phago-embeddings

Embedding backends for Phago semantic intelligence.

Overview

This crate provides vector embedding support for Phago:

• SimpleEmbedder — Hash-based embeddings (no external dependencies)
• OnnxEmbedder — Local ONNX runtime embeddings (optional local feature)
• ApiEmbedder — API-based embeddings (OpenAI, Voyage, Cohere) (optional api feature)
• Chunker — Document chunking with configurable overlap
• Similarity functions — cosine, euclidean, dot product

Usage

use phago_embeddings::{SimpleEmbedder, Embedder, cosine_similarity};

// Create a simple hash-based embedder
let embedder = SimpleEmbedder::new(256);

// Generate embeddings
let v1 = embedder.embed("cell membrane").unwrap();
let v2 = embedder.embed("phospholipid bilayer").unwrap();

// Compute similarity
let similarity = cosine_similarity(&v1, &v2).unwrap();

Features

Part of Phago

This is a subcrate of phago. For most use cases, depend on the main phago crate with the semantic feature instead.

License

MIT