use crate::pattern_matching::shift_and::masks;
use crate::utils::TextSlice;
use std::borrow::Borrow;
#[derive(Copy, Clone, Eq, PartialEq, Ord, PartialOrd, Hash, Debug)]
pub struct BNDM {
m: usize,
masks: [u64; 256],
accept: u64,
}
impl BNDM {
pub fn new<C, P>(pattern: P) -> Self
where
C: Borrow<u8>,
P: IntoIterator<Item = C>,
P::IntoIter: DoubleEndedIterator + ExactSizeIterator,
{
let pattern = pattern.into_iter();
let m = pattern.len();
assert!(m <= 64, "Expecting a pattern of at most 64 symbols.");
let (masks, accept) = masks(pattern.rev());
BNDM { m, masks, accept }
}
pub fn find_all<'a>(&'a self, text: TextSlice<'a>) -> Matches<'a> {
Matches {
bndm: self,
window: self.m,
text,
}
}
}
#[derive(Copy, Clone, Eq, PartialEq, Ord, PartialOrd, Hash, Debug)]
pub struct Matches<'a> {
bndm: &'a BNDM,
window: usize,
text: TextSlice<'a>,
}
impl<'a> Iterator for Matches<'a> {
type Item = usize;
fn next(&mut self) -> Option<usize> {
while self.window <= self.text.len() {
let mut occ = None;
let mut active = (1u64 << self.bndm.m) - 1;
let (mut j, mut lastsuffix) = (1, 0);
while active != 0 {
active &= self.bndm.masks[self.text[self.window - j] as usize];
if active & self.bndm.accept != 0 {
if j == self.bndm.m {
occ = Some(self.window - self.bndm.m);
break;
} else {
lastsuffix = j;
}
}
j += 1;
active <<= 1;
}
self.window += self.bndm.m - lastsuffix;
if occ.is_some() {
return occ;
}
}
None
}
}
#[cfg(test)]
mod tests {
use super::*;
use itertools::Itertools;
#[test]
fn test_find_all() {
let text = b"dhjalkjwqnnnannanaflkjdklfj";
let pattern = b"qnnnannan";
let bndm = BNDM::new(pattern);
assert_eq!(bndm.find_all(text).collect_vec(), [8]);
}
#[test]
fn test_find_all_at_start() {
let text = b"dhjalkjwqnnnannanaflkjdklfj";
let pattern = b"dhjalk";
let bndm = BNDM::new(pattern);
assert_eq!(bndm.find_all(text).collect_vec(), [0]);
}
}