[Question] Dimension Mismatch: 'state_body' is 34-dim in code but 35-dim in dataset

[kiossupersure-lgtm]

Description:
I noticed a dimension inconsistency while reproducing the project:
In Code (server_low_level_*.py): The state_body is constructed as 34 dimensions:
ang_vel (3) + rpy (2) + dof_pos (29) = 34
While in Dataset (data.json): The provided dataset contains a 35-dimensional state_body. This matches the logic in server_motion_lib.py, which includes base position/rotation + joint positions.

My questions:

1. Is the 34-dim implementation in the deployment code intentional (e.g., to handle the Sim2Real gap by ignoring root states)?
2. How should I handle the misalignment of data dimensions to produce standard TWIST datasets?

Thanks!