spellchecker

A spell checker based on the algorithm described by Peter Norvig, Director of Research at Google.

Features

• Can be trained one or more times with a large corpus of words
• Probabilistic model for determining the most likely correct spelling

Installation and Running

To use the library, add the following to your Cargo.toml manifest file:

[dependencies]
regex = { git = "https://github.com/tomconder/spellchecker.git" }

Then in your code, you can use the library as follows:

extern crate spellchecker;
use spellchecker::Checker;

Now you can create a Checker instance and use it to check spelling:

let mut spellchecker = Checker::new();

// train the spellchecker with a large text file 
let contents = fs::read_to_string("big.txt").expect("Something went wrong");
spellchecker.train(&contents);

// the expected value is "poetry"
println!("{} -> {}", "peotryy", spellchecker.correct("peotryy"));

Algorithm

Norvig's algorithm

Peter Norvig's spell checker uses a probabilistic model based on a large corpus of words to determine the most likely correct spelling. It works by generating a set of possible corrections, or edits, for a misspelled word. Then it ranks those corrections based on their likelihood of being the correct spelling.