Struct ContentRecommender 

pub struct ContentRecommender { /* private fields */ }

Content-based recommender using TF-IDF + HNSW.

Type Parameters

None - uses f64 internally for maximum precision

Configuration

• m: HNSW connections per node (12-48 recommended)
• ef_construction: HNSW construction parameter (100-200 recommended)
• decay_factor: IDF decay factor (0.9-0.99 recommended)

Implementations

impl ContentRecommender

pub fn new(m: usize, ef_construction: usize, decay_factor: f64) -> Self

Create a new content-based recommender.

Arguments

• m - HNSW connections per node (12-48 recommended)
• ef_construction - HNSW construction parameter (100-200 recommended)
• decay_factor - IDF decay factor (0.9-0.99 recommended)

Examples

use aprender::recommend::ContentRecommender;

let rec = ContentRecommender::new(16, 200, 0.95);

pub fn add_item( &mut self, item_id: impl Into<String>, content: impl Into<String>, )

Add an item to the recommender.

Arguments

• item_id - Unique identifier for the item
• content - Text content to extract features from

Examples

use aprender::recommend::ContentRecommender;

let mut rec = ContentRecommender::new(16, 200, 0.95);
rec.add_item("item1", "machine learning tutorial");
rec.add_item("item2", "deep learning guide");

pub fn recommend( &self, item_id: &str, k: usize, ) -> Result<Vec<(String, f64)>, AprenderError>

Recommend similar items.

Arguments

• item_id - Item to find similar items for
• k - Number of recommendations to return

Returns

List of (item_id, similarity_score) pairs, sorted by similarity (highest first)

Examples

use aprender::recommend::ContentRecommender;

let mut rec = ContentRecommender::new(16, 200, 0.95);
rec.add_item("a", "machine learning");
rec.add_item("b", "deep learning");
rec.add_item("c", "machine intelligence");

let similar = rec.recommend("a", 2).expect("recommend should succeed");
assert_eq!(similar.len(), 2);

pub fn len(&self) -> usize

Number of items in the recommender.

Examples

use aprender::recommend::ContentRecommender;

let mut rec = ContentRecommender::new(16, 200, 0.95);
assert_eq!(rec.len(), 0);

rec.add_item("item1", "content");
assert_eq!(rec.len(), 1);

pub fn is_empty(&self) -> bool

Check if recommender is empty.

Examples

use aprender::recommend::ContentRecommender;

let rec = ContentRecommender::new(16, 200, 0.95);
assert!(rec.is_empty());

Trait Implementations

impl Debug for ContentRecommender

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

Formats the value using the given formatter.