Struct RouterConfig 

pub struct RouterConfig {
    pub dimension: usize,
    pub metric: DistanceMetric,
    pub max_connections: usize,
    pub ef_construction: usize,
    pub ef_search: usize,
    pub cache_size: usize,
}

Configuration for semantic router

Fields

dimension: usize

Vector dimension for embeddings

metric: DistanceMetric

Distance metric to use

max_connections: usize

Maximum connections per HNSW layer

ef_construction: usize

HNSW construction parameter

ef_search: usize

Default search parameter

cache_size: usize

Query result cache size (number of queries to cache)

Implementations

impl RouterConfig

pub fn low_latency(dimension: usize) -> Self

Create configuration optimized for low latency queries

Best for: Real-time applications, interactive search, chat systems Trade-offs: Slightly lower recall (~90%), faster queries

Example

use ipfrs_semantic::RouterConfig;

let config = RouterConfig::low_latency(768);
assert_eq!(config.dimension, 768);
// Optimized for speed with reasonable accuracy

pub fn high_recall(dimension: usize) -> Self

Create configuration optimized for high recall (accuracy)

Best for: Research applications, critical retrieval, high-quality recommendations Trade-offs: Slower queries, higher memory usage

Example

use ipfrs_semantic::RouterConfig;

let config = RouterConfig::high_recall(768);
assert_eq!(config.dimension, 768);
// Optimized for accuracy with acceptable latency

pub fn memory_efficient(dimension: usize) -> Self

Create configuration optimized for memory efficiency

Best for: Edge devices, constrained environments, large datasets with limited RAM Trade-offs: Lower recall (~85%), smaller cache

Example

use ipfrs_semantic::RouterConfig;

let config = RouterConfig::memory_efficient(384);
assert_eq!(config.dimension, 384);
// Smaller connections and cache for low memory footprint

pub fn large_scale(dimension: usize) -> Self

Create configuration optimized for large-scale datasets (100k+ vectors)

Best for: Production systems, large knowledge bases, document collections Trade-offs: Higher memory usage, optimized for throughput

Example

use ipfrs_semantic::RouterConfig;

let config = RouterConfig::large_scale(768);
assert_eq!(config.dimension, 768);
// Balanced for large datasets with good performance

pub fn balanced(dimension: usize) -> Self

Create configuration for balanced performance (alias for default)

Best for: General purpose, getting started, typical applications Trade-offs: Balanced recall (~95%) and latency

Example

use ipfrs_semantic::RouterConfig;

let config = RouterConfig::balanced(768);
assert_eq!(config.dimension, 768);
// Good all-around configuration

pub fn with_metric(dimension: usize, metric: DistanceMetric) -> Self

Create configuration with custom distance metric

Example

use ipfrs_semantic::{RouterConfig, DistanceMetric};

let config = RouterConfig::with_metric(768, DistanceMetric::L2);
assert_eq!(config.dimension, 768);

pub fn with_cache_size(self, size: usize) -> Self

Set the query result cache size

Example

use ipfrs_semantic::RouterConfig;

let config = RouterConfig::balanced(768).with_cache_size(5000);
assert_eq!(config.cache_size, 5000);

pub fn with_ef_search(self, ef_search: usize) -> Self

Set the ef_search parameter for query-time search

Higher values improve recall but increase latency

Example

use ipfrs_semantic::RouterConfig;

let config = RouterConfig::balanced(768).with_ef_search(100);
assert_eq!(config.ef_search, 100);

Trait Implementations

impl Clone for RouterConfig

fn clone(&self) -> RouterConfig

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for RouterConfig

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

Formats the value using the given formatter.

impl Default for RouterConfig

fn default() -> Self

Returns the “default value” for a type.