RBush is a high-performance JavaScript library for 2D spatial indexing of points and rectangles. It's based on an optimized R-tree data structure with bulk insertion support.
Spatial index is a special data structure for points and rectangles that allows you to perform queries like "all items within this bounding box" very efficiently (e.g. hundreds of times faster than looping over all items). It's most commonly used in maps and data visualizations.
The demos contain visualization of trees generated from 50k bulk-loaded random points. Open web console to see benchmarks; click on buttons to insert or remove items; click to perform search under the cursor.
Install with NPM (npm install rbush), or use CDN links for browsers:
rbush.js,
rbush.min.js
// as a ES module
import RBush from 'rbush';
// as a CommonJS module
const RBush = require('rbush');const tree = new RBush();An optional argument to RBush defines the maximum number of entries in a tree node.
9 (used by default) is a reasonable choice for most applications.
Higher value means faster insertion and slower search, and vice versa.
const tree = new RBush(16);Insert an item:
const item = {
minX: 20,
minY: 40,
maxX: 30,
maxY: 50,
foo: 'bar'
};
tree.insert(item);Remove a previously inserted item:
tree.remove(item);By default, RBush removes objects by reference.
However, you can pass a custom equals function to compare by value for removal,
which is useful when you only have a copy of the object you need removed (e.g. loaded from server):
tree.remove(itemCopy, (a, b) => {
return a.id === b.id;
});Remove all items:
tree.clear();By default, RBush assumes the format of data points to be an object
with minX, minY, maxX and maxY properties.
You can customize this by overriding toBBox, compareMinX and compareMinY methods like this:
class MyRBush extends RBush {
toBBox([x, y]) { return {minX: x, minY: y, maxX: x, maxY: y}; }
compareMinX(a, b) { return a.x - b.x; }
compareMinY(a, b) { return a.y - b.y; }
}
const tree = new MyRBush();
tree.insert([20, 50]); // accepts [x, y] pointsIf you're indexing a static list of points (you don't need to add/remove points after indexing), you should use kdbush which performs point indexing 5-8x faster than RBush.
Bulk-insert the given data into the tree:
tree.load([item1, item2, ...]);Bulk insertion is usually ~2-3 times faster than inserting items one by one. After bulk loading (bulk insertion into an empty tree), subsequent query performance is also ~20-30% better.
Note that when you do bulk insertion into an existing tree, it bulk-loads the given data into a separate tree and inserts the smaller tree into the larger tree. This means that bulk insertion works very well for clustered data (where items in one update are close to each other), but makes query performance worse if the data is scattered.
const result = tree.search({
minX: 40,
minY: 20,
maxX: 80,
maxY: 70
});Returns an array of data items (points or rectangles) that the given bounding box intersects.
Note that the search method accepts a bounding box in {minX, minY, maxX, maxY} format
regardless of the data format.
const allItems = tree.all();Returns all items of the tree.
const result = tree.collides({minX: 40, minY: 20, maxX: 80, maxY: 70});Returns true if there are any items intersecting the given bounding box, otherwise false.
// export data as JSON object
const treeData = tree.toJSON();
// import previously exported data
const tree = rbush(9).fromJSON(treeData);Importing and exporting as JSON allows you to use RBush on both the server (using Node.js) and the browser combined, e.g. first indexing the data on the server and and then importing the resulting tree data on the client for searching.
Note that the nodeSize option passed to the constructor must be the same in both trees for export/import to work properly.
For "k nearest neighbors around a point" type of queries for RBush, check out rbush-knn.
The following sample performance test was done by generating
random uniformly distributed rectangles of ~0.01% area and setting maxEntries to 16
(see debug/perf.js script).
Performed with Node.js v6.2.2 on a Retina Macbook Pro 15 (mid-2012).
| Test | RBush | old RTree | Improvement |
|---|---|---|---|
| insert 1M items one by one | 3.18s | 7.83s | 2.5x |
| 1000 searches of 0.01% area | 0.03s | 0.93s | 30x |
| 1000 searches of 1% area | 0.35s | 2.27s | 6.5x |
| 1000 searches of 10% area | 2.18s | 9.53s | 4.4x |
| remove 1000 items one by one | 0.02s | 1.18s | 50x |
| bulk-insert 1M items | 1.25s | n/a | 6.7x |
- single insertion: non-recursive R-tree insertion with overlap minimizing split routine from R*-tree (split is very effective in JS, while other R*-tree modifications like reinsertion on overflow and overlap minimizing subtree search are too slow and not worth it)
- single deletion: non-recursive R-tree deletion using depth-first tree traversal with free-at-empty strategy (entries in underflowed nodes are not reinserted, instead underflowed nodes are kept in the tree and deleted only when empty, which is a good compromise of query vs removal performance)
- bulk loading: OMT algorithm (Overlap Minimizing Top-down Bulk Loading) combined with Floyd–Rivest selection algorithm
- bulk insertion: STLT algorithm (Small-Tree-Large-Tree)
- search: standard non-recursive R-tree search
- R-trees: a Dynamic Index Structure For Spatial Searching
- The R*-tree: An Efficient and Robust Access Method for Points and Rectangles+
- OMT: Overlap Minimizing Top-down Bulk Loading Algorithm for R-tree
- Bulk Insertions into R-Trees Using the Small-Tree-Large-Tree Approach
- R-Trees: Theory and Applications (book)
npm install # install dependencies
npm test # lint the code and run tests
npm run perf # run performance benchmarks
npm run cov # report test coverageRBush should run on Node and all major browsers that support ES5.
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent