Struct BigramFilter 

pub struct BigramFilter { /* private fields */ }

Inverted bigram index with optional “skip-1” extension Copmressed into bitset for minimal usage, the layout of this struct actually matters

Implementations

impl BigramFilter

pub fn query(&self, pattern: &[u8]) -> Option<Vec<u64>>

AND the posting lists for all query bigrams (consecutive + skip). Returns None if no query bigrams are tracked.

pub fn set_skip_index(&mut self, skip: BigramFilter)

Attach a skip-1 bigram index for tighter candidate filtering.

pub fn is_candidate(candidates: &[u64], file_idx: usize) -> bool

pub fn count_candidates(candidates: &[u64]) -> usize

pub fn is_ready(&self) -> bool

pub fn file_count(&self) -> usize

pub fn columns_used(&self) -> usize

pub fn heap_bytes(&self) -> usize

Total heap bytes used by this index (lookup + dense data + skip).

pub fn has_key(&self, key: u16) -> bool

Check whether a bigram key is present in this index.

pub fn lookup(&self) -> &[u16]

Raw lookup table (65536 entries mapping bigram key → column index).

pub fn dense_data(&self) -> &[u64]

Flat dense bitset data at fixed stride words.

pub fn words(&self) -> usize

Number of u64 words per column (= ceil(file_count / 64)).

pub fn dense_count(&self) -> usize

Number of dense columns retained after compression.

pub fn populated(&self) -> usize

Number of files that contributed content to the index.

pub fn skip_index(&self) -> Option<&BigramFilter>

Reference to the optional skip-1 bigram sub-index.

pub fn new( lookup: Vec<u16>, dense_data: Vec<u64>, dense_count: usize, words: usize, file_count: usize, populated: usize, ) -> Self

Create a new bigram filter from the internal data

Trait Implementations

impl Debug for BigramFilter

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

Formats the value using the given formatter.