ofxRapidLib

ofxRapidLib is an openFrameworks wrapper for the RapidLib library. RapidLib is a lightweight, interactive machine learning library intended to be used in interactive music and visual projects. It was directly inspired by Rebecca Fiebrink's Wekinator, and was written in collaboration with her at Goldsmiths, University of London, as part of the RAPID-MIX project.

RapidLib is an open source, cross-platform project and is avaiable under a BSD license.

The master branch of ofxRapidLib has been tested with:

• ofx_0.10.1, 0.11.0
• MacOS 10.14 with XCode 10
• Windows 10 with Visual Studio 2017, 2019

Documentation

RapidLib DOxygen documentation

Interactive machine learning

The interactive machine learning API has the following classes:

• classification (k-Nearest Neighbor)
• regression (Neural Network)
• seriesClassification (Dynamic Time Warping)

There are also two classes for holding the data that are used to train machine learning models:

• trainingExample
• trainingSeries

Basic signal processing

In addition to machine learning, ofxRapidLib provides users with some basic signal processing algorithms for pre-processing incoming sensor data. This is centered around a buffer class, called rapidStream. It offers the following functions:

• rapidStream.velocity() (aka first-order difference)
• rapidStream.acceleration() (aka second-order difference)
• rapidStream.minimum() _The smallest value in the buffer_
• rapidStream.maximum() _The largest value in the buffer_
• rapidStream.sum() _sum of all buffered values_
• rapidStream.mean()
• rapidStream.standardDeviation()
• rapidStream.rms() _root mean square of values in the buffer_
• rapidStream.bayesfilter(input) _Bayesian filter for EMG envelope detection_
• rapidStream.minVelocity()
• rapidStream.maxVelocity()
• rapidStream.minAcceleration()
• rapidStream.maxAcceleration()

Examples

Description of examples

JavaScript

RapidLib has been ported to JavaScript. A node module is maintained here Add it to your node app with:

npm install rapidlib

The RapidLib JavaScript library also runs client side. It is extensively documented on CodeCircle. Search for the tag "#RapidLib"

• RapidLib_001: Basic