semanticObjectsFor function

List<SemanticObject> semanticObjectsFor(List<SceneAnnouncement> announcements, { required CameraNode camera, required Size size, })

announcements projected through camera for a viewport of size.

Anything with no bounds, or entirely behind the eye, is dropped: a node a screen reader is told about but cannot be given a position for is a node whose focus ring has nowhere to go.

Implementation

List<SemanticObject> semanticObjectsFor(
  List<SceneAnnouncement> announcements, {
  required CameraNode camera,
  required Size size,
}) {
  if (size.isEmpty) return const <SemanticObject>[];
  final viewProjection = camera.viewProjection(size.width / size.height);
  return <SemanticObject>[
    for (final it in announcements)
      if (it.node.subtreeBounds case final Aabb3 box)
        if (screenBoundsOfBox(
              viewProjection,
              box,
              width: size.width,
              height: size.height,
            )
            case final ScreenBounds at)
          SemanticObject(
            id: it.id,
            label: it.label,
            hint: it.hint,
            selected: it.selected,
            onTap: it.onTap,
            bounds: Rect.fromLTRB(at.left, at.top, at.right, at.bottom),
          ),
  ];
}