3 releases

0.1.2 Feb 9, 2024
0.1.1 Nov 24, 2023
0.1.0 Apr 12, 2023

#305 in Machine learning

26 downloads per month
Used in zbusdg

MIT/Apache

25KB
381 lines

tests Crates.io docs.rs

About rhai-ml

This crate provides some basic machine learning and artificial intelligence utilities for the Rhai scripting language. For a complete API reference, check the docs.

Install

To use the latest released version of rhai-ml, add this to your Cargo.toml:

rhai-ml = "0.1.2"

To use the bleeding edge instead, add this:

rhai-ml = { git = "https://github.com/cmccomb/rhai-ml" }

Usage

Using this crate is pretty simple! If you just want to evaluate a single line of Rhai, then you only need:

use rhai::FLOAT;
use rhai_ml::eval;
let result = eval::<FLOAT>("\
let xdata = [[1.0, 2.0], [2.0, 3.0], [3.0, 4.0]]; \
let ydata = [1.0, 2.0, 3.0]; \
let model = train(xdata, ydata, \"linear\"); \
let ypred = predict(xdata, model);
ypred[0]
").unwrap();

If you need to use rhai-ml as part of a persistent Rhai scripting engine, then do this instead:

use rhai::{Engine, packages::Package, FLOAT};
use rhai_ml::MLPackage;

// Create a new Rhai engine
let mut engine = Engine::new();

// Add the rhai-ml package to the new engine
engine.register_global_module(MLPackage::new().as_shared_module());

// Now run your code
let value = engine.eval::<FLOAT>("\
let xdata = [[1.0, 2.0], [2.0, 3.0], [3.0, 4.0]]; \
let ydata = [1.0, 2.0, 3.0]; \
let model = train(xdata, ydata, \"linear\"); \
let ypred = predict(xdata, model);
ypred[0]
").unwrap();

Features

Feature Default Description
metadata Disabled Enables exporting function metadata and is necessary for running doc-tests on Rhai examples.

Dependencies

~7.5MB
~140K SLoC