Struct ApproxFilter 

pub struct ApproxFilter {
    pub trigrams: Vec<u64>,
    pub min_required: usize,
}

Sound trigram filter for an approximate-regex query.

Fields

trigrams: Vec<u64>

Distinct pattern trigram hashes, sorted ascending. Empty means the pattern has no extractable literal runs (no filter possible).

min_required: usize

Minimum number of Self::trigrams that must appear in any matching doc. Always <= trigrams.len(). When zero the filter degenerates to “no constraint” and the caller must full-scan.

Implementations

impl ApproxFilter

pub fn build(ast: &Ast, k: u16) -> Self

Build the filter for a regex AST allowing up to k edit operations.

pub fn is_active(&self) -> bool

Whether the filter imposes any constraint. A degenerate filter (no required trigrams or min_required == 0) passes every doc and the caller should full-scan.

pub fn candidates(&self, postings: &Postings) -> Vec<u32>

Compute the candidate doc-id set for the filter.

The strategy is the postings UNION of every pattern trigram. Any doc that contains at least one pattern trigram is in the union; any doc with min_required >= 1 must be in the union; this is therefore a sound upper bound. Per-doc bloom filtering is applied separately by the index in a tighter inner loop that already has the doc record in hand.

The output is the surviving doc ids in ascending order.

pub fn passes(&self, bloom: &BloomFilter) -> bool

Test the filter against a per-doc bloom filter.

Returns true if the doc may match the pattern under the configured edit budget; false if the pattern imposes more required-trigram constraints than the doc can plausibly satisfy.

Trait Implementations

impl Clone for ApproxFilter

fn clone(&self) -> ApproxFilter

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for ApproxFilter

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

Formats the value using the given formatter.