-- truncate anything after a space
	url = url:gsub("%%20", " "):gsub(" .*", "")
	return p._url(url, text, msg)
end

return p