Module:NUMBEROFSECTIONS: Difference between revisions

Content deleted Content added
Inline
Revision as of 10:43, 4 May 2020

local p = {}

local function trimArg(arg)
	if arg == "" or arg == nil then
		return nil
	else
		return mw.ustring.lower(mw.text.trim(arg))
	end
end

local function tableLength(T)
	local count = 0
	for _ in pairs(T) do count = count + 1 end
	return count
end

-- Credit: http://lua-users.org/wiki/FormattingNumbers
local function comma_value(amount)
	local formatted = amount
	while true do
		local k
		formatted, k = string.gsub(formatted, "^(-?%d+)(%d%d%d)", '%1,%2')
		if k == 0 then
			break
		end
	end
	return formatted
end

function p.numberof(frame)
	local pframe = frame:getParent()
	local args = pframe.args
	local tname = "NUMBEROF"
	local result = -1
	local data = mw.ext.data.get("Wikipedia_statistics/data.tab") -- https://commons.wikimedia.org/wiki/Data:Wikipedia_statistics/data.tab

	local action = trimArg(args[1]) -- "ARTICLES, PAGES, EDITS, USERS, ACTIVEUSERS, ADMINS or FILES"
	if not action then
		error("Parameter 1 is missing. See [[Template:" .. tname .. "]] documentation")
	end
	local site = trimArg(args[2]) -- "en" or "en.wikipedia" or "en.wikiquote" etc..
	if not site then
		error("Parameter 2 is missing. See [[Template:" .. tname .. "]] documentation")
	end
	local site2split = mw.text.split(site, "%.")
	if tableLength(site2split) == 1 then
		site = site .. ".wikipedia"
	end
	local comma = trimArg(args[3]) -- If set to "N" use commas in output

	for k,v in pairs(data['data']) do
		if site == mw.ustring.lower(v[1]) then
			if action == "activeusers" then
				result = v[2]
			elseif action == "admins" then
				result = v[3]
			elseif action == "articles" then
				result = v[4]
			elseif action == "edits" then
				result = v[5]
			elseif action == "files" then
				result = v[6]
			elseif action == "pages" then
				result = v[7]
			elseif action == "users" then
				result = v[8]
			end
		end
	end
	if result > -1 and comma then
		result = comma_value(result)
	end
	return result
end

-- Unescape functionality grabbed from https://stackoverflow.com/a/14899740/1832568
local function unescape(str)
	str = string.gsub(str, '&#(%d+);', string.char)
	str = string.gsub(str, '&#x(%d+);', function(n) return string.char(tonumber(n, 16)) end)
	return str
end

-- Counting function accepting a string haystack and table of needles
local function count(haystack, needles)
	local number = 0
	-- While we have needles to look for
	for index, needle in ipairs(needles) do
		-- find them all in our haystack
		for m in string.gmatch(haystack, needle) do
			number = number + 1
		end
	end
	return number
end

-- Function takes any number of # delimited page names and section level numbers
function p.sections(frame)
	local total = 0
	local needles = {}
	local haystack = ''
	-- Separate page names from # delimited string into table
	local pages = mw.text.split(unescape(frame.args[1]), '%s?#%s?')
	-- Separate whitespace delimited section level numbers into table
	local levels = mw.text.split(frame.args['level'], '%s*')
	-- Iterate through levels
	for level in mw.text.gsplit(table.concat(levels), '') do
		-- and add the level needle to needles
		needles[#needles + 1] = '\n'..string.rep('=', tonumber(level))..'[^=]'
	end
	-- For each page name in pages
	for index, page in ipairs(pages) do
		-- create a haystack to search from the page content
		haystack = mw.title.new(page):getContent()
		-- If we've requested the content of a legitimate page
		if haystack then
			--[[ pass the raw markup and needles to count
				 and add the return to total ]]
			total = total + count('\n' .. haystack, needles)
		end
	end
	--[[ then return how many sections of the required level
		 are in all the pages passed ]]
	return total
end

return p
Revision as of 10:43, 4 May 2020

Usage
