cmudict-fast 0.8.0

An updated and improved fork of the library for using the CMUSphinx pronunciation dictionary

CMUSphinx pronunciation dictionary

This is a fork of the original rust library for getting pronunciations from the CMUSphinx pronunciation dictionary.

:warning: 0.6.0 onward is not api compatible with the original crate. 0.5.0 or before is mostly api compatible but notably doesn't have the download utility function.


  • 0.8.0
    • implement FromStr for cmudict
  • 0.7.1
    • update version in readme :facepalm:
  • 0.7.0
    • added ability to be serialized and deserialized with serde behind "serialization" feature
  • 0.6.0
    • changed get function to return a slice of all potential pronunciations
  • 0.5.0
    • merged cmudict_core and cmudict into one crate
    • added back rhyme function that got lost in the chaos
  • 0.4.1
    • patch to bring fixed readme onto
  • 0.4.0
    • switched from the depreciated failure crate to thiserror based errors
    • switched to rust 2018 in the cargo config
    • removed the utility function to download
    • removed a lot of dependencies that were either no longer needed or depreciated
    • fixed hung lookup bug (due to the new internals)
    • changed the internals so file i/o wasn't needed for every lookup (and as a result simplified them)

The last of those changes is where the name comes from - as a result of the entire dictionary being loaded into volitile memory, each individual lookup is much faster than the original crate (should be O(1) on average with worst case O(n)). Of course, there's the downside that it takes longer to createthe object and uses more memory, but for the application I initially created this for the lookup time was essential.


To use in your rust project, add the following to your Cargo.toml:

cmudict-fast = "0.8"


To use the dictionary, you have to get an instance of the Cmudict struct:

use cmudict_fast::Cmudict;

fn main() {
  let dict = Cmudict::new("./path/to/a/cmudict/file").expect("Couldn't make Cmudict");

You can retrieve the pronunciation for a word like this:

use cmudict_fast::Cmudict;

fn main() {
  let dict = Cmudict::new("path/to/cmudict").expect("Couldn't get/make Cmudict");
  let word = dict.get("apple").unwrap().pronunciation();
  println!("{:?}", word); // &[Symbol::AE(Stress::Primary), Symbol::P, Symbol::AH(Stress::None), Symbol::L]