k-dimensional tree

Implementation of k-dimensional tree with no dependencies for multi purpose. Common uses are: search in multisemensional spaces (range searches and nearest neighbor searches).

Introduction

k-dimensional tree in data structure is a type of binary tree in which each leaf of the tree represents a point in a space of k dimensions.

This structure allows for very useful kinds of operations with an interesting computational cost. For instance, finding the post office closest to a certain point can be a hard task if the number of post offices is very large. A search for the nearest neighbor in a k-dimensional solves this problem with a computational cost of O (log n) in the average case.

Usage

$ npm i k-dimensional_tree

const {KdTree, Point, Rect} = require('k-dimensional_tree');

// makes a KdTree for two dimensions
const kdt = new KdTree(2);

kdt.insert(new Point([0.5, 0.3]));
kdt.insert(new Point([0.4, 0.01]));

console.log(kdt.nearest(new Point([0.01, 2])));
console.log(kdt.range(new Rect(new Point([0.01, 0.1]), new Point([0.5, 0.35]))));
console.log(kdt.pointsInRadius(new Point([0.01, 2]), 0.075));

API

insert

Create point in tree

kdt.insert(new Point([0.5, 0.3]));

contains

Check if point p exists in k-dimention tree

kdt.contains(new Point([0.5, 0.3]));

size

Get numbet of points in k-dimentional tree

kdt.contains();

isEmpty

Check if k-dimention tree is empty

kdt.isEmpty();

nodes

Get all points in k-dimentional tree

kdt.nodes();

nearest

Get nearest neighbor point of p point

kdt.nearest(new Point([0.01, 2]));

Range searchs

range

Query points inside rectangle

kdt.range(new Rect(new Point([0.01, 0.1], new Point([0.5, 0.35]))));

pointsInRadius

Query all points inside radius from a p point

kdt.pointsInRadius(new Point([0.01, 2]), 0.075)