Module:CIA World Factbook

local p = {}
local getArgs = require('Module:Arguments').getArgs

-- prefix of all World Factbook pages
local factbookPrefix = 'https://web.archive.org/web/%04d%02d%02d000000/https://www.cia.gov/the-world-factbook/'
local archiveURL = 'https://web.archive.org/web/20211205000000/https://www.cia.gov/the-world-factbook/about/archives/download/factbook-%d.zip'

-- Function to turn a string into a URL fragment appropriate for CIA website
local function parseFragment(s)
	if not s then
		return ''
	end
	s = mw.ustring.lower(s)
	s = mw.ustring.gsub(s,' ','-')
	s = mw.ustring.gsub(s,',','')
	return s
end

local monthMap = {january=1,february=2,march=3,april=4,may=5,june=6,
	              july=7,august=8,september=9,october=10,november=11,december=12,
	              jan=1,feb=2,mar=3,apr=4,may=5,jun=6,jul=7,aug=8,sep=9,oct=10,nov=11,dec=12}

local function parseDate(date)
	local year, month, day
	if not date then return nil, nil, nil end
	_, _, year, month, day = mw.ustring.find(date,"(20%d%d)-(%d+)-(%d+)")
	if year then return year, month, day end
	_, _, day, month, year = mw.ustring.find(date,"(%d+)%s+(%a+)%s+(20%d%d)")
	month = month and monthMap[mw.ustring.lower(month)]
	if year then return year, month, day end
	_, _, month, day, year = mw.ustring.find(date,"(%a+)%s(%d+),%s+(20%d%d)")
	month = month and monthMap[mw.ustring.lower(month)]
	if year then return year, month, day end
	_, _, year = mw.ustring.find(date,"(20%d%d)")
	if year then return year, nil, nil end
	return nil, nil, nil
end

local function archiveDate(args)
	local year, month, day = parseDate(args.date)
	if not args.year or args.year == year then return tonumber(year), month, day end
	return tonumber(args.year), nil, nil
end

-- Function to fill in factbook link:
-- Arguments:
--    args.country: topic of page (optional)
--    args.section: section of page (optional)
-- Returns:
--    link to World Factbook page about country, with section anchor
function p._country(args)
	local year, month, day = archiveDate(args)
	if not year or year > 2025 then
		year = 2025
		month = 12
		day = 5
	end
	if year < 2021 then
		return mw.ustring.format(archiveURL,year)
	end
	if not month then month = 12 end
	if not day then day = 5 end
	local prefix = mw.ustring.format(factbookPrefix,year,month,day)
	if not args.country then
		return prefix
	end
	local result = prefix..'countries/'..parseFragment(args.country)
	if args.section then
		return result..'/#'..parseFragment(args.section)
	end
	return result
end


function p.country(frame)
	local args = getArgs(frame)
	return p._country(args)
end

return p