Struct Comparator 

pub struct Comparator { /* private fields */ }

Pairwise record comparator with automatic GPU/CPU selection.

Wraps FieldComparator (CPU) or DeviceComparator (GPU) depending on the Backend. Implements ComparatorTrait identically in both cases.

Implementations

impl Comparator

pub fn from_cpu(fc: FieldComparator) -> Self

Wrap an already-constructed zer_compare::FieldComparator directly.

Use this when you want to override default similarity functions via zer_compare::FieldComparator::with_fns before creating the comparator. Always uses the CPU path; GPU acceleration is not available this way.

pub fn new(schema: &Schema, backend: &Backend) -> Self

Build a comparator from a schema and backend.

pub fn backend_name(&self) -> &'static str

Name of the active backend, for diagnostics.

pub fn compare_batch_from_pool( &self, pool: &RecordPool, pair_indices: &[(usize, usize)], schema: &Schema, ) -> ComparisonBatch

Primary hot-path: pool-native batch comparison.

pool is a RecordPool built from the candidate records; pair_indices holds (i, j) pairs where i and j are indices into the pool. Avoids all Record::clone() and HashMap lookups, the fastest path for large BRP-style jobs where records are already loaded into a pool.

pub fn compare_batch_indexed( &self, records: &[Record], pair_indices: &[(usize, usize)], schema: &Schema, ) -> ComparisonBatch

Convenience wrapper: builds a pool from a flat records slice and compares the pair_indices pairs. No Record::clone().

Trait Implementations

impl Comparator for Comparator

fn compare(&self, a: &Record, b: &Record, schema: &Schema) -> ComparisonVector

Compare a single pair, always CPU, returns an individual vector.

fn compare_batch_from_pool( &self, pool: &RecordPool, indices: &[(usize, usize)], schema: &Schema, ) -> ComparisonBatch

Pool-native batch comparison, the primary hot path.