Struct EmbeddingSettings

pub struct EmbeddingSettings {

    pub source: Setting<EmbedderSource>,
    pub model: Setting<String>,
    pub revision: Setting<String>,
    pub pooling: Setting<OverridePooling>,
    pub api_key: Setting<String>,
    pub dimensions: Setting<usize>,
    pub binary_quantized: Setting<bool>,
    pub document_template: Setting<String>,
    pub document_template_max_bytes: Setting<usize>,
    pub url: Setting<String>,
    pub request: Setting<Value>,
    pub response: Setting<Value>,
    pub headers: Setting<BTreeMap<String, String>>,
    pub search_embedder: Setting<SubEmbeddingSettings>,
    pub indexing_embedder: Setting<SubEmbeddingSettings>,
    pub distribution: Setting<DistributionShift>,
}

Fields

source: Setting<EmbedderSource>

The source used to provide the embeddings.

Which embedder parameters are available and mandatory is determined by the value of this setting.

🔄 Reindexing

• 🏗️ Changing the value of this parameter always regenerates embeddings.

Defaults

• Defaults to openAi

model: Setting<String>

The name of the model to use.

Mandatory

• This parameter is mandatory for source ollama

Availability

• This parameter is available for sources openAi, huggingFace, ollama

🔄 Reindexing

• 🏗️ Changing the value of this parameter always regenerates embeddings.

Defaults

• For source openAi, defaults to text-embedding-3-small
• For source huggingFace, defaults to BAAI/bge-base-en-v1.5

revision: Setting<String>

The revision (commit SHA1) of the model to use.

If unspecified, Meilisearch picks the latest revision of the model.

Availability

• This parameter is available for source huggingFace

🔄 Reindexing

• 🏗️ Changing the value of this parameter always regenerates embeddings

Defaults

• When model is set to default, defaults to 617ca489d9e86b49b8167676d8220688b99db36e
• Otherwise, defaults to null

pooling: Setting<OverridePooling>

The pooling method to use.

Availability

• This parameter is available for source huggingFace

🔄 Reindexing

• 🏗️ Changing the value of this parameter always regenerates embeddings

Defaults

• Defaults to useModel

Compatibility Note

• Embedders created before this parameter was available default to forceMean to preserve the existing behavior.

api_key: Setting<String>

The API key to pass to the remote embedder while making requests.

Availability

• This parameter is available for source openAi, ollama, rest

🔄 Reindexing

• 🌱 Changing the value of this parameter never regenerates embeddings

Defaults

• For source openAi, the key is read from OPENAI_API_KEY, then MEILI_OPENAI_API_KEY.
• For other sources, no bearer token is sent if this parameter is not set.

Note

• This setting is partially hidden when returned by the settings

dimensions: Setting<usize>

The expected dimensions of the embeddings produced by this embedder.

Mandatory

• This parameter is mandatory for source userProvided

Availability

• This parameter is available for source openAi, ollama, rest, userProvided

🔄 Reindexing

• 🏗️ When the source is openAi, changing the value of this parameter always regenerates embeddings
• 🌱 For other sources, changing the value of this parameter never regenerates embeddings

Defaults

• For source openAi, the dimensions is the maximum allowed by the model.
• For sources ollama and rest, the dimensions are inferred by embedding a sample text.

binary_quantized: Setting<bool>

Whether to binary quantize the embeddings of this embedder.

Binary quantized embeddings are smaller than regular embeddings, which improves disk usage and retrieval speed, at the cost of relevancy.

Availability

• This parameter is available for all embedders

🔄 Reindexing

• 🏗️ When set to true, embeddings are not regenerated, but they are binary quantized, which takes time.

Defaults

• Defaults to false

Note

As binary quantization is a destructive operation, it is not possible to disable again this setting after first enabling it. If you are unsure of whether the performance-relevancy tradeoff is right for you, we recommend to use this parameter on a test index first.

document_template: Setting<String>

A liquid template used to render documents to a text that can be embedded.

