short-tree

short-tree

ShortTree is a class extending RBTree from bintrees, and works explicitly on nodes of arrays.

The ShortTree< T > class extends RBTree< Array< T > >.

insert is overloaded and behaves differently. When adding a node, it will first check if there is another shorter node being the beginning of the to-be-inserted node, and if so, won't insert. It also checks if there are existing longer nodes which begin with the newly inserted node, and deletes them.

A new function is added values() which returns Array< Array< T > >, i.e. an array of all nodes (and again, each node is an array of T).

Versions

• Since v2 this is a pure ESM package, and requires Node.js >=12.20. It cannot be used from CommonJS.
• Since v3 requires Node.js >= 14.13.1.

Algorithm

When inserting [ 'a', 'b', 'c', 'd' ], one node is inserted with this value.

Inserting [ 'x', 'y' ], will insert one new node.

If later, [ 'a', 'b', 'c', 'd', 'e' ] is inserted, it won't be - there's already a "shorter" version of this node (the first one inserted).

If later, [ 'a', 'b' ] is inserted, the first node [ 'a', 'b', 'c', 'd' ] will be removed (or "chopped off" after b).

API

Construct a ShortTree by giving the comparison function for T.

If T is number e.g., this could be (a, b) => a - b.

Order

The order when traversing the values is going to depend on the comparison function provided to the constructor.

E.g. a traditional (a: string, b: string) => a.localeCompare(b) will ensure an order for your current locale. You can use Intl to define string comparison orders for other situations.

If the human friendly order isn't that important, but speed is critical, use fast-string-compare.

Example

import { compare } from 'fast-string-compare'
import { ShortTree } from 'short-tree'

// T is deduced to {string}
const tree = new ShortTree( compare );

tree.insert( [ 'a', 'b', 'c', 'd' ] );
tree.insert( [ 'x', 'y' ] );
// This will "chop off" (i.e. remove) [ 'a', 'b', 'c', 'd' ]
tree.insert( [ 'a', 'b' ] );

tree.values( ); // [ [ 'a', 'b' ], [ 'x', 'y' ] ]