Struct FastEmbedder 

pub struct FastEmbedder { /* private fields */ }

FastEmbed-backed semantic embedder.

Supports multiple ONNX models with configurable dimensions and pooling.

Implementations

impl FastEmbedder

pub fn embedder_id_static() -> &'static str

Stable embedder identifier for MiniLM (matches vector index naming).

pub fn model_id_static() -> &'static str

Stable model identifier for MiniLM.

pub fn required_model_files() -> &'static [&'static str]

Required non-model files for any ONNX embedder.

The ONNX model itself can live at onnx/model.onnx (modern) or model.onnx (legacy) — use [select_model_file] to find it.

pub fn model_file_candidates() -> &'static [&'static str]

Candidate ONNX model locations, ordered from preferred to legacy.

pub fn select_model_file(model_dir: &Path) -> Option<PathBuf>

Select the ONNX model file, preferring onnx/model.onnx over model.onnx.

pub fn default_model_dir(data_dir: &Path) -> PathBuf

Default MiniLM model directory relative to the cass data dir.

pub fn model_dir_for(data_dir: &Path, embedder_name: &str) -> Option<PathBuf>

Get model directory for a specific embedder name.

pub fn runtime_model_dir_for( data_dir: &Path, embedder_name: &str, ) -> Option<PathBuf>

Resolve the runtime model directory for an embedder.

model_dir_for is the cass-managed cache location. This variant honors the explicit FRANKENSEARCH_MODEL_DIR override used by operators who pre-stage a model bundle outside the cass data directory.

pub fn canonical_name(embedder_name: &str) -> Option<&'static str>

pub fn config_for(embedder_name: &str) -> Option<OnnxEmbedderConfig>

Get config for a specific embedder by name.

pub fn load_from_dir(model_dir: &Path) -> EmbedderResult<Self>

Load the MiniLM model (convenience wrapper).

pub fn load_with_config( model_dir: &Path, config: OnnxEmbedderConfig, ) -> EmbedderResult<Self>

Load an ONNX embedder with custom configuration.

pub fn load_by_name( data_dir: &Path, embedder_name: &str, ) -> EmbedderResult<Self>

Load an embedder by name from the data directory.

pub fn model_id(&self) -> &str

Stable model identifier for compatibility checks.

Trait Implementations

impl SyncEmbed for FastEmbedder

fn embed_sync(&self, text: &str) -> EmbedderResult<Vec<f32>>

Synchronously embed a single text into a vector.

fn embed_batch_sync(&self, texts: &[&str]) -> EmbedderResult<Vec<Vec<f32>>>

Synchronously embed a batch of texts.

fn dimension(&self) -> usize

The output dimensionality of embedding vectors.

fn id(&self) -> &str

Unique, stable identifier for this embedder (stored in index headers).

fn model_name(&self) -> &str

Human-readable model name.

fn is_semantic(&self) -> bool

Whether this embedder produces semantically meaningful vectors.

fn category(&self) -> ModelCategory

The speed/quality category of this embedder.

fn tier(&self) -> ModelTier

Default progressive tier assignment.

fn is_ready(&self) -> bool

Whether the embedder is loaded and operational.

fn supports_mrl(&self) -> bool

Whether this model supports Matryoshka Representation Learning.