Add API for screenshots

This adds an API for taking screenshots of the current tab manually and when certain actions are trracked.

The ScreenshotStream interface defines an API for capturing screenshots. Screenshots are taken via the MediaDevices API, specifically the getDisplayMedia() function. After starting a stream of frames from the browser, individual shots can be saved at any point, either manually or when triggering actions fire.

The Registry API has been updated to add 2 fields: triggersScreenshot, which defines whether an action should fire a screenshot, and transitionTime, which defines how long it takes for an action to be reflected in the visualization and therefore the delay before screenshotting.

The Trrack object has been updated with a new field, screenshots, an instance of ScreenshotStream. This exposes methods for starting capture (necessary to obtain a stream of frames from the browser), which asks the user for permission to share the current tab. There are also methods to capture a screenshot immediately, capture after a delay, stop the MediaStream that provides frames from the screenshare, and get saved screenshots. The current implementation of ScreenshotStream saves all captured screenshots to an array. The Trrack object does not take screenshots due to actions with triggersScreenshot = true until the ScreenshotStream has been started.

Trrack / trrackjs

Add API for screenshots #70