🌪️ Vortex Search Engine

A powerful, lightweight, persistent vector-space search engine with NLP capabilities.

Vortex is a pure JavaScript search engine designed for speed and flexibility. It features BM25 ranking, fuzzy search (typo tolerance), synonym expansion, and persistence.

Features 🚀

• BM25 Ranking: State-of-the-art probabilistic ranking (better than TF-IDF).
• Fuzzy Search: Auto-corrects typos (e.g., "cmoputer" -> "computer").
• Conversational NLP: Ignores filler words ("hey", "please") and handles natural language questions.
• Synonym Expansion: Automatically expands terms (e.g., "fix" -> "repair").
• Persistence: Save and load your search index to JSON.
• Zero Dependencies: Pure Node.js.

Installation 📦

npm install vortex-search-engine

Or use the browser bundle via CDN (or download from dist/):

<script src="dist/vortex.browser.js"></script>

Usage 🛠️

Node.js

const { VortexEngine } = require("vortex-search-engine");
const engine = new VortexEngine();

Browser

<script src="dist/vortex.browser.js"></script>
<script>
    const engine = new VortexEngine();
</script>

Advanced Features

Fuzzy Search & Typos

Vortex automatically detects and corrects typos.

// User types "fiannce" -> Matches "Finance"
const results = engine.search("fiannce");

Persistence (Save/Load)

Save your index to disk so you don't have to re-index every time.

// Save
engine.save("search_index.json");

// Load
const newEngine = new VortexEngine();
newEngine.load("search_index.json");

API Reference 📚

addDocument(id, title, content)

Adds a document to the index.

• id: Unique identifier for the document.
• title: Document title (boosted in search).
• content: The main text content.

search(query)

Searches for the query string. Returns an array of results sorted by relevance score.

save(filepath)

Saves the current engine state (index, documents, model) to a JSON file.

load(filepath)

Loads the engine state from a JSON file.

License 📄

MIT © Sethun Thunder