Struct NgramFilter 

pub struct NgramFilter { /* private fields */ }

A filter based on required n-grams (2-byte subsequences).

A block passes if all n-grams from at least one pattern might be present.

Example

use flashsieve::NgramFilter;

let filter = NgramFilter::from_patterns(&[b"abc".as_slice()]);
// Blocks that might contain "ab" AND "bc" will match

Implementations

impl NgramFilter

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

Build a filter from a set of patterns.

Patterns shorter than two bytes contribute zero n-grams and therefore match any bloom filter.

Examples

use flashsieve::{NgramFilter, NgramBloom};

let filter = NgramFilter::from_patterns(&[b"hello".as_slice()]);
let bloom = NgramBloom::from_block(b"hello world", 1024).unwrap();
assert!(filter.matches_bloom(&bloom));

pub fn matches_bloom(&self, bloom: &NgramBloom) -> bool

Check if a bloom filter passes this filter.

Returns true when all 2-byte n-grams from at least one source pattern might be present in the bloom filter.

Example

use flashsieve::{NgramFilter, NgramBloom};

let filter = NgramFilter::from_patterns(&[b"hello".as_slice()]);
let bloom = NgramBloom::from_block(b"hello world", 1024).unwrap();
assert!(filter.matches_bloom(&bloom));

pub fn quick_reject(&self, data: &[u8]) -> bool

Fast-path heuristic that checks only the first 4KB of data.

This avoids building a full bloom filter for large files when checking a few n-grams.

This is a heuristic with possible false positives, but never false negatives for patterns that appear entirely within the first 4KB.

Example

use flashsieve::NgramFilter;

let filter = NgramFilter::from_patterns(&[b"hello".as_slice()]);
let data = b"hello world";
assert!(filter.quick_reject(data));

pub fn matches_bloom_pair(&self, b1: &NgramBloom, b2: &NgramBloom) -> bool

Check if the union of two bloom filters passes this filter.

Example

use flashsieve::{NgramFilter, NgramBloom};

let filter = NgramFilter::from_patterns(&[b"ab".as_slice()]);
let b1 = NgramBloom::from_block(b"ab", 1024).unwrap();
let b2 = NgramBloom::from_block(b"xy", 1024).unwrap();
assert!(filter.matches_bloom_pair(&b1, &b2));

pub fn matches_bloom_multi(&self, blooms: &[NgramBloom]) -> bool

Check if the union of multiple consecutive bloom filters passes this filter.

Returns true when all 2-byte n-grams from at least one source pattern might be present in the union of the provided blooms.

Example

use flashsieve::{NgramFilter, NgramBloom};

let filter = NgramFilter::from_patterns(&[b"ab".as_slice()]);
let b1 = NgramBloom::from_block(b"ab", 1024).unwrap();
let b2 = NgramBloom::from_block(b"cd", 1024).unwrap();
let b3 = NgramBloom::from_block(b"ef", 1024).unwrap();
assert!(filter.matches_bloom_multi(&[b1, b2, b3]));

Trait Implementations

impl Clone for NgramFilter

fn clone(&self) -> NgramFilter

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for NgramFilter

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

Formats the value using the given formatter.