@timohausmann/quadtree-ts
v2.2.2
Published
Quadtree Typescript Implementation
Downloads
5,609
Readme
quadtree-ts
This library can store and retrieve Rectangles, Circles and Lines in a recursive 2D quadtree. Every node can hold a maximum number of objects before it splits into four subnodes. Objects are only stored on leaf nodes (the lowest level). If an object overlaps into multiple leaf nodes, a reference to the object is stored in each node.
This is a fork of @timohausmann/quadtree-js using Typescript and supporting primitives and overall better extensibility.
Docs and examples
Install
Install this module via npm and import or require it:
npm install --save-dev @timohausmann/quadtree-ts
// ES6
import { Quadtree } from '@timohausmann/quadtree-ts';
// CommonJS
const { Quadtree } = require('@timohausmann/quadtree-ts');
Alternatively, download the source and include it the old-fashioned way, or use an awesome CDN like jsdelivr or unpkg. (If you only need Rectangles and want to save some bytes, use quadtree.umd.basic.js
instead):
<!-- self-hosted -->
<script src="quadtree.umd.full.js"></script>
<!-- CDN jsdelivr -->
<script src="https://cdn.jsdelivr.net/npm/@timohausmann/quadtree-ts/dist/quadtree.umd.full.js"></script>
<!-- CDN unpkg -->
<script src="https://unpkg.com/@timohausmann/quadtree-ts/dist/quadtree.umd.full.js"></script>
Use
Create a new Quadtree:
import { Quadtree } from '@timohausmann/quadtree-ts';
const myTree = new Quadtree({
width: 800,
height: 600,
x: 0, // optional, default: 0
y: 0, // optional, default: 0
maxObjects: 10, // optional, default: 10
maxLevels: 4 // optional, default: 4
});
The tree and each node may have four subnodes that are arranged like this:
| II | I | |-------|-------| | III | IV |
Optional properties:
maxObjects
– defines how many objects a node can hold before it splitsmaxLevels
– defines the deepest level subnodex
andy
– coordinate offset
I recommend using low values for maxLevels
because each level will quadruple the possible amount of nodes. Using lower values for maxLevels
increases performance but may return more candidates. Finetuning these values depends on your 2D space, the amount and size of the objects and your retrieving areas.
Insert elements in the Quadtree:
import { Rectangle, Circle, Line } from '@timohausmann/quadtree-ts';
const rectangle = new Rectangle({
x: 100,
y: 100,
width: 100,
height: 100
});
myTree.insert(rectangle);
const line = new Line({
x1: 25,
y1: 25,
x2: 75,
y2: 25
})
myTree.insert(line);
const circle = new Circle({
x: 100,
y: 100,
r: 50
})
myTree.insert(circle);
Retrieve elements from nodes that intersect with the given shape:
const area = new Rectangle({
x: 150,
y: 150,
width: 100,
height: 100
})
const elements = myTree.retrieve(area);
Reset the Quadtree:
myTree.clear();
Shapes
The supported built-in primitive shapes are Rectangle
, Circle
and Line
.
All shapes can be inserted or used for retrieval.
Each shape requires properties specific to their geometry.
| Shape | Required Properties | |-----------|---------------------| | Rectangle | x, y, width, height | | Circle | x, y, r | | Line | x1, y1, x2, y2 |
You can use these classes directly, extend them or integrate them in your own objects (see below).
Note: if you are using the UMD bundles, the classes are available as Quadtree.Rectangle
, Quadtree.Circle
and Quadtree.Line
.
// Class usage as-is
const player = new Rectangle({
x: 67,
y: 67,
width: 100,
height: 100
});
myTree.insert(player);
// Class extension
class Explosion extends Circle {
constructor(props) {
super(props);
}
explode() {
console.log('boom!');
}
}
const explosion = new Explosion({ x: 100, y: 100, r: 100 });
const affectedObjects = myTree.retrieve(explosion);
Custom data
All shapes support an optional data property that you can use however you like.
const rectangle = new Rectangle({
x: 24,
y: 24,
width: 100,
height: 100,
data: 'custom data here'
});
const circle = new Circle({
x: 128,
y: 128,
r: 50,
data: {
name: 'Stanley',
health: 100
}
});
Custom integration
Under the hood all shape classes implement a qtIndex
method that is crucial for determining in which quadrant a shape belongs. You can think of it as a shape identifier. An alternative to using the class constructors is to supply your own qtIndex
function for objects you want the Quadtree to interact with.
