Resolve body name patterns (exact or regex) against available names.

Send a command string to a tmux pane.

Returns torch.sqrt(torch.max(0, x)) but with a zero subgradient where x is 0.

Unicode representation of the formatted message. String is decoded with the codeset used by the filesystem of the operating system.

Accumulate consecutive steps above the threshold and return done mask.

Update all robot states to match the current replay frames

! @brief public accept interface @param[in] strict whether to expect the last token to be EOF @return whether the input is a proper JSON

! @brief public accept interface @param[in] strict whether to expect the last token to be EOF @return whether the input is a proper JSON

activate the teleop policy

r"""Convert 3d vector of axis-angle rotation to 3x3 rotation matrix. Args: angle_axis: tensor of 3d vector of axis-angle rotations.

Convert rotations given as axis/angle to rotation matrices. Args: axis_angle: Rotations given as a vector in axis angle form,

Perform batch rigid transformation on a skeletal structure. Args: rot_mats: Local rotation matrices for each joint: (B, J, 3, 3)

Overwrite the entire buffer for ``key`` with ``data``. Useful for writing pre-computed values (e.g. advantages, returns) after a full

Updates the given vertex and/or face normals. :param recalcVertexNormals: A flag to indicate whether or not the vertex normals shoul

Capture frame for video recording if enabled, or render the environment.

Finds contact between two body groups. Any geom in the body is considered to be in contact. Args: mj_model (MujocoModel): Current sim

Close and clean up this environment

Close the camera connection.

Compute root position and rotation from the representation. Returns: torch.Tensor: [..., T, 3] global root position t

Convert DOF order between conventions. Auto-detects input format: - qpos: [..., 7 + num_dof] -> reorder DOF p

Fast batch conversion from motion features to mujoco qpos format. Args: motion_features: [batch, numFrames, motion_dim] Motion fe

Create a sortable HTML table for metrics logging using DataTables. Args: metrics_dict: Dictionary containing metrics data with keys

Create a plotter for real-time visualization with pre-created actors. This method initializes a plotter that can be updated efficien

Computes differences between angles. Args: angles (Tensor): [..., T] the batched sequences of rotation angles in radians. Returns:

@brief return whether pointer points to the root document @sa https://json.nlohmann.me/api/json_pointer/empty/

@brief return whether pointer points to the root document @sa https://json.nlohmann.me/api/json_pointer/empty/

Compute the value estimate for given observations. Normalizes the ``critic_obs`` entry if a normalizer is configured, then forwards t

Extracts body motion features from input. Args: motion (torch.Tensor): Motion features of shape [..., D] Returns:

@brief: extract the root / localPose / boundary features from the original full 353 feature

Convert a full configuration to the reduced configuration space. :param q_full: Configuration in full space :return: Configu

Generate and save a frame to LeRobot dataset if enabled.

Path to the system folder where MakeHuman is installed (it is possible that data is stored in another path). Writing to this folder or mo

Compute log-probability of actions under the current distribution. Args: actions: Action tensor of shape ``(batch, act_dim)``.

Get the configuration of body actuated joints from a full configuration. :param q: Configuration in full space :return: Conf

Get the full configuration from the body and hand actuated joint configurations. Can specify either both hands together or left and r

Get poses (position + orientation) of G1 key frames using Pinocchio FK. This is a **display-independent** function — it only needs a Pinocch

Returns the trigger/grip value by name (float). Valid names: "left_trigger", "right_trigger", "left_grip", "right_grip".

Get network interfaces with their IP addresses

Get network interfaces with their IP addresses

This function returns a dictionary of random textures for cabinet, counter top, floor, and wall. Args: rng (np.random.Generator): Ra

Get the current state of the Joy-Con by reading all available events.

Initialize the observation buffer for rollout phase.

Converts motion features back to joint rotations and positions. Args: features (torch.Tensor): Motion features of shape [..., D]

@brief return whether value is a boolean @sa https://json.nlohmann.me/api/basic_json/is_boolean/

@brief return whether value is a boolean @sa https://json.nlohmann.me/api/basic_json/is_boolean/

Load and merge wbc yaml with dataclass overrides

Logs timing information based on thresholds and averages.

Plot tensor data. Args: data: Dictionary mapping keys to tensor values timestamp: Timestamp for the data (if None, us

* @brief Publish (or re-publish) the robot configuration. * * Called during init and periodically so that late-joining subscribers * al

Construct 3D identity rotation given shape

Normalize quaternion to unit length.

Convert quaternion vector to angle axis of rotation. The quaternion should be in (x, y, z, w) or (w, x, y, z) format. Adapted from ceres C++

Convert rotations given as quaternions to rotation matrices. Args: quaternions: quaternions with real part first, as ten

Queue an action to be executed Args: action: A dictionary of action parameters