Struct VectorSearchRequest 

pub struct VectorSearchRequest {

    pub index_name: String,
    pub query: Vec<f32>,
    pub k: u32,
    pub ef_search: u32,
    pub filter: Option<VectorFilter>,
    pub mmr: bool,
    pub mmr_lambda: f32,
    pub mmr_pool: u32,
    pub rerank: Option<RerankSpec>,
    pub planner_override: u32,
    pub group_field: String,
    pub group_size: u32,
    pub groups: u32,
    pub group_overfetch: u32,
    pub group_missing_as_own: bool,
    pub query_payload: Option<QueryPayload>,
}

Fields

index_name: String

query: Vec<f32>

Query vector (single-vector indices).

k: u32

Number of nearest neighbors to return

ef_search: u32

Optional: override ef_search (0 = use default)

filter: Option<VectorFilter>

Optional: metadata/attribute pre/post filter (null = no filter)

mmr: bool

── MMR diversity reranking (optional, single-vector path only) ── Maximal-marginal-relevance post-step: over-fetch candidates, then greedily select k maximizing lambda*sim(q,d) - (1-lambda)*max_{s in selected} sim(d,s). Additive & wire-compatible: old clients omit these ⇒ MMR off, behavior unchanged. The returned distance is always the original query distance; only the selected set and ordering change.

enable MMR post-step

mmr_lambda: f32

0..1 relevance↔diversity tradeoff; 0 ⇒ default 0.5

mmr_pool: u32

over-fetch multiplier; 0 ⇒ default 4 (fetch_k = k*mmr_pool)

rerank: Option<RerankSpec>

Optional second-stage rerank (cross-encoder or model-free score-fusion). Absent ⇒ no rerank.

planner_override: u32

── Filtered-search planner override (optional) ── Pin the selectivity-adaptive filtered-search strategy instead of letting the cardinality planner choose. Only meaningful when filter is set and the target index is graph-backed (Plain/PQ/SQ HNSW); ignored otherwise. 0 = auto (planner chooses from the estimated selectivity) [default] 1 = brute-force exact scan over the matching ids 2 = filtered-HNSW (traverse-through-non-matching, admit-if-match) 3 = ACORN two-hop neighbor expansion Additive & wire-compatible: old clients omit it ⇒ 0 ⇒ unchanged behavior.

group_field: String

── Result grouping / field-collapse (optional, single-vector path only) ── Collapse results to at most group_size hits per distinct value of the payload attribute group_field, returning groups distinct group keys total (“one best chunk per document”). Pure post-search collapse over the post-filter candidate pool, structurally identical to the MMR over-fetch / re-select path. Additive & wire-compatible: old clients omit these ⇒ grouping off ⇒ behavior byte-identical. Grouping and mmr are mutually exclusive (orthogonal selection rules) and requesting both ⇒ InvalidArgument.

payload field to group by; empty ⇒ grouping off

group_size: u32

max hits per group; 0 ⇒ default 1 (one-best-per-group)

groups: u32

number of distinct group keys to return; 0 ⇒ fall back to k

group_overfetch: u32

over-fetch multiplier; 0 ⇒ default 4, capped at MAX_POOL

group_missing_as_own: bool

Missing-field policy. By default (false) candidates that lack group_field are dropped from grouped results (matches Qdrant/Weaviate). When true, each missing-field candidate is returned as its own singleton group (group_key empty), counting against the groups cap. Additive/wire-compatible: old clients omit it ⇒ false ⇒ drop behavior unchanged from Phase 1.

query_payload: Option<QueryPayload>

Query payload selector. query (field 2) carries a single-vector query for the classic ANN path; multi_vector carries a token matrix for the late-interaction MaxSim path against a multi-vector index. At most one of the two should be populated.

Trait Implementations

impl Clone for VectorSearchRequest

fn clone(&self) -> VectorSearchRequest

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for VectorSearchRequest

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

Formats the value using the given formatter.

impl Default for VectorSearchRequest

fn default() -> Self

Returns the “default value” for a type.

impl Message for VectorSearchRequest

fn encoded_len(&self) -> usize

Returns the encoded length of the message without a length delimiter.

fn clear(&mut self)

Clears the message, resetting all fields to their default.

fn encode(&self, buf: &mut impl BufMut) -> Result<(), EncodeError>

where Self: Sized,

Encodes the message to a buffer.

fn encode_to_vec(&self) -> Vec<u8>

where Self: Sized,

Encodes the message to a newly allocated buffer.

fn encode_length_delimited( &self, buf: &mut impl BufMut, ) -> Result<(), EncodeError>

where Self: Sized,

Encodes the message with a length-delimiter to a buffer.

fn encode_length_delimited_to_vec(&self) -> Vec<u8>

where Self: Sized,

Encodes the message with a length-delimiter to a newly allocated buffer.

fn decode(buf: impl Buf) -> Result<Self, DecodeError>

where Self: Default,

Decodes an instance of the message from a buffer.

fn decode_length_delimited(buf: impl Buf) -> Result<Self, DecodeError>

where Self: Default,

Decodes a length-delimited instance of the message from the buffer.

fn merge(&mut self, buf: impl Buf) -> Result<(), DecodeError>

where Self: Sized,

Decodes an instance of the message from a buffer, and merges it into self.

fn merge_length_delimited(&mut self, buf: impl Buf) -> Result<(), DecodeError>

where Self: Sized,

Decodes a length-delimited instance of the message from buffer, and merges it into self.

impl PartialEq for VectorSearchRequest

fn eq(&self, other: &VectorSearchRequest) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for VectorSearchRequest