Struct SemanticQuery 

pub struct SemanticQuery;

Builds SphereQL Regions from semantic constraints on embeddings.

Implementations

impl SemanticQuery

pub fn within_angle<P: Projection>( query: &Embedding, projection: &P, max_angular_distance: f64, ) -> Region

Spherical cap: all points within max_angular_distance radians of the query.

pub fn above_similarity<P: Projection>( query: &Embedding, projection: &P, min_similarity: f64, ) -> Region

Spherical cap from a cosine similarity threshold. cos_sim >= threshold ↔ angular_distance <= arccos(threshold).

pub fn in_shell(inner: f64, outer: f64) -> Region

Radial shell: embeddings whose projected radius falls in [inner, outer].

Panics

Panics if inner > outer or either bound is negative. Both are caller contracts; use Shell::new directly if you need a Result.

pub fn similar_in_shell<P: Projection>( query: &Embedding, projection: &P, min_similarity: f64, shell_inner: f64, shell_outer: f64, ) -> Region

Intersection of a similarity cap with a radial shell. “Semantically similar AND within a magnitude/metadata range.”