#version 450

layout(local_size_x = 8, local_size_y = 8, local_size_z = 1) in;

layout(set = 0, binding = 0) buffer wData {
    int buf[];
} write_buffer;

layout(set = 0, binding = 1) buffer rData {
    int buf[];
} read_buffer;

layout(set = 0, binding = 2) buffer Settings {
    int settings[];
} settings;

ivec4 separate(int pix){
    ivec4 r = ivec4(
    (pix & 0x000000FF),
    (pix & 0x0000FF00) >> 8,
    (pix & 0x00FF0000) >> 16,
    (pix & 0xFF000000) >> 24
    );
    return r;
}

uint get_idx(int offset_x, int offset_y){

    uint x = min(settings.settings[0], max(0, gl_GlobalInvocationID.x + offset_x));
    uint y = min(settings.settings[1], max(0, gl_GlobalInvocationID.y + offset_y));
    return ((y * settings.settings[0]) + x);
}

void main() {

    uint idx = get_idx(0,0);

    ivec4 p  = separate(read_buffer.buf[get_idx(0 , 0)]);

    ivec4 p0 = separate(read_buffer.buf[get_idx(0 , 1)]);
    ivec4 p1 = separate(read_buffer.buf[get_idx(0 ,-1)]);
    ivec4 p2 = separate(read_buffer.buf[get_idx(1 , 1)]);
    ivec4 p3 = separate(read_buffer.buf[get_idx(-1,-1)]);
    ivec4 p4 = separate(read_buffer.buf[get_idx(1 , 0)]);
    ivec4 p5 = separate(read_buffer.buf[get_idx(-1, 0)]);
    ivec4 p6 = separate(read_buffer.buf[get_idx(1 ,-1)]);
    ivec4 p7 = separate(read_buffer.buf[get_idx(-1, 1)]);

    ivec4 avg = p0 + p1 + p2 + p3 + p4 + p5 + p6 + p7;
    avg = max(avg, 1) / 8;

    int total_of_avg = avg.x + avg.y + avg.z;

    ivec4 pixels[8] = ivec4[](p0, p1, p2, p3, p4, p5, p6, p7);

    int current_best = 999;
    ivec4 current_best_p = p;

    for (int i = 0; i < 8; i++){
        int v = pixels[i].x + pixels[i].y + pixels[i].z;
        if (abs(total_of_avg - v) < current_best) {
            current_best = abs(total_of_avg - v);
            current_best_p = pixels[i];
        }
    }

    p = current_best_p;

   write_buffer.buf[idx] = (write_buffer.buf[idx] & (~0x000000FF) ) | (p.x);
   write_buffer.buf[idx] = (write_buffer.buf[idx] & (~0x0000FF00) ) | (p.y << 8);
   write_buffer.buf[idx] = (write_buffer.buf[idx] & (~0x00FF0000) ) | (p.z << 16);
   write_buffer.buf[idx] = (write_buffer.buf[idx] & (~0xFF000000) ) | (p.w << 24);
}