natsort-rs 0.1.10

A blazing fast natural sorting library for Python

natsort-rs

🚀 A blazing fast natural sorting library for Python written in Rust 🦀

Installation

pip install natsort-rs

Usage

from natsort_rs import natsort

Sort a list of strings

items = ['item 1', 'item 10', 'item 3']
print(natsort(items))  
# ['item 1', 'item 3', 'item 10']

Sort case insensitively

items = ['Item 1', 'Item 3', 'item 2']
print(natsort(items, ignore_case=True))
# ['Item 1', 'item 2', 'Item 3']

Sort complex objects based on property

items = [
    {'name': 'item 1', 'id': 1},
    {'name': 'item 3', 'id': 3},
    {'name': 'item 2', 'id': 2}
]
print(natsort(items, key=lambda d: d['name']))
# [{'name': 'item 1', 'id': 1}, {'name': 'item 2', 'id': 2}, {'name': 'item 3', 'id': 3}]

Return the sorting indices

This can be helpful if you only want to get the sorted indices returned, that makes the performance-critical part useful for custom sorting use cases:

items = ['item 1', 'item 10', 'item 3']
print(natsort(items, return_indices=True))  
# [0, 2, 1]

Turn off multithreading

Multithreading is currently experimental and can be turned on by setting parallel = True.

Benchmark

No. of items	Duration natsort [s]	Duration natsort-rs [s]	Relative speedup
10	0.00006	0.00000	16.8
100	0.00094	0.00002	44.3
1000	0.00281	0.00022	12.7
10000	0.02835	0.00262	10.8
100000	0.29712	0.03334	8.9
1000000	3.31207	0.45333	7.3

Execute benchmark.py to reproduce the results.

Credits

This Python module is build on top of the natord crate and inspired by natsort.

License

MIT License