https://github.com/tkluck/accumulation_tree

Red/black tree with support for fast accumulation of values in a key range
https://github.com/tkluck/accumulation_tree

Last synced: 19 days ago
JSON representation

Red/black tree with support for fast accumulation of values in a key range

• Host: GitHub
• URL: https://github.com/tkluck/accumulation_tree
• Owner: tkluck
• License: mit
• Created: 2017-02-05T23:32:49.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2020-05-09T18:00:09.000Z (over 4 years ago)
• Last Synced: 2024-05-02T06:13:54.992Z (7 months ago)
• Language: Python
• Size: 23.4 KB
• Stars: 17
• Watchers: 4
• Forks: 2
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
`accumulation_tree`

===================

[![CircleCI][circleci-img]][circleci-url]

[circleci-url]: https://circleci.com/gh/tkluck/accumulation_tree/

[circleci-img]: https://img.shields.io/circleci/project/github/tkluck/pac4cli.svg

A red/black tree which also stores partial aggregations at each node, making

getting aggregations of key range slices an O(log(N)) operation.

This data structure is very similar to a [Fenwick tree][wiki] and can be used

for the same purpose. The main difference with the description on Wikipedia is

that we allocate nodes on the heap instead of using an implicit tree structure

on a flat array.

This implementation was written specifically for use in [tdigest][tdigest-github],

and borrows code from [bintrees][bintrees-github].

[tdigest-github]: https://github.com/CamDavidsonPilon/tdigest/

[bintrees-github]: https://github.com/mozman/bintrees/

[wiki]: https://en.wikipedia.org/wiki/Fenwick_tree

Synopsis

--------

```python

>>> from accumulation_tree import AccumulationTree

>>> t = AccumulationTree(lambda x: x)

>>> N = 10000

>>> for x in range(N):

...    t.insert(x,x)

>>> t.get_accumulation(0, 2)

1

>>> t.get_accumulation(0, 5)

10

>>> all(t.get_accumulation(0, x) == x*(x-1)/2 for x in range(N))

True

```