Enum TextOp 

pub enum TextOp {
    Search {
        collection: QualifiedCollection,
        field: Option<String>,
        query: String,
        top_k: usize,
        mode: QueryMode,
        fuzzy: bool,
        prefilter: Option<SurrogateBitmap>,
        filters: Vec<u8>,
        rls_filters: Vec<u8>,
        scores: Vec<TextScoreSpec>,
    },
    BM25ScoreScan {
        collection: QualifiedCollection,
        filters: Vec<u8>,
        rls_filters: Vec<u8>,
        scores: Vec<TextScoreSpec>,
        bound: Option<ScoreScanBound>,
    },
    PhraseSearch {
        collection: QualifiedCollection,
        field: Option<String>,
        terms: Vec<String>,
        top_k: usize,
        prefilter: Option<SurrogateBitmap>,
        filters: Vec<u8>,
        rls_filters: Vec<u8>,
        scores: Vec<TextScoreSpec>,
    },
    HybridSearch {

        collection: QualifiedCollection,
        vector_field: String,
        query_vector: Vec<f32>,
        text_field: Option<String>,
        query_text: String,
        filters: Vec<u8>,
        top_k: usize,
        ef_search: usize,
        mode: QueryMode,
        fuzzy: bool,
        vector_weight: f32,
        filter_bitmap: Option<SurrogateBitmap>,
        rls_filters: Vec<u8>,
        score_alias: Option<String>,
    
},
    FtsIndexDoc {
        collection: QualifiedCollection,
        surrogate: Surrogate,
        fields: Vec<(String, String)>,
        provenance: Option<SyncProvenance>,
    },
    FtsDeleteDoc {
        collection: QualifiedCollection,
        surrogate: Option<Surrogate>,
        provenance: Option<SyncProvenance>,
    },
    HybridSearchTriple {

        collection: QualifiedCollection,
        vector_field: String,
        query_vector: Vec<f32>,
        text_field: Option<String>,
        query_text: String,
        filters: Vec<u8>,
        graph_seed_id: String,
        graph_depth: usize,
        graph_edge_label: Option<String>,
        top_k: usize,
        ef_search: usize,
        mode: QueryMode,
        fuzzy: bool,
        rrf_k: (f64, f64, f64),
        filter_bitmap: Option<SurrogateBitmap>,
        rls_filters: Vec<u8>,
        score_alias: Option<String>,
    
},
    SetTextConfig {
        collection: QualifiedCollection,
        analyzer_name: Option<String>,
        fuzzy_default: Option<bool>,
    },
}

Full-text search physical operations.

Variants

Search

BM25 full-text search on the inverted index.

Fields

collection: QualifiedCollection

field: Option<String>

Field index the query reads. None reads the whole-document index.

query: String

top_k: usize

Hits returned, best first. usize::MAX returns every match.

mode: QueryMode

Boolean combination of the query terms.

fuzzy: bool

Enable fuzzy matching (Levenshtein) for typo tolerance.

prefilter: Option<SurrogateBitmap>

Pre-computed bitmap of eligible surrogates (from prefilter evaluation). None = no prefilter; all postings are eligible.

filters: Vec<u8>

Residual WHERE predicates (serialized Vec<ScanFilter>). They restrict candidates before ranking, so top_k counts only rows that satisfy them.

rls_filters: Vec<u8>

RLS filters (serialized Vec<ScanFilter>). Like filters, they restrict candidates before ranking, so top_k counts only rows the policy admits.

scores: Vec<TextScoreSpec>

Score columns injected into each hit.

BM25ScoreScan

Every row filters admit, each with its score columns.

The physical plan for bm25_score(field, term) with no text_match: every admitted row is present. A row the score’s index holds but its query does not match carries 0.0. A row the index does not hold carries null.

Fields

collection: QualifiedCollection

filters: Vec<u8>

Residual WHERE predicates (serialized Vec<ScanFilter>).

rls_filters: Vec<u8>

RLS filters (serialized Vec<ScanFilter>). A row that fails one is dropped.

scores: Vec<TextScoreSpec>

Score columns injected into each row.

bound: Option<ScoreScanBound>

The query’s LIMIT pushed into the scan. None returns every admitted row. The relational tail still applies its own ORDER BY and LIMIT over the rows returned.

PhraseSearch

Exact phrase search: all terms must appear consecutively in the document.

Unlike Search (BM25 scoring), phrase search returns only documents where the query terms appear as an exact contiguous sequence. Scoring is positional: documents with the phrase closer to the start rank higher.

Fields

collection: QualifiedCollection

field: Option<String>

Field index the phrase reads. None reads the whole-document index.

terms: Vec<String>

The phrase words as written, in order. The Data Plane analyzes them once with the collection’s analyzer and matches the tokens as a contiguous sequence.

top_k: usize

Hits returned, best first. usize::MAX returns every match.

prefilter: Option<SurrogateBitmap>

Pre-computed bitmap of eligible surrogates (from prefilter evaluation).

filters: Vec<u8>

Residual WHERE predicates (serialized Vec<ScanFilter>), applied before ranking.

rls_filters: Vec<u8>

RLS filters (serialized Vec<ScanFilter>), applied before ranking.

scores: Vec<TextScoreSpec>

Score columns injected into each hit.

HybridSearch

Hybrid search: vector similarity + BM25 text, fused via RRF.

