Utils¶

vista.utils.misc.agent2poly(agent: vista.entities.agents.Car.Car, ref_dynamics: Optional[vista.entities.agents.Dynamics.StateDynamics] = None) → shapely.geometry.geo.box[source]¶

Convert Agent object to polygon w.r.t. a reference dynamics.

Parameters

agent (Car) – An agent with valid dynamics (pose and vehicle state).
ref_dynamics (StateDynamics) – A reference dynamics for computing the polygon representation of the agent. Default set to None that uses human dynamics of the agent.

Returns

A polygon that describes the agent.

Return type

Box

vista.utils.misc.merge_dict(dict1: Dict, dict2: Dict) → Dict[source]¶

Merge two dict, where dict1 has higher priority.

Parameters

dict1 (Dict) – The first dictionary.
dict2 (Dict) – The second dictionary.

Returns

The merged dictionary.

Return type

Dict

vista.utils.misc.fetch_agent_info(agent: vista.entities.agents.Car.Car) → Dict[str, Any][source]¶

Get info from agent class.

Parameters: agent (Car) – The agent to be extracted information from.
Returns: A dictionary contains various information of the agent.
Return type: Dict

vista.utils.misc.img2flow(img: numpy.ndarray, mag_minmax: Union[List, Tuple, numpy.ndarray], flow_size: Optional[Union[List, Tuple, numpy.ndarray]] = None) → numpy.ndarray[source]¶

Convert HSV-encoded flow image to optical flow.

Parameters

img (np.ndarray) – An image with channel order BGR.
mag_minmax (Vec) – The minmum and maximum when normalizing the flow magnitude to [0,1].
flow_size (Vec) – Size of the output flow array. If set, resize the image before converting to flow; default to None.

Returns

A HxWx2 array with the two channels as magnitude and the angle of the flow.

Return type

np.ndarray

vista.utils.misc.biinterp(I0: numpy.ndarray, I1: numpy.ndarray, F_0_1: numpy.ndarray, F_1_0: numpy.ndarray, ts: float, t0: float, t1: float) → numpy.ndarray[source]¶

Interpolate frame with bidirectional flow.

Parameters

I0 (np.ndarray) – A RGB image at time t0.
I1 (np.ndarray) – A RGB image at time t1.
F_0_1 (np.ndarray) – The flow from time t0 to t1.
F_1_0 (np.ndarray) – The flow from time t1 to t0.
ts (float) – The timestamp to be interpolated to.
t0 (float) – The timestamp of I0.
t1 (float) – The timestamp of I1.

Returns

An interpolated RGB image at time ts.

Return type

np.ndarray

vista.utils.misc.flow_backwarp(img: numpy.ndarray, flow: numpy.ndarray, use_pytorch: Optional[bool] = False) → numpy.ndarray[source]¶

Warp image based on optical flow.

Parameters

img (np.ndarray) – An image to be warped.
flow (np.ndarray) – Optical flow to warp the image.
use_pytorch (bool) – Whether to use pytorch for warping; default to False.

Returns

A warped image.

Return type

np.ndarray

Helper functions for spatial transformation. Follow right handed (OpenGL) coordinate system.

vista.utils.transform.rot2mat(rot: Union[numpy.ndarray, List[Any], Tuple[Any]], seq: Optional[str] = 'xyz') → numpy.ndarray[source]¶

Convert euler vector (with sequence order) to rotation matrix.

Parameters

rot (Vec) – A 3-dimensional rotation vector in Euler angle.
seq (str) – The order of the rotation vector, e.g., xzy; default to xyz.

Returns

A rotation matrix equivalent to the given rotation vector.

Return type

np.ndarray

vista.utils.transform.vec2mat(trans: Union[numpy.ndarray, List[Any], Tuple[Any]], rot: Union[numpy.ndarray, List[Any], Tuple[Any]]) → numpy.ndarray[source]¶

Convert translation and rotation vector to transformation matrix.

Parameters

trans (Vec) – A 3-dimensional translation vector.
rot (Vec) – A 3-dimensional rotation vector.

Returns

A 4-by-4 transformation matrix in SE(3).

Return type

np.ndarray

vista.utils.transform.euler2quat(euler: Union[numpy.ndarray, List[Any], Tuple[Any]], seq: Optional[str] = 'xyz', degrees: Optional[bool] = False) → Union[numpy.ndarray, List[Any], Tuple[Any]][source]¶

Convert Euler rotation to quaternion.

Parameters

euler (Vec) – A 3-dimensional rotation vector in Euler angle.
seq (str) – The order of the rotation vector, e.g., xzy; default to xyz.

Returns

A 4-dimensional vector that describes a quaternion.

Return type

Vec

vista.utils.transform.quat2euler(quat: Union[numpy.ndarray, List[Any], Tuple[Any]], seq: Optional[str] = 'xyz', degrees: Optional[bool] = False) → Union[numpy.ndarray, List[Any], Tuple[Any]][source]¶

Convert quaternion to Euler rotation.

