embedding.js's Introduction

embedding.js

NOTE: This repo has been archived and split into two: embeddings.js and vectordb.js

A simple in-memory embedding database that works with OpenAI's text-embedding-ada-002 text embeddings, built on top of hnswlib-node. Useful for finding relevant documents to include in gpt-3.5-turbo and gpt-4 context windows.

Features

Fast approximate nearest neighbor search using hierarchical navigable small world graphs.
Utilizes OpenAI's text-embedding-ada-002 model for text embeddings.
Easy-to-use API for adding and searching data in the database.

Installation

npm install --save @themaximalist/embedding.js

Configuration

To use this module, you will need an API key from OpenAI. Set the OPENAI_API_KEY environment variable with your API key:

export OPENAI_API_KEY=<your-openai-api-key>

Usage

const embedding = require("@themaximalist/embedding.js");

(async function () {
    const embeddings = new embedding.EmbeddingDatabase();
    await embeddings.add({
        name: "Cat",
        attributes: "It's a cat",
        sound: "meow",
    });

    await embeddings.add({
        name: "Dog",
        attributes: "It's a dog",
        sound: "woof",
    });

    await embeddings.add({
        name: "Cow",
        attributes: "It's a cow",
        sound: "moo",
    });

    let result;

    result = await embeddings.search("moo");
    console.log(result[0]); // cow

    result = await embeddings.search("woof");
    console.log(result[0]); // dog

    result = await embeddings.search("bark");
    console.log(result[0]); // dog

    result = await embeddings.search("roar");
    console.log(result[0]); // cat
})();