Enum LogicalNode 

pub enum LogicalNode {
    VectorScan {
        collection: String,
        k: usize,
        oversample: usize,
        alpha: f32,
        strategy: FusionStrategy,
    },
    SparseScan {
        collection: String,
        k: usize,
    },
    Filter {
        predicate: String,
        input: Box<LogicalNode>,
    },
    Join {
        left: Box<LogicalNode>,
        right_collection: String,
        left_key: String,
        right_key: String,
    },
    Project {
        columns: Vec<String>,
        input: Box<LogicalNode>,
    },
    Sort {
        by: SortKey,
        descending: bool,
        input: Box<LogicalNode>,
    },
    Limit {
        n: usize,
        input: Box<LogicalNode>,
    },
    HybridScan {
        collection: String,
        k: usize,
        alpha: f32,
        strategy: FusionStrategy,
        oversample: usize,
    },
}

Variants

VectorScan

Scan the vector index for nearest neighbors.

Fields

collection: String

k: usize

oversample: usize

Oversampling multiplier (dense fetches k * oversample candidates).

alpha: f32

Alpha for hybrid fusion (1.0 = pure dense, 0.0 = pure sparse).

strategy: FusionStrategy

Fusion strategy.

SparseScan

Full sparse BM25 scan (no dense pre-filtering).

Fields

collection: String

k: usize

Filter

Apply a scalar metadata filter.

Fields

predicate: String

Predicate string (serialised JSON or SQL-like).

input: Box<LogicalNode>

Join

Join vector candidates with a metadata table.

Fields

left: Box<LogicalNode>

right_collection: String

Name of the metadata table / collection to join against.

left_key: String

Join key on the left side (field in vector payload).

right_key: String

Join key on the right side (field in metadata).

Project

Project (select) specific fields from results.

Fields

columns: Vec<String>

input: Box<LogicalNode>

Sort

Sort results by a field or score.

Fields

by: SortKey

descending: bool

input: Box<LogicalNode>

Limit

Limit number of results.

Fields

n: usize

input: Box<LogicalNode>

HybridScan

Hybrid scan: dense HNSW + sparse BM25 combined.

Fields

collection: String

k: usize

alpha: f32

strategy: FusionStrategy

oversample: usize

Dense oversampling factor before sparse re-scoring.

Trait Implementations

impl Clone for LogicalNode

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for LogicalNode

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

Formats the value using the given formatter.