Unity plugin using on-screen object position as proxy for player attention. Objects are assigned a "tendency" which, together with attention measurements, enables custom loading of prefabs to adapt narrative environments over time.
There is a hack fix implemented where code tries to find render component on same object, then looks up if you have specified one manually, then if those fail gets the first render component it finds on child.
Not ideal, is there a way to composite child object render components together?
Sometimes Time.time.now comparison on enter and exit focus in FocusMeasures component gives wildly larger times than feasible. Seems to happen very rarely. Will test a variety of setups of objects and see if can track down.
A basic version of Nimbus interactions in development. This needs finishing and implementing in a similar way to Focus and FocusMeasures, but might need a separate data handler / etc. Is there a relevant use-case beyond immediate interactions for Nimbus in this plugin's system?
Rather than a discrete / Start-based loading, offer a toggle to enable continuously updated loading of objects and other elements based on tendency. Might want to get the current locale after the specified locale for this, so a locale can evolve its config over time as you attend to objects within it.
Also consider: Timeline and signals system expansion to handle tendency tracking over time to make reactions more dynamic.
Investigate cross-fading of local HDRP volumes between scenes/locales (and tendencies?) and/or combining values of volume profiles? single profile with tendency values applied/combined and applied to different parts of it?
There is a hack fix implemented where code tries to find collider component on same object, then looks up if you have specified one manually, then if those fail gets the first collider component it finds on child. Similar issue to Issue #3.
Is there a way to composite child object collider components together to get an overall collider shape/size?
Getting an index out of bounds error sometimes when loading objects proportionally with prefab loader. Suspect an error in the list generation in proportional / inverse proportion algos.
Write object states out to JSON or binary blob or other serialization to save memory if required, may also help with reloading a scene with previous states.
No errors when there is a mismatch between the settings file on the attention data manager and focus objects, loaders, etc. Need to either find a way to set the config file at the attention data manager script / singleton level, or do error handling so it doesn't fail silently but gives proper feedback to diagnose the issue.