Camera orientation mismatch in scaling robot size

[MhdMohammadi]

I'm trying to import "fetch" into one of the scenes, and because the robot's size is relatively too big, I want to rescale it to avoid possible collisions. However, it looks like the camera orientation cannot be set correctly.

File "/omnigibson-src/omnigibson/prims/xform_prim.py", line 141, in _set_xform_properties
    assert np.allclose(new_position, current_position, atol=1e-4) and np.allclose(r1, r2, atol=1e-4), \
AssertionError: /World/robot0/eyes/Camera: old_pos: [1.67729348e-01 5.24871461e-17 1.04592872e+00], new_pos: [1.67729348e-01 5.24871477e-17 1.04592872e+00], old_orn: [ 0.43045944 -0.43045944 -0.56098545  0.56098545], new_orn: [-0.47755187  0.20257453  0.62235748 -0.58615608]

And this part of the code is asserting the mismatch between the new and old orientation values.

self.set_position_orientation(position=current_position, orientation=current_orientation)
new_position, new_orientation = self.get_position_orientation()
r1 = R.from_quat(current_orientation).as_matrix()
r2 = R.from_quat(new_orientation).as_matrix()
# Make sure setting is done correctly
assert np.allclose(new_position, current_position, atol=1e-4) and np.allclose(r1, r2, atol=1e-4), \
    f"{self.prim_path}: old_pos: {current_position}, new_pos: {new_position}, " \
    f"old_orn: {current_orientation}, new_orn: {new_orientation}"

And this is the config file to reproduce the settings.

env:
    action_frequency: 60                  # (int): environment executes action at the action_frequency rate
    physics_frequency: 60                 # (int): physics frequency (1 / physics_timestep for physx)
    # device: null                          # (None or str): specifies the device to be used if running on the gpu with torch backend
    automatic_reset: false                # (bool): whether to automatic reset after an episode finishes
    flatten_action_space: false           # (bool): whether to flatten the action space as a sinle 1D-array
    flatten_obs_space: false              # (bool): whether the observation space should be flattened when generated
    use_external_obs: false               # (bool): Whether to use external observations or not
    initial_pos_z_offset: 0.1
    external_sensors: null

scene:
    type: InteractiveTraversableScene
    scene_model: Benevolence_1_int
    trav_map_resolution: 0.1
    default_erosion_radius: 0.0
    trav_map_with_objects: true
    num_waypoints: 1
    waypoint_resolution: 0.2
    load_room_types: null
    load_room_instances: null
    load_task_relevant_only: false
    seg_map_resolution: 0.1
    scene_source: OG
    include_robots: true

robots:
    - type: Fetch
      obs_modalities: [proprioception, scan, rgb, depth_linear, seg_semantic, seg_instance]
      scale: 0.5
      action_type: continuous
      action_normalize: True
      visual_only: true

task:
    type: DummyTask