Struct SearchService 

pub struct SearchService<S, E>
where
    S: DatasetStore,
    E: EmbeddingProvider,
{ /* private fields */ }

Service for semantic search operations.

This service encapsulates the search business logic, coordinating between the embedding provider and the dataset store.

Type Parameters

• S - Dataset store implementation (e.g., DatasetRepository)
• E - Embedding provider implementation (e.g., GeminiClient)

Example

use ceres_core::search::SearchService;

let search_service = SearchService::new(repo, gemini);
let results = search_service.search("climate data", 10).await?;

for result in results {
    println!("{}: {:.2}", result.dataset.title, result.similarity_score);
}

Implementations

impl<S, E> SearchService<S, E>

where S: DatasetStore, E: EmbeddingProvider,

pub fn new(store: S, embedding: E) -> Self

Creates a new search service.

Arguments

• store - Dataset store for vector search queries
• embedding - Embedding provider for generating query embeddings

pub async fn search( &self, query: &str, limit: usize, ) -> Result<Vec<SearchResult>, AppError>

Performs semantic search and returns ranked results.

This method:

1. Generates an embedding vector from the query text
2. Searches the database using cosine similarity
3. Returns results ordered by similarity (highest first)

Arguments

• query - The search query text
• limit - Maximum number of results to return

Returns

A vector of SearchResult, ordered by similarity score (highest first).

Errors

Returns an error if:

• The embedding generation fails (API error, network error, etc.)
• The database query fails

Trait Implementations

impl<S, E> Clone for SearchService<S, E>

where S: DatasetStore + Clone, E: EmbeddingProvider + Clone,

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.