PHP Pixel Map Group Efficiency - php

Question

I currently have a map that is 1600 x 1600 stored in MySQL(2,560,000 records). I'm rendering a simple 25x25 map out to users for interaction. Users can "claim" tiles on this map. I'd like to be able to calculate the number of open faces for tiles owned by a given user. I can divide this out by the total tiles owned to determine an arbitrary efficientcy rating.
All map coordinates are simply stored as X/Y values.
I'm looking for something that can potentially process an array of said X/Y values and determine how many open faces are accessible for each owned group. For example...
0 = player
x x x x x
x x 0 x x
x x x x x
4 open faces
x x x x x
x x 0 x x
x x 0 x x
x x x x x
6 open faces
x x x x x
x x x 0 x
x x 0 x x
x x x x x
8 open faces
Right now I'm doing some inefficient array looping to calculate this out. I have a simple counter, then I'm looping through an array of all values and am looking for values +-1 in each direction of X and Y to reduce the count. Each loop either adds 0-4 to the total counter based on the number of finds. The inherent problem with this method is that as a group grows, it will take longer and longer to calculate out. Since it's possible for one group to consume 20,000 points o rmore, it's quite a burden.
Any help is greatly appreciated.

Answer 1

One approach would involve creating a Point class. For example:
class Point {
public $x;
public $y;
public function __construct($x, $y){
$this->x = $x;
$this->y = $y;
}
public function getNeighbors(){
// TODO: What if we are at the edge of the board?
return array(
new Point($x+1, $y+1),
new Point($x+1, $y-1),
new Point($x-1, $y+1),
new Point($x-1, $y-1),
);
}
}
Create instances from that class for each point occupied by a user:
// Generate array of Points from database
$user_points = array(new Point(134, 245), new Point(146, 456));
Iterate through to generate all the neighbors:
// Get a flat array of neighbor Points
$neighbors = array_merge(array_map(function($point){
return $point->getNeighbors();
}, $user_points));
// TOOD: Remove Points that are equal to values in $user_points
Then, lastly, submit a COUNT query for the "neighbor" points to determine how many are occupied by other users and remove those from the total.
(Note: I've added TODOs where more work is to be done.)
The inherent problem with this method is that as a group grows, it will take longer and longer to calculate out.
You should consider using an in-memory key-value store, like Redis. But yes, the look-up time (for occupied blocks), in time complexity, appears linear with regard to the number of entries.

Answer 2

Here's the final block of simple code I came up with to determine the geo efficiency. Some names of things have been changed. :P
I'm running with notices on, and everything's ajax, so I decided to go with single isset checks on a multi-dimensional instead of something else.
$sql = 'SELECT map_x, map_y FROM Map WHERE person_id = :person_id';
$query = $DB->prepare($sql);
$query->execute(array(':nation_id' => $this->person_id));
$counter = 0;
$coords = array();
while($row = $query->fetch())
{
++$counter;
$coords[$row['map_x']][$row['map_y']] = 1;
}
$faces = 0;
foreach($coords as $x => $batch)
{
foreach($batch as $y => $junk)
{
$hits = 4;
if(isset($coords[$x + 1][$y]))
{
--$hits;
}
if(isset($coords[$x - 1][$y]))
{
--$hits;
}
if(isset($coords[$x][$y - 1]))
{
--$hits;
}
if(isset($coords[$x][$y + 1]))
{
--$hits;
}
$faces += $hits;
}
}