fst

fst copied to clipboard

FST walker for forward/backward transitions

2

For some uses, it would make sense to statefully walk through the FST, stepping forward with input bytes, but also popping off the last input and returning to the previous...

joshtriplett

consider adding a tokenizer/scanning routine

5

@llogiq brought up this use case where one might have a really big set of tokens that one wants to use to scan some text. It turns out that this...

BurntSushi

Motivation - Get distance while using `Map::search_with_state`. Breaking changes - `Levenshtein::new(query: &str, distance: u32)` changed to `distance: usize`. - Original automaton state wrapped into `LevenshteinState`. Tests - `cargo test --features...

MartinKavik

Please provide a FromIterator impl

3

fst has all the necessary infrastructure for `FromIterator` to work; please consider implementing it, so that `collect` works, and in particular to make it easy to collect into `Result` or...

joshtriplett

Methods to filter fst into new fst based on bytes

5

For some uses of an FST, before applying a large number of queries to the fst, the caller may be able to prune out states and transitions from the FST...

joshtriplett

Optimization for sparse automatons

9

Hello The fst can be queried with a specific key or with an automaton. The key is optimized by just „going forward“, while the automaton is „offered“ one byte at...

vorner

levenshtein automata not matching Japanese Characters correctly

15

Hi, awesome library and documentation. The explanations are really nice to read. I have an example that should probably match but, doesn't. I guess it has to do something with...

PSeitz

I added a `contains` automaton, then you can search keyword contains in some terms. ```Rust use fst::{automaton::Contains, IntoStreamer, Set}; fn main() -> Result { let paths = vec!["a foo bar",...

hengfeiyang

Resolves #169

jfolz

Create Str and Subsequence automata from bytes

1

I was wondering if there is any particular reason why `Str` and `Subsequence` automata can only be created from strings. The other parts of the API that I've looked at...

jfolz