Lsys GNN encoding method - Expansion algorithm for connectivity matrix (pseudo code)

Variable: n = number of neurons in GNN
Input:  Genotype G = {cS[1..4], cABCD[1..NumLayers][1..16], ca_p[1..64]}
        where NumLayers = log₂(n) - 2
Output: Connectivity matrix M[1..n][1..n]

// Expand L-system productions
current_string = cS[1..4]

for level = 1 to NumLayers do
    next_string = []
    for each symbol s in current_string do
        if s = A then append cABCD[level][1..4]  to next_string
        if s = B then append cABCD[level][5..8]  to next_string
        if s = C then append cABCD[level][9..12] to next_string
        if s = D then append cABCD[level][13..16] to next_string
    end
    current_string = next_string
end

// Expand terminal symbols to binary connection values
bit_string = []
for each symbol s in current_string do
    index = ord(s) - ord('a')        // a=0, b=1, ..., p=15
    append bits(index, 4) to bit_string
end

// Fill connectivity matrix row by row
k = 1
for row = 1 to n do
    for col = 1 to n do
        M[row][col] = bit_string[k]
        k = k + 1
    end
end