Struct NFA

pub struct NFA { /* private fields */ }

A contiguous NFA implementation of Aho-Corasick.

When possible, prefer using AhoCorasick instead of this type directly. Using an NFA directly is typically only necessary when one needs access to the Automaton trait implementation.

This NFA can only be built by first constructing a noncontiguous::NFA. Both NFA::new and Builder::build do this for you automatically, but Builder::build_from_noncontiguous permits doing it explicitly.

The main difference between a noncontiguous NFA and a contiguous NFA is that the latter represents all of its states and transitions in a single allocation, where as the former uses a separate allocation for each state. Doing this at construction time while keeping a low memory footprint isn’t feasible, which is primarily why there are two different NFA types: one that does the least amount of work possible to build itself, and another that does a little extra work to compact itself and make state transitions faster by making some states use a dense representation.

Because a contiguous NFA uses a single allocation, there is a lot more opportunity for compression tricks to reduce the heap memory used. Indeed, it is not uncommon for a contiguous NFA to use an order of magnitude less heap memory than a noncontiguous NFA. Since building a contiguous NFA usually only takes a fraction of the time it takes to build a noncontiguous NFA, the overall build time is not much slower. Thus, in most cases, a contiguous NFA is the best choice.

Since a contiguous NFA uses various tricks for compression and to achieve faster state transitions, currently, its limit on the number of states is somewhat smaller than what a noncontiguous NFA can achieve. Generally speaking, you shouldn’t expect to run into this limit if the number of patterns is under 1 million. It is plausible that this limit will be increased in the future. If the limit is reached, building a contiguous NFA will return an error. Often, since building a contiguous NFA is relatively cheap, it can make sense to always try it even if you aren’t sure if it will fail or not. If it does, you can always fall back to a noncontiguous NFA. (Indeed, the main AhoCorasick type employs a strategy similar to this at construction time.)

Example

This example shows how to build an NFA directly and use it to execute Automaton::try_find:

use aho_corasick::{
    automaton::Automaton,
    nfa::contiguous::NFA,
    Input, Match,
};

let patterns = &["b", "abc", "abcd"];
let haystack = "abcd";

let nfa = NFA::new(patterns).unwrap();
assert_eq!(
    Some(Match::must(0, 1..2)),
    nfa.try_find(&Input::new(haystack))?,
);

It is also possible to implement your own version of try_find. See the Automaton documentation for an example.

Implementations

impl NFA

pub fn new<I, P>(patterns: I) -> Result<NFA, BuildError>

where I: IntoIterator<Item = P>, P: AsRef<[u8]>,

Create a new Aho-Corasick contiguous NFA using the default configuration.

Use a Builder if you want to change the configuration.

pub fn builder() -> Builder

A convenience method for returning a new Aho-Corasick contiguous NFA builder.

This usually permits one to just import the NFA type.

Trait Implementations

impl Automaton for NFA

fn start_state(&self, anchored: Anchored) -> Result<StateID, MatchError>

Returns the starting state for the given anchor mode.

fn next_state(&self, anchored: Anchored, sid: StateID, byte: u8) -> StateID

Performs a state transition from sid for byte and returns the next state.

fn is_special(&self, sid: StateID) -> bool

Returns true if the given ID represents a “special” state. A special state is a dead, match or start state.

fn is_dead(&self, sid: StateID) -> bool

Returns true if the given ID represents a dead state.

fn is_match(&self, sid: StateID) -> bool

Returns true if the given ID represents a match state.

fn is_start(&self, sid: StateID) -> bool

Returns true if the given ID represents a start state.

fn match_kind(&self) -> MatchKind

Returns the match semantics that this automaton was built with.

fn patterns_len(&self) -> usize

Returns the total number of patterns compiled into this automaton.

fn pattern_len(&self, pid: PatternID) -> usize

Returns the length of the pattern for the given ID.

