LARGE NUMBER ADDER

-- The function get_name() should return a single string that is the name of the puzzle.
--
function get_name()
    return "LARGE NUMBER ADDER"
end

-- The function get_description() should return an array of strings, where each string is
-- a line of description for the puzzle. The text you return from get_description() will
-- be automatically formatted and wrapped to fit inside the puzzle information box.
--
function get_description()
    return { "LARGE NUMBERS ARE GIVEN WITH TWO INPUTS WITH UPPER AND LOWER THREE DIGITS OF THE NUMBER", "WRITE THE SUM OF TWO INPUT NUMBERS TO OUTPUTS AS LARGE NUMBERS" }
end

-- The function get_streams() should return an array of streams. Each stream is described
-- by an array with exactly four values: a STREAM_* value, a name, a position, and an array
-- of integer values between -999 and 999 inclusive.
--
-- STREAM_INPUT: An input stream containing up to 39 numerical values.
-- STREAM_OUTPUT: An output stream containing up to 39 numerical values.
-- STREAM_IMAGE: An image output stream, containing exactly 30*18 numerical values between 0
--               and 4, representing the full set of "pixels" for the target image.
--
-- NOTE: Arrays in Lua are implemented as tables (dictionaries) with integer keys that start
--       at 1 by convention. The sample code below creates an input array of 39 random values
--       and an output array that doubles all of the input values.
--
-- NOTE: To generate random values you should use math.random(). However, you SHOULD NOT seed
--       the random number generator with a new seed value, as that is how TIS-100 ensures that
--       the first test run is consistent for all users, and thus something that allows for the
--       comparison of cycle scores.
--
-- NOTE: Position values for streams should be between 0 and 3, which correspond to the far
--       left and far right of the TIS-100 segment grid. Input streams will be automatically
--       placed on the top, while output and image streams will be placed on the bottom.
--
function get_streams()
    input1 = {{},{}}
    input2 = {{},{}}
    output = {{},{}}
    for i = 1,39 do
        repeat
            n1 = math.random(0, 999999)
            n2 = math.random(0, 999999)
            n3 = n1 + n2
        until n3 < 1000000
        input1[1][i], input1[2][i] = math.floor(n1/1000), n1 % 1000
        input2[1][i], input2[2][i] = math.floor(n2/1000), n2 % 1000
        output[1][i], output[2][i] = math.floor(n3/1000), n3 % 1000
    end
    return {
        { STREAM_INPUT, "IN.AU", 0, input1[1] },
        { STREAM_INPUT, "IN.BU", 1, input2[1] },
        { STREAM_INPUT, "IN.AL", 2, input1[2] },
        { STREAM_INPUT, "IN.BL", 3, input2[2] },
        { STREAM_OUTPUT, "OUT.U", 1, output[1] },
        { STREAM_OUTPUT, "OUT.L", 2, output[2] },
    }
end

-- The function get_layout() should return an array of exactly 12 TILE_* values, which
-- describe the layout and type of tiles found in the puzzle.
--
-- TILE_COMPUTE: A basic execution node (node type T21).
-- TILE_MEMORY: A stack memory node (node type T30).
-- TILE_DAMAGED: A damaged execution node, which acts as an obstacle.
--
function get_layout()
    return {
        TILE_COMPUTE,   TILE_COMPUTE,   TILE_COMPUTE,   TILE_COMPUTE,
        TILE_COMPUTE,   TILE_COMPUTE,   TILE_COMPUTE,   TILE_COMPUTE,
        TILE_COMPUTE,   TILE_COMPUTE,   TILE_COMPUTE,   TILE_COMPUTE,
    }
end