Parameters

quat (Vec) – A 4-dimensional vector that describes a quaternion.
seq (str) – The order of the output Euler rotation vector; default to xyz.
degrees (bool) – Whether to convert the output rotation to degrees instead of radians; default to False.

Returns

A 3-dimensional rotation vector in Euler angle.

Return type

Vec

vista.utils.transform.latlongyaw2vec(latlongyaw: Union[numpy.ndarray, List[Any], Tuple[Any]]) → Tuple[Union[numpy.ndarray, List[Any], Tuple[Any]], Union[numpy.ndarray, List[Any], Tuple[Any]]][source]¶

Convert lateral, longitudinal, yaw compoenents to translational and rotational vectors. Note that the longitudinal component will be negated to follow right-handed OpenGL coordinate system.

Parameters: latlongyaw (Vec) – A 3-dimensional vector with entries as lateral shift, longitudinal shift, and yaw difference.
Returns: Return a tuple (Vec_a, Vec_b), where Vec_a is a 3-dimensional translation vector and Vec_b is a 3-dimensional rotation vector in Euler angle xyz.

vista.utils.transform.vec2latlongyaw(trans: Union[numpy.ndarray, List[Any], Tuple[Any]], rot: Union[numpy.ndarray, List[Any], Tuple[Any]]) → Union[numpy.ndarray, List[Any], Tuple[Any]][source]¶

Convert translational and rotational vectors to lateral, longitudinal, yaw compoenents. Note that the y translational component will be negated to follow right-handed OpenGL coordinate system.

Parameters

trans (Vec) – A 3-dimensional translation vector.
rot (Vec) – A 3-dimensional Euler rotation vector.

Returns

A 3-dimensional vector with entries as lateral shift,: longitudinal shift, and yaw difference.

Return type

Vec

vista.utils.transform.compute_relative_latlongyaw(latlongyaw: Union[numpy.ndarray, List[Any], Tuple[Any]], latlongyaw_ref: Union[numpy.ndarray, List[Any], Tuple[Any]]) → Union[numpy.ndarray, List[Any], Tuple[Any]][source]¶

Compute relative lateral, longitudinal, yaw compoenents.

Parameters

latlongyaw (Vec) – A 3-dimensional vector with lateral, longitudinal, and yaw component.
latlongyaw_ref (Vec) – A reference 3-dimensional vector with lateral, longitudinal, and yaw component.

Returns

A 3-dimensional vector that describe relative lateral,: longitudinal shift and yaw difference between the given two vectors.

Return type

Vec

vista.utils.transform.SE3_inv(T_in: numpy.ndarray) → numpy.ndarray[source]¶

More efficient matrix inversion for SE(3).

Parameters: T_in (np.ndarray) – a 4-by-4 transformation matrix in SE(3).
Returns: The inverse of T_in.
Return type: np.ndarray

vista.utils.transform.mat2vec(mat: numpy.ndarray, seq: Optional[str] = 'xyz', degrees: Optional[bool] = False) → Tuple[Union[numpy.ndarray, List[Any], Tuple[Any]], Union[numpy.ndarray, List[Any], Tuple[Any]]][source]¶

Convert transformation matrix to translational and rotational vectors.

Parameters

mat (np.ndarray) – A 4-by-4 transformation matrix in SE(3).
seq (str) – The order of the output Euler rotation vector; default to xyz.
degrees (bool) – Whether to convert the output rotation to degrees instead of radians; default to False.

Returns

Return a tuple (Vec_a, Vec_b), where Vec_a is a 3-dimensional translation vector and Vec_b is a 3-dimensional rotation vector in Euler angle xyz.

vista.utils.transform.pi2pi(angle: float) → float[source]¶

Make sure angle is within -pi to pi.

Parameters: angle (float) – Input angle in radians.
Returns: Output angle that is within range [-pi, pi]
Return type: float

Logging utilities of the simulator.

class vista.utils.logging.CustomFormatter(fmt=None, datefmt=None, style='%', validate=True)[source]¶

A custom logger class inherited from python logging module. It basically define a custom formatting as <time>::<level>::<module>::<msg> with colored text (DEBUG: cyan, INFO: regular, WARNING: yellow, ERROR: bold red, CRITICAL: underlined bold red). Otherwise, it can be used as if using regular python logging.

Example Usage:

>>> from vista.utils import logging
>>> logging.setLevel(logging.ERROR)
>>> logging.debug('This is a debug message.')
>>> logging.warning('This is a warning message.')

format(record)[source]¶

Format the specified record as text.

The record’s attribute dictionary is used as the operand to a string formatting operation which yields the returned string. Before formatting the dictionary, a couple of preparatory steps are carried out. The message attribute of the record is computed using LogRecord.getMessage(). If the formatting string uses the time (as determined by a call to usesTime(), formatTime() is called to format the event time. If there is exception information, it is formatted using formatException() and appended to the message.

vista.utils.logging.setLevel(level: Union[str, int]) → None[source]¶

Set the logging level of this handler. Check CustomFormatter for example usage.

Parameters: level (str or int) – Logging level.