Struct EngineBuilder 

pub struct EngineBuilder { /* private fields */ }

Builder for constructing an Engine with custom configuration.

Use this when you need to specify a custom text analyzer or embedding model. For simple cases with default settings (StandardAnalyzer, no embedder), use Engine::new directly.

Example

use std::sync::Arc;

let schema = Schema::builder()
    .add_field("content", FieldOption::Text(TextOption::default()))
    .add_field("content_vec", FieldOption::Flat(FlatOption { dimension: 384, ..Default::default() }))
    .build();

let engine = Engine::builder(storage, schema)
    .analyzer(Arc::new(StandardAnalyzer::default()))
    .embedder(Arc::new(MyEmbedder))
    .build()
    .await?;

Implementations

impl EngineBuilder

pub fn new(storage: Arc<dyn Storage>, schema: Schema) -> Self

Create a new builder with the given storage and schema.

pub fn analyzer(self, analyzer: Arc<dyn Analyzer>) -> Self

Set the analyzer for text fields.

Both simple analyzers (e.g., StandardAnalyzer) and PerFieldAnalyzer are supported. When a PerFieldAnalyzer is passed, it is used directly (with _id automatically set to KeywordAnalyzer if not already configured).

If not set, StandardAnalyzer is used as the default.

pub fn embedder(self, embedder: Arc<dyn Embedder>) -> Self

Set the embedder for vector fields.

Both simple embedders and PerFieldEmbedder are supported. When a PerFieldEmbedder is passed, each vector field will use the embedder registered for that field name, falling back to the default.

If not set, no embedder is configured.

pub async fn build(self) -> Result<Engine>

Build the Engine.

Creates the lexical store, vector store, and document log (WAL), then runs WAL recovery to replay any uncommitted changes from a previous session.

Errors

Returns an error if storage initialization, index creation, WAL opening, or recovery replay fails.