Enum EmbeddingSpec 

pub enum EmbeddingSpec {
    Dense {
        model: Option<String>,
        vector: Option<String>,
        field: Option<String>,
    },
    Sparse {
        model: Option<String>,
        vector: Option<String>,
        field: Option<String>,
    },
    Hybrid {
        dense_model: Option<String>,
        dense_vector: Option<String>,
        dense_field: Option<String>,
        sparse_model: Option<String>,
        sparse_vector: Option<String>,
        sparse_field: Option<String>,
    },
    MultiVector {
        model: Option<String>,
        vector: Option<String>,
        field: Option<String>,
    },
    Image {
        model: Option<String>,
        vector: Option<String>,
        field: Option<String>,
    },
    Multi(Vec<EmbeddingSpec>),
}

Upsert-level USING embedding clause.

Variants

Dense

USING DENSE / USING MODEL / USING VECTOR — dense embedding.

Fields

model: Option<String>

Optional embedding model.

vector: Option<String>

Optional target vector name.

field: Option<String>

Optional payload field to embed.

Sparse

USING SPARSE — sparse embedding.

Fields

model: Option<String>

Optional sparse model (e.g. BM25-compatible).

vector: Option<String>

Optional target vector name.

field: Option<String>

Optional payload field to embed.

Hybrid

USING HYBRID — parallel dense + sparse embedding.

Fields

dense_model: Option<String>

Optional dense embedding model.

dense_vector: Option<String>

Optional dense target vector name.

dense_field: Option<String>

Optional dense input payload field.

sparse_model: Option<String>

Optional sparse embedding model.

sparse_vector: Option<String>

Optional sparse target vector name.

sparse_field: Option<String>

Optional sparse input payload field.

MultiVector

Multivector / ColBERT: text → bag of token vectors for a named multi slot.

Fields

model: Option<String>

Optional multivector embedding model.

vector: Option<String>

Optional target multivector name.

field: Option<String>

Optional payload field to embed.

Image

Image / CLIP vision: payload field holds a path or URL → dense vector.

Fields

model: Option<String>

Optional CLIP vision model.

vector: Option<String>

Optional target dense vector name.

field: Option<String>

Optional payload field holding the image path or URL.

Multi(Vec<EmbeddingSpec>)

Combined specs (e.g. DENSE + SPARSE + MULTI VECTOR colbert).

Trait Implementations

impl Clone for EmbeddingSpec

fn clone(&self) -> EmbeddingSpec

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for EmbeddingSpec

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

Formats the value using the given formatter.

impl PartialEq for EmbeddingSpec

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

Equality operator ==.

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

Inequality operator !=.

impl StructuralPartialEq for EmbeddingSpec