Document / Configure 3D Robot-to-Camera Transform for Limelights

The Limelights need to know their exact 3D position (forward/backward, left/right, up/down, and roll/pitch/yaw) relative to the physical center of the robot in order for Odometry and MegaTag localization (botpose) to calculate correctly. Currently, this transform is expected to be entered manually into each Limelight's Web Interface.

Tasks:
- [x] Measure the physical 3D offset of both the `Drive` and `Shooter` limelights relative to the robot's center.
- [x] Enter these 3D offsets into constants in RobotContainer for both cameras.
- [x] Investigate if we should push these coordinates from Java code (e.g. `RobotMap`) on initialization to prevent data loss if a Limelight is wiped.

Start by reviewing the completed camera-offset constants in RobotContainer and the RobotMap initialization path. Confirm how Drive and Shooter Limelight transforms are currently configured, then determine whether initialization should write them to the cameras. Done means the offsets are preserved after a Limelight reset and the intended configuration is documented.

Written by the indexing model from the issue text.