node-trees is a library containing tree data structures. It currently only contains quad-tree, but more will be on the way.
Install from NPM in your terminal
npm install node-trees
Require it
var QuadTree = require('node-trees').QuadTree;
var myQuadTree = QuadTree();
Quad trees are great data structures for 2d positional data. When the quad tree is created, it contains a single node. As data is inserted and the contents of the node grow until a threshold is reached. Once the threshold is reached, the node splits into four smaller nodes, one for each quadrant of the original node. This process also occurs in the smaller nodes. As more and more data is added, the tree grows deeper and deeper until a node can no longer be split. Nodes that are either too small, or too deep are prevented from splitting.
Quad trees are great for indexing and storing large amounts of 2d spacial elements. To demonstrate how to use them let's start with an example. Let's say we have a map with a handful of locations marked on it. Each location has an x, y, width, and height.
Without a quad tree we might store our data in an array.
var locations = [
{ label: 'Home' id: 0, x: 3455, y: 12711, width: 243, height: 299 },
{ label: 'Work' id: 1, x: -654, y: 2044, width: 600, height: 546 },
{ label: 'The Park' id: 2, x: 31, y: 34127, width: 1091, height: 3117 }
...
];
Now lets suppose we want to render some of the locations to the screen. In order to retrieve all locations with in the viewable area, we need to loop through each item and make sure its within the viewable area.
var location, viewableLocations = [];
for(var i = 0; i < locations.length; i += 1) {
location = locations[i];
if(
location.x >= view.x &&
location.y >= view.y &&
location.x + location.width <= view.x + view.width &&
location.y + location.height <= view.y + view.height
) {
viewableLocations.push(location);
}
}
This is fine when there are only a few locations, but as the number of locations on the map increase the above code becomes less and less efficent. Eventually this loop will become a major performance problem.
The solution is a quad tree instead of an array. Each location gets inserted into the tree using its location as a key.
var locations = new QuadTree();
locations.insert({ x: 3455, y: 12711, width: 243, height: 299 }, { label: 'Home' id: 0 });
locations.insert({ x: -654, y: 2044, width: 600, height: 546 }, { label: 'Work' id: 1 });
locations.insert({ x: 31, y: 34127, width: 1091, height: 3117 }, { label: 'The Park' id: 2 });
...
We can get all the locations within the view by asking the locations quad tree for all locations within the view rectangle.
var visibleLocations = locations.get({ x: view.x, y: view.y, width: view.width, height: view.height });
Unlike with the array, the quad tree does not slow to a crawl when large amounts of data are inserted. Instead the quad tree only loops over the locations within the view area. It does not require checking every location.
new QuadTree(
[number size = 8192]
[, number maxLeafs = 32]
[, number maxDepth = 8]
[, number x = size / 2]
[, number y = size / 2�]
) => QuadTree quadTree
Creates a new QuadTree instance.
| Argument Name | Description |
|---|---|
| size | The inital height and width of the QuadTree instance. Should be larger than the height/width of the area most items in the tree occupy. May help reduce tree growth. |
| maxLeaf | The maximum number of leaves allowed within a node. Once exceeded the containing node splits into four, and the leaves are distributed into each new node. |
| maxDepth | The maximum depth of the tree. Once a node reaches the maxDepth it can no longer split. Instead it will continue to grow as more leaves are inserted into it. |
| x | The x offset of the tree. If most of your items occupy a specific area setting the x may help reduce tree growth. |
| y | The y offset of the tree. If most of your items occupy a specific area setting the y may help reduce tree growth. |
insert(Rect rect[, * data = undefined])
Adds a rectangle, and optionally associated data, to a QuadTree instance. Takes a Rect instance (or and object with x, y, width, height properties) as a key, and any data to link the rectangle position to.
| Argument Name | Description |
|---|---|
| rect | A rectangle; Rect instance or an object containing an x, y, width, height. This will be the location of the rectangle within the tree. |
| data | Data to associate with the rectangle. Can literally be anything you want. |
get(Rect rect[, * data]) => Array results
Retrieves all rectangles and their associated data within a given rectangle. If a second argument is given, only rectangles that match the second arguments data object will be returned, allowing one to check the location of a rectangle.
| Argument Name | Description |
|---|---|
| rect | A Rect instance or an object containing an x, y, width, height. This will be the location in the tree that will be searched for rectangles. |
| data | only results with this data will be returned as results. This allows you to check the location of a rectangle associated with a specific piece of data. |
remove(Rect rect[, * data]) => Array results
Exactly like get but the data returned as results is removed from the tree.
| Argument Name | Description |
|---|---|
| rect | A Rect instance or an object containing an x, y, width, height. This will be the location in the tree that will be searched for rectangles. |
| data | only results with this data will be returned as results. This allows you to check the location of a rectangle associated with a specific piece of data. |
truncate()
Truncate deletes all data in the tree.
I made this library to improve the performance of a game engine I'm working on for Unicode Games. I'd like to share it with fellow devs. Feel free to fork this project and make your own changes.