Module:Sandbox/urastats

local p = {}

function p.sums()

	local function add_days_to_date(date_str, days)
    	local y, m, d = date_str:match("(%d+)%-(%d+)%-(%d+)")
    	local ts = os.time{ year = y, month = m, day = d, hour = 12 }
    	local new_ts = ts + days * 24 * 60 * 60
    	return os.date("%Y-%m-%d", new_ts)
	end

    -- Initialise language
    local lang = mw.language.getContentLanguage()
    
    -- Load the data array
    -- https://commons.wikimedia.org/wiki/Data:URAstats_Data.tab
    local t = mw.ext.data.get('URAstats_Data.tab')
    local rows = t.data

    -- get the necessary rows
	local day0 = rows[#rows]    -- today
	local day1 = rows[#rows-1]  -- yesterday
	local dayz = rows[1]        -- when records began

    -- Extract numeric fields (columns 2–6)
    local a0 = tonumber(day0[2]) or 0
    local b0 = tonumber(day0[3]) or 0
    local c0 = tonumber(day0[4]) or 0
    local d0 = tonumber(day0[5]) or 0
    local e0 = tonumber(day0[6]) or 0

    local a1 = tonumber(day1[2]) or 0
    local b1 = tonumber(day1[3]) or 0
    local c1 = tonumber(day1[4]) or 0
    local d1 = tonumber(day1[5]) or 0
    local e1 = tonumber(day1[6]) or 0
    
    local az = tonumber(dayz[2]) or 0
    local bz = tonumber(dayz[3]) or 0
    local cz = tonumber(dayz[4]) or 0
    local dz = tonumber(dayz[5]) or 0
    local ez = tonumber(dayz[6]) or 0
    
    -- Compute sums
    local def0 = a0 + b0 + c0
    local prob0 = d0 + e0

    local def1 = a1 + b1 + c1
    local prob1 = d1 + e1

    local defz = az + bz + cz
    local probz = dz + ez

	-- Compute rate
	local articles_per_day = math.floor((((defz + probz) - (def0 + prob0)) / #rows))
	local days_to_clear_backlog = math.ceil(((def0 + prob0) / articles_per_day))

    -- Retrieve dates and format them nicely
	local date0 = lang:formatDate( "j F Y", day0[1], nil )
	local finish_date = lang:formatDate( "j F Y", add_days_to_date(day0[1], days_to_clear_backlog), nil)

	-- Compose return string
    return string.format(
    	"At the beginning of the day (UTC) on %s there were:\n" ..
        "* %s tagged unreferenced articles, a change of %s since the day before\n" ..
        "* %s probably unreferenced articles, a change of %s since the day before\n" ..
        "* That is, a '''total backlog of %s articles''' still to deal with, a change of %s since the day before\n" ..
        "At the current [[arithmetic mean|average]] rate of '''%s articles per day''', " ..
        "it will take '''%s days to clear the backlog''', " ..
        "finishing on '''%s'''.",
        date0, 
        lang:formatNum(def0), lang:formatNum(def0 - def1), 
        lang:formatNum(prob0), lang:formatNum(prob0 - prob1), 
        lang:formatNum(def0 + prob0), lang:formatNum((def0 + prob0) - (def1 + prob1)),
        lang:formatNum(articles_per_day), lang:formatNum(days_to_clear_backlog), finish_date
    )
end

return p