Fields

collection: QualifiedCollection

vector_field: String

Vector column the vector leg searches.

query_vector: Vec<f32>

text_field: Option<String>

Field index the text leg reads. None reads the whole-document index.

query_text: String

filters: Vec<u8>

Residual WHERE predicates (serialized Vec<ScanFilter>). They restrict both legs before fusion.

top_k: usize

ef_search: usize

mode: QueryMode

Boolean combination of the text leg’s query terms.

fuzzy: bool

vector_weight: f32

Weight for vector results in RRF (0.0–1.0). Default: 0.5.

filter_bitmap: Option<SurrogateBitmap>

rls_filters: Vec<u8>

RLS post-fusion filters.

score_alias: Option<String>

SELECT-list alias the response should use for the RRF score column. None means the executor uses the fixed internal name rrf_score. Set by the planner from the SELECT alias for the rrf_score(...) call.

FtsIndexDoc

Index a document into the inverted FTS index.

Used by the sync path when a Lite client sends an FtsIndex frame. Origin assigns a surrogate for (collection, doc_id) on the Control Plane before dispatch; surrogate is the pre-assigned value.

Fields

collection: QualifiedCollection

surrogate: Surrogate

Pre-assigned global surrogate for (collection, doc_id).

fields: Vec<(String, String)>

(field, text) per top-level string field. Empty removes the document from every index.

provenance: Option<SyncProvenance>

Sync provenance: identifies the originating peer and sequence for idempotency.

FtsDeleteDoc

Remove a document from the inverted FTS index.

Used by the sync path when a Lite client sends an FtsDelete frame.

Fields

collection: QualifiedCollection

surrogate: Option<Surrogate>

The global surrogate (collection, doc_id) is bound to. None when the key’s home binds none: the delete removes nothing and still commits its sync provenance.

provenance: Option<SyncProvenance>

Sync provenance: identifies the originating peer and sequence for idempotency.

HybridSearchTriple

Three-source hybrid search: vector + BM25 text + graph BFS, fused via weighted RRF.

Extends HybridSearch with an optional graph BFS leg. The graph leg performs a BFS from graph_seed_id up to graph_depth hops, filtering edges by graph_edge_label when set. All three ranked lists are passed to reciprocal_rank_fusion_weighted with per-source k-constants.

Fields

collection: QualifiedCollection

vector_field: String

Vector column the vector leg searches.

query_vector: Vec<f32>

text_field: Option<String>

Field index the text leg reads. None reads the whole-document index.

query_text: String

filters: Vec<u8>

Residual WHERE predicates (serialized Vec<ScanFilter>). They restrict every leg before fusion.

graph_seed_id: String

Node id used as the BFS seed for the graph leg.

graph_depth: usize

Maximum BFS depth from the seed node.

graph_edge_label: Option<String>

Edge label filter for graph BFS. None = all edges.

top_k: usize

ef_search: usize

mode: QueryMode

Boolean combination of the text leg’s query terms.

fuzzy: bool

rrf_k: (f64, f64, f64)

Per-source RRF k constants: (vector_k, text_k, graph_k).

filter_bitmap: Option<SurrogateBitmap>

rls_filters: Vec<u8>

RLS post-fusion filters.

score_alias: Option<String>

SELECT-list alias for the fused RRF score column.

SetTextConfig

Bind a collection’s per-collection FTS analyzer.

Persists to backend metadata (InvertedIndex::set_collection_analyzer) so InvertedIndex::analyze_for_collection resolves it for every subsequent tokenization of the collection’s text — forward indexing, the in-transaction staged-write overlay, and query-time scoring alike. Config-only, single-node, non-WAL-durable — the same dispatch shape VectorOp::SetParams uses for CREATE VECTOR INDEX.

Fields

collection: QualifiedCollection

analyzer_name: Option<String>

Analyzer name, already checked against nodedb_fts::index::analyzer_config::analyzer_exists by the DDL layer (e.g. “standard”, “english”, “japanese”). None leaves the collection’s current analyzer in place.

fuzzy_default: Option<bool>

Whether searches over this collection fall back to fuzzy matching by default. None leaves the current setting in place.

Trait Implementations

impl Clone for TextOp

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 TextOp

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for TextOp

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl<'__msgpack_de> FromMessagePack<'__msgpack_de> for TextOp

fn read<R: Read<'__msgpack_de>>(reader: &mut R) -> Result<Self, Error>

where Self: Sized,

Reads the MessagePack representation of this value from the provided reader.

impl PartialEq for TextOp

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

Equality operator ==.

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

Inequality operator !=.

impl Serialize for TextOp

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for TextOp

impl ToMessagePack for TextOp

fn write<W: Write>(&self, writer: &mut W) -> Result<(), Error>

Writes the MessagePack representation of this value into the provided writer.

fn write_slice<W>(values: &[Self], writer: &mut W) -> Result<(), Error>

where W: Write, Self: Sized,

Writes the MessagePack representation of a slice of values into the provided writer.

fn size_hint(&self) -> Option<TrustedSizeHint>

Returns a trusted upper bound for the number of bytes written by the next call to Self::write, or None when it cannot be determined cheaply.

fn max_size() -> Option<TrustedSizeHint>

where Self: Sized,

Returns a trusted upper bound valid for every value of this type.