Struct MaxScoreExecutor 

pub struct MaxScoreExecutor<'a> { /* private fields */ }

Unified Block-Max MaxScore executor for top-k retrieval

Works with both full-text (BM25) and sparse vector (dot product) queries through the polymorphic TermCursor. Combines three optimizations:

1. MaxScore partitioning (Turtle & Flood 1995): terms split into essential (must check) and non-essential (only scored if candidate is promising)
2. Block-max pruning (Ding & Suel 2011): skip blocks where per-block upper bounds can’t beat the current threshold
3. Conjunction optimization (Lucene/Grand 2023): progressively intersect essential terms as threshold rises, skipping docs that lack enough terms

Implementations

impl<'a> MaxScoreExecutor<'a>

pub fn set_predicate(&mut self, predicate: Option<DocPredicate<'a>>)

Set a filter predicate that rejects documents before scoring.

pub fn sparse( sparse_index: &'a SparseIndex, query_terms: Vec<(u32, f32)>, k: usize, heap_factor: f32, ) -> Self

Create an executor for sparse vector queries.

Builds TermCursor::Sparse for each matched dimension.

pub fn text( posting_lists: Vec<(BlockPostingList, f32)>, avg_field_len: f32, k: usize, ) -> Self

Create an executor for full-text BM25 queries.

Builds TermCursor::Text for each posting list.

pub async fn execute(self) -> Result<Vec<ScoredDoc>>

Execute Block-Max MaxScore and return top-k results (async).

pub fn execute_sync(self) -> Result<Vec<ScoredDoc>>

Synchronous execution — works when all cursors are text or mmap-backed sparse.