Meillisearch interpolates the template for each document and sends the resulting text to the embedder. The embedder then generates document vectors based on this text.

Availability

• This parameter is available for source openAi, huggingFace, ollama and `rest

🔄 Reindexing

• 🏗️ When modified, embeddings are regenerated for documents whose rendering through the template produces a different text.

document_template_max_bytes: Setting<usize>

Rendered texts are truncated to this size.

Availability

• This parameter is available for source openAi, huggingFace, ollama and rest

🔄 Reindexing

• 🏗️ When increased, embeddings are regenerated for documents whose rendering through the template produces a different text.
• 🌱 When decreased, embeddings are never regenerated

Default

• Defaults to 400

url: Setting<String>

URL to reach the remote embedder.

Mandatory

• This parameter is mandatory for source rest

Availability

• This parameter is available for source openAi, ollama and rest

🔄 Reindexing

• 🌱 When modified for source openAi, embeddings are never regenerated
• 🏗️ When modified for sources ollama and rest, embeddings are always regenerated

request: Setting<Value>

Template request to send to the remote embedder.

Mandatory

• This parameter is mandatory for source rest

Availability

• This parameter is available for source rest

🔄 Reindexing

• 🏗️ Changing the value of this parameter always regenerates embeddings

response: Setting<Value>

Template response indicating how to find the embeddings in the response from the remote embedder.

Mandatory

• This parameter is mandatory for source rest

Availability

• This parameter is available for source rest

🔄 Reindexing

• 🏗️ Changing the value of this parameter always regenerates embeddings

headers: Setting<BTreeMap<String, String>>

Additional headers to send to the remote embedder.

Availability

• This parameter is available for source rest

🔄 Reindexing

• 🌱 Changing the value of this parameter never regenerates embeddings

search_embedder: Setting<SubEmbeddingSettings>

indexing_embedder: Setting<SubEmbeddingSettings>

distribution: Setting<DistributionShift>

Affine transformation applied to the semantic score to make it more comparable to the ranking score.

Availability

• This parameter is available for all embedders

🔄 Reindexing

• 🌱 Changing the value of this parameter never regenerates embeddings

Trait Implementations

impl Clone for EmbeddingSettings

fn clone(&self) -> EmbeddingSettings

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl ComposeSchema for EmbeddingSettings

fn compose(generics: Vec<RefOr<Schema>>) -> RefOr<Schema>

impl Debug for EmbeddingSettings

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

Formats the value using the given formatter.

impl Default for EmbeddingSettings

fn default() -> EmbeddingSettings

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for EmbeddingSettings

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

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl<__Deserr_E> Deserr<__Deserr_E> for EmbeddingSettings

where __Deserr_E: DeserializeError,

fn deserialize_from_value<V: IntoValue>( deserr_value__: Value<V>, deserr_location__: ValuePointerRef<'_>, ) -> Result<Self, __Deserr_E>

Attempts to deserialize Self from the given value. Note that this method is an implementation detail. You probably want to use the deserialize function directly instead.

impl From<EmbeddingConfig> for EmbeddingSettings

fn from(value: EmbeddingConfig) -> Self

Converts to this type from the input type.

impl From<EmbeddingSettings> for EmbeddingConfig

fn from(value: EmbeddingSettings) -> Self

Converts to this type from the input type.

impl From<EmbeddingSettings> for SubEmbeddingSettings

fn from(value: EmbeddingSettings) -> Self

Converts to this type from the input type.

impl PartialEq for EmbeddingSettings

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for EmbeddingSettings

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

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl ToSchema for EmbeddingSettings

fn name() -> Cow<'static, str>

Return name of the schema.

fn schemas(schemas: &mut Vec<(String, RefOr<Schema>)>)

Implement reference utoipa::openapi::schema::Schemas for this type.

impl Eq for EmbeddingSettings

impl StructuralPartialEq for EmbeddingSettings