Fast English word segmentation in Rust
Go to file
David Hotham 29cafd0db1 tweak the non-word penalty 2024-02-12 10:26:21 +01:00
.cargo py: initial version of Python bindings 2021-03-24 11:57:29 +01:00
.github Publish wheels for Intel-based Macs (#52) 2023-12-12 20:09:00 +01:00
data Remove old data files 2021-06-03 16:09:24 +02:00
instant-segment tweak the non-word penalty 2024-02-12 10:26:21 +01:00
instant-segment-py Use adapted version of generated workflow 2023-11-09 12:00:14 +01:00
.gitignore Add script to download word list input data 2021-06-03 16:09:24 +02:00
Cargo.toml py: initial version of Python bindings 2021-03-24 11:57:29 +01:00
LICENSE Add license files (fixes #15) 2021-04-29 15:34:02 +02:00
Makefile Make `test-python` Make-target work on Linux (#51) 2023-12-01 17:20:37 +01:00
README.md Update Readme to reflect recent perf improvements 2023-11-16 01:23:26 +01:00
cover.svg replace cover image (#48) 2023-11-15 10:13:42 -08:00
deny.toml Restrict bad dependencies in cargo deny config 2023-10-27 11:08:22 +02:00

README.md

Cover logo

Instant Segment: fast English word segmentation in Rust

Documentation Crates.io PyPI Build status License: Apache 2.0

Instant Segment is a fast Apache-2.0 library for English word segmentation. It is based on the Python wordsegment project written by Grant Jenks, which is in turn based on code from Peter Norvig's chapter Natural Language Corpus Data from the book Beautiful Data (Segaran and Hammerbacher, 2009).

For the microbenchmark included in this repository, Instant Segment is ~500x faster than the Python implementation. The API was carefully constructed so that multiple segmentations can share the underlying state to allow parallel usage.

How it works

Instant Segment works by segmenting a string into words by selecting the splits with the highest probability given a corpus of words and their occurrences.

For instance, provided that choose and spain occur more frequently than chooses and pain, and that the pair choose spain occurs more frequently than chooses pain, Instant Segment can help identify the domain choosespain.com as ChooseSpain.com which more likely matches user intent.

Read about how we built and improved Instant Segment for use in production at Instant Domain Search to help our users find relevant domains they can register.

Using the library

Python (>= 3.9)

pip install instant-segment

Rust

[dependencies]
instant-segment = "0.8.1"

Examples

The following examples expect unigrams and bigrams to exist. See the examples (Rust, Python) to see how to construct these objects.

import instant_segment

segmenter = instant_segment.Segmenter(unigrams, bigrams)
search = instant_segment.Search()
segmenter.segment("instantdomainsearch", search)
print([word for word in search])

--> ['instant', 'domain', 'search']
use instant_segment::{Search, Segmenter};
use std::collections::HashMap;

let segmenter = Segmenter::new(unigrams, bigrams);
let mut search = Search::default();
let words = segmenter
    .segment("instantdomainsearch", &mut search)
    .unwrap();
println!("{:?}", words.collect::<Vec<&str>>())

--> ["instant", "domain", "search"]

Check out the tests for more thorough examples: Rust, Python

Testing

To run the tests run the following:

cargo t -p instant-segment --all-features

You can also test the Python bindings with:

make test-python