trackormouse

A library for detecting whether the current input is from a physical mouse or trackpad based on javascript wheel events.

Since browsers do not provide a direct API signal indicating the type of pointing device, this library uses heuristics to determine that.

Usage

import { WheelClassifier } from 'trackormouse'

const classifier = new WheelClassifier()

window.addEventListener('wheel', (e) => {
    classifier.addEvent(e)

    const deviceType = classifier.inferDeviceType()
    console.log('Detected device type:', deviceType)
})

Signals

The signals used to determine the pointing device type is exposed via the WheelClassifier class, which allows usages to customize the classification based on their specific needs.

For example, some use cases may want to set a minimum number of events before changing the configuration based on the inferred device type, or disable filter by horizontal scrolling.

• deltaXEvents: Number of events that contains non-zero delta X.
• deltaXAndYEvents: Number of events that contain both non-zero delta X and non-zero delta Y at the same time.
• deltaYLooksLikeTick: Number of events where the deltaY looks like a wheel tick. This is determined by checking deltaY against known constants per-platform that translates a traditional mouse wheel tick into pixel-deltas.
• deltaYFractional: Number of events where the deltaY contains non-integer values.
• deltaModeNotPixels: Number of events where the deltaMode is not pixels.
• numEvents: Total number of accumulated events.
• peakEventsPerSec: Peak number of events received within a 1-second window.

The default classification heuristic used by inferDeviceType() is as follows:

1. Any events with deltaMode != PIXEL → 'mouse'
2. Any events with fractional deltaY → 'trackpad'
3. Any events with both non-zero deltaX and deltaY → 'trackpad'
4. Any events with non-zero deltaX → 'trackpad'
5. Any events with deltaY that doesn't look like a tick → 'trackpad'
6. Otherwise → 'mouse'