Struct PatternSplitter 

pub struct PatternSplitter { /* private fields */ }

A compiled multi-pattern splitter for efficient repeated splitting.

Unlike split_at_patterns which rebuilds the Aho-Corasick automaton on each call, PatternSplitter compiles the automaton once and reuses it. This is ~25x faster when splitting multiple texts with the same patterns.

Example

use chunk::{PatternSplitter, IncludeDelim};

// Compile once
let splitter = PatternSplitter::new(&[b". ", b"? ", b"! "]);

// Reuse for multiple texts
let offsets1 = splitter.split(b"Hello. World?", IncludeDelim::Prev, 0);
let offsets2 = splitter.split(b"Another. Text!", IncludeDelim::Prev, 0);

Implementations

impl PatternSplitter

pub fn new(patterns: &[&[u8]]) -> Self

Create a new PatternSplitter with the given patterns.

This compiles the Aho-Corasick automaton, which takes O(total pattern length) time.

pub fn split( &self, text: &[u8], include_delim: IncludeDelim, min_chars: usize, ) -> Vec<(usize, usize)>

Split text using the compiled patterns.

This is O(text length) after compilation.