This is also helpful if your existing object properties don't match the required shape properties and you have to map them.
// Custom integration without mapping properties
// In this case the custom object has all the
// expected shape properties (x1, y1, x2, y2)
// So we can simply reference the qtIndex method
const redLaser = {
color: 'red',
brightness: 10,
x1: 67,
y1: 67,
x2: 128,
y2: 128,
qtIndex: Line.prototype.qtIndex
}
myTree.insert(redLaser);
// Custom integration with mapping properties
// In this case the coordinates are arrays so they need to be mapped
const greenLaser = {
color: 'green',
brightness: 10,
startPoint: [50, 50],
endPoint: [100, 50],
qtIndex: function(node) {
return Line.prototype.qtIndex.call({
x1: this.startPoint[0],
y1: this.startPoint[1],
x2: this.endPoint[0],
y2: this.endPoint[1],
}, node);
}
};
myTree.insert(greenLaser);
// Custom integration with mapping in a class
// If you have many instances of the same thing,
// I recommend adding the qtIndex to your class/prototype
class Bomb {
constructor() {
this.position = [50, 50];
this.radius = 100;
}
qtIndex(node) {
return Circle.prototype.qtIndex.call({
x: this.position[0],
y: this.position[1],
r: this.radius,
}, node);
}
}
const bombOmb = new Bomb();
myTree.insert(bombOmb);
Check out the examples for more information.
Typescript
All types can be imported and are documented in the API docs.
The Quadtree class accepts an optional type argument <ObjectsType>
for all inserted/retrieved objects:
class GameEntity extends Rectangle {
...
}
const myTree = new Quadtree<GameEntity>({
width: 800,
height: 600
});
const rocket = new Rocket(...); // extends GameEntity
myTree.insert(rocket);
const results = myTree.retrieve(...); // GameEntity[]
The shape classes accept an optional type argument <CustomDataType>
for the custom data:
interface PlayerData {
name: string
health: number
}
const hero = new Rectangle<PlayerData>({
x: 100,
y: 100,
width: 24,
height: 48,
data: {
name: 'Shiffman',
health: 100,
}
});
Browser Support
As of 2.0.0 the UMD bundles use ES6 features (e.g. classes) that are not supported by IE11 and below. For legacy browser support, please polyfill the code on your own or use quadtree-js.
Development scripts
npm run dev
to watch and build the sourcenpm run build
execute rollup, docs, dtsnpm run rollup
to build the source onlynpm run test
to run the test suitenpm run lint
to run eslintnpm run docs
to create docsnpm run dts
to create definition filesnpx jest -i './test/Quadtree/Quadtree.remove.test.ts' -t 'removes objects from subnodes'
run a single test
Folder structure
/dist
auto-generated bynpm run build
/docs
github pages/docs/documentation
auto-generated bynpm run docs
/docs/examples
the demo examples/src
source code/test
jest test suite/types
Auto-generated bynpm run dts
Migration from quadtree-js
- Named exports only
- Change import to
import { Quadtree } from ...
- Change import to
- Quadtree constructor
maxObjects
andmaxLevels
are now named properties.- Also,
x
andy
are now optional. - Change
new Quadtree({x: 0, y: 0, width: 100, height: 100}, 5, 2);
tonew Quadtree({width: 100, height: 100, maxObjects: 5, maxLevels: 2});
- Objects interacting with the Quadtree need to be a shape class or implement a
qtIndex
method (see above) - Bundle filename has changed:
quadtree.umd.full.js
- Typescript: Use
Rectangle
instead ofRect