fn min_pattern_len(&self) -> usize

Returns the length, in bytes, of the shortest pattern in this automaton.

fn max_pattern_len(&self) -> usize

Returns the length, in bytes, of the longest pattern in this automaton.

fn match_len(&self, sid: StateID) -> usize

Returns the total number of matches for the given state ID.

fn match_pattern(&self, sid: StateID, index: usize) -> PatternID

Returns the pattern ID for the match state given by sid at the index given.

fn memory_usage(&self) -> usize

Returns the heap memory usage, in bytes, used by this automaton.

fn prefilter(&self) -> Option<&Prefilter>

Returns a prefilter, if available, that can be used to accelerate searches for this automaton.

fn try_find(&self, input: &Input<'_>) -> Result<Option<Match>, MatchError>

Executes a non-overlapping search with this automaton using the given configuration.

fn try_find_overlapping( &self, input: &Input<'_>, state: &mut OverlappingState, ) -> Result<(), MatchError>

Executes a overlapping search with this automaton using the given configuration.

fn try_find_iter<'a, 'h>( &'a self, input: Input<'h>, ) -> Result<FindIter<'a, 'h, Self>, MatchError>

where Self: Sized,

Returns an iterator of non-overlapping matches with this automaton using the given configuration.

fn try_find_overlapping_iter<'a, 'h>( &'a self, input: Input<'h>, ) -> Result<FindOverlappingIter<'a, 'h, Self>, MatchError>

where Self: Sized,

Returns an iterator of overlapping matches with this automaton using the given configuration.

fn try_replace_all<B>( &self, haystack: &str, replace_with: &[B], ) -> Result<String, MatchError>

where Self: Sized, B: AsRef<str>,

Replaces all non-overlapping matches in haystack with strings from replace_with depending on the pattern that matched. The replace_with slice must have length equal to Automaton::patterns_len.

fn try_replace_all_bytes<B>( &self, haystack: &[u8], replace_with: &[B], ) -> Result<Vec<u8>, MatchError>

where Self: Sized, B: AsRef<[u8]>,

Replaces all non-overlapping matches in haystack with strings from replace_with depending on the pattern that matched. The replace_with slice must have length equal to Automaton::patterns_len.

fn try_replace_all_with<F>( &self, haystack: &str, dst: &mut String, replace_with: F, ) -> Result<(), MatchError>

where Self: Sized, F: FnMut(&Match, &str, &mut String) -> bool,

Replaces all non-overlapping matches in haystack by calling the replace_with closure given.

fn try_replace_all_with_bytes<F>( &self, haystack: &[u8], dst: &mut Vec<u8>, replace_with: F, ) -> Result<(), MatchError>

where Self: Sized, F: FnMut(&Match, &[u8], &mut Vec<u8>) -> bool,

Replaces all non-overlapping matches in haystack by calling the replace_with closure given.

fn try_stream_find_iter<'a, R: Read>( &'a self, rdr: R, ) -> Result<StreamFindIter<'a, Self, R>, MatchError>

where Self: Sized,

Returns an iterator of non-overlapping matches with this automaton from the stream given.

fn try_stream_replace_all<R, W, B>( &self, rdr: R, wtr: W, replace_with: &[B], ) -> Result<()>

where Self: Sized, R: Read, W: Write, B: AsRef<[u8]>,

Replaces all non-overlapping matches in rdr with strings from replace_with depending on the pattern that matched, and writes the result to wtr. The replace_with slice must have length equal to Automaton::patterns_len.

fn try_stream_replace_all_with<R, W, F>( &self, rdr: R, wtr: W, replace_with: F, ) -> Result<()>

where Self: Sized, R: Read, W: Write, F: FnMut(&Match, &[u8], &mut W) -> Result<()>,

Replaces all non-overlapping matches in rdr by calling the replace_with closure given and writing the result to wtr.

impl Clone for NFA

fn clone(&self) -> NFA

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for NFA

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.