Crate embedcache 

EmbedCache - High-performance text embedding library with caching capabilities

This library provides functionality for generating text embeddings with various state-of-the-art models and caching the results for improved performance.

Features

• Multiple embedding models (BGE, MiniLM, Nomic, etc.)
• Modular text chunking strategies with extensible trait-based architecture
• LLM-based intelligent chunking (concept and introspection modes)
• SQLite-based caching
• Asynchronous operation

Installation

Add this to your Cargo.toml:

[dependencies]
embedcache = "0.1"

Examples

Using as a library

use embedcache::{FastEmbedder, Embedder};
use fastembed::{InitOptions, EmbeddingModel};

let embedder = FastEmbedder {
    options: InitOptions::new(EmbeddingModel::BGESmallENV15),
};

let texts = vec![
    "This is an example sentence.".to_string(),
    "Another example sentence for embedding.".to_string(),
];

let embeddings = embedder.embed(&texts).await?;

Implementing custom chunking strategies

use embedcache::ContentChunker;
use async_trait::async_trait;

struct MyCustomChunker;

#[async_trait]
impl ContentChunker for MyCustomChunker {
    async fn chunk(&self, content: &str, size: usize) -> Vec<String> {
        // Your custom chunking logic here
        vec![content.to_string()]
    }

    fn name(&self) -> &str {
        "my-custom-chunker"
    }
}

Running the embedcache service

The library also includes a binary that can be run as a standalone service:

cargo install embedcache
embedcache

Re-exports

pub use cache::cache_result;

pub use cache::get_from_cache;

pub use cache::initialize_db_pool;

pub use chunking::ContentChunker;

pub use chunking::LLMConceptChunker;

pub use chunking::LLMConfig;

pub use chunking::LLMIntrospectionChunker;

pub use chunking::WordChunker;

pub use config::ServerConfig;

pub use embedding::get_embedding_model;

pub use embedding::initialize_models;

pub use embedding::Embedder;

pub use embedding::FastEmbedder;

pub use embedding::SUPPORTED_MODELS;

pub use handlers::embed_text;

pub use handlers::list_supported_features;

pub use handlers::process_url;

pub use models::get_default_config;

pub use models::AppState;

pub use models::Config;

pub use models::InputData;

pub use models::InputDataText;

pub use models::ProcessedContent;

pub use utils::fetch_content;

pub use utils::generate_hash;

Modules

cache

Cache module for storing processed content

chunking

Content chunking module

config

Server configuration module

embedding

Embedding generation module

handlers

HTTP request handlers module

models

Data models module

utils

Utility functions

Functions

initialize_chunkers

Initialize chunkers