Add MultiPhotonPoseEstimator (from Robot2025)

[kcooney]

This copies MultiPhotonPoseEstimator from the Robot2025 code, and makes
the following improvements:

1. Improve Javadoc
2. Rename methods for clarity
   • update() -> processAllUnreadResults()
   • addToSim() -> addCamerasToSimulator()
   • setDrivePose() - > publishCameraPosesRelativeTo()
3. Introduce PoseEstimateConsumer
4. Simplify the API
   • Remove addCamera(String, Transform3d, String description) overload
     (if the camera names describe where they are located, no descriptions
     are needed)
   • Simplify API of addCamerasToSimulator() by allowing callers to add a
     Supplier<SimCameraProperties> to addCamera()
   • Publish additional topics under photonvision/[cameraName]/LatestPose
     (vs Vision/[cameraName])

The first commit simply copies the code as-is, only changing package
statements.

[kcooney]

@vdikov friendly ping

[cuttestkittensrule]

lgtm.
We could probably handle some of the TODOs, but this should probably just be merged in