fst 0.3.0

Use finite state transducers to compactly represents sets or maps of many strings (> 1 billion is possible).


This crate provides a fast implementation of ordered sets and maps using finite state machines. In particular, it makes use of finite state transducers to map keys to values as the machine is executed. Using finite state machines as data structures enables us to store keys in a compact format that is also easily searchable. For example, this crate leverages memory maps to make range queries very fast.

Check out my blog post Index 1,600,000,000 Keys with Automata and Rust for extensive background, examples and experiments.

Linux build status Windows build status

Dual-licensed under MIT or the UNLICENSE.


Full API documentation and examples.

The fst-regex and fst-levenshtein crates provide regular expression matching and fuzzy searching on FSTs, respectively.


Simply add a corresponding entry to your Cargo.toml dependency list:

fst = "0.2"

And add this to your crate root:

extern crate fst;


This example demonstrates building a set in memory and executing a fuzzy query against it. Check out the documentation for a lot more examples!

extern crate fst;
extern crate fst_levenshtein;

use std::error::Error;
use std::process;

use fst::{IntoStreamer, Streamer, Set};
use fst_levenshtein::Levenshtein;

fn try_main() -> Result<(), Box<Error>> {
  // A convenient way to create sets in memory.
  let keys = vec!["fa", "fo", "fob", "focus", "foo", "food", "foul"];
  let set = Set::from_iter(keys)?;

  // Build our fuzzy query.
  let lev = Levenshtein::new("foo", 1)?;

  // Apply our fuzzy query to the set we built.
  let mut stream = set.search(lev).into_stream();

  let keys = stream.into_strs()?;
  assert_eq!(keys, vec!["fo", "fob", "foo", "food"]);

fn main() {
  if let Err(err) = try_main() {
    eprintln!("{}", err);