Europa 🤖

A lightweight AI utilities library for Rust, providing essential tools for working with embeddings, vector operations, and text processing.

Features

• Vector Embeddings: Work with AI model embeddings

  • Cosine similarity calculations
  • Euclidean distance measurements
  • Serialization support via serde

• Text Processing: Simple tokenization utilities

  • Whitespace tokenization
  • Token counting
  • Lowercase normalization

Notice

⚠️ This crate is in early development.

Installation

Add Europa to your Cargo.toml:

[dependencies]
europa = { version = "0.0.2", default-features = false, features = ["vector-embedding"] }

Quick Start

use europa::prelude::*;

// Create embeddings
let embedding1 = VectorEmbedding::new(vec![1.0, 0.5, 0.2], "my-model");
let embedding2 = VectorEmbedding::new(vec![0.9, 0.6, 0.1], "my-model");

// Calculate similarity
let similarity = embedding1.cosine_similarity(&embedding2);
println!("Cosine similarity: {}", similarity);

// Calculate distance
let distance = embedding1.euclidean_distance(&embedding2);
println!("Euclidean distance: {}", distance);

use europa::prelude::*;

// Tokenize text
let tokens = Tokenizer::tokenize("Hello, AI World!");
println!("Tokens: {:?}", tokens);

// Count tokens
let count = Tokenizer::count_tokens("This is a test");
println!("Token count: {}", count);

Use Cases

• Semantic search and similarity matching
• Vector database operations
• AI model output processing
• Text preprocessing pipelines
• Embedding comparison and analysis

Roadmap

• Batch embedding operations
• Additional distance metrics (Manhattan, dot product)
• Advanced tokenization (BPE, WordPiece)
• Prompt templates and chains
• Vector database integration helpers
• Model quantization utilities

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

License

Licensed under either of:

• Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.