spatialgeometry.Shape

class Shape(pose: ndarray | SE3 = array([[1., 0., 0., 0.], [0., 1., 0., 0.], [0., 0., 1., 0.], [0., 0., 0., 1.]]), color: list | ndarray | tuple | set | None = None, stype: str | None = None, base: ndarray | SE3 | None = None, **kwargs)

Bases: SceneNode, ABC

Abstract base class for a renderable 3D shape in a scene graph.

It is a SceneNode instance with atributes for its type, shape, and color. The collision attribute is a read-only bool, used for objects that are drawn in the scene but take no part in collision detection (see CollisionShape for that).

Parameters:

• pose – Local reference frame of the shape, defaults to the identity transform.

• color – Color as (r, g, b) or (r, g, b, a) in [0-1] (or [0-255], auto-normalised), or a matplotlib color name. Defaults to a mid-grey (0.3, 0.3, 0.3, 1.0).

• stype – Shape type identifier used by the renderer/wire protocol (e.g. "cuboid", "mesh") – set by each concrete subclass, not normally passed directly by a caller.

• base – Deprecated alias for pose.

bounds(world: bool = False) → ndarray

Min/max extent of this shape’s axis-aligned bounding box along each axis.

Parameters:

world – See corners().

Return type:

ndarray(3,2)

property collision: bool

True if this shape is used for collision checking rather than (or as well as) visual rendering, as set by the collision argument of a CollisionShape subclass’ constructor.

This is a read-only property.

Return type:

bool

property color: tuple[float, float, float, float]

shape.color returns a four length tuple representing (red, green, blue, opacity) where opacity represents transparency. Values returned are in the range [0-1]. See opacity for a convenient way to get/set just this last channel.

copy() → Shape

Copy of Shape object

Returns:

Shallow copy of Shape object

Return type:

Shape

corners(world: bool = False) → ndarray

The 8 corners of this shape’s axis-aligned bounding box.

Parameters:

world – If True, apply this shape’s current pose and return the axis-aligned envelope of the posed shape (its corners will move as the shape is re-posed, and a rotated shape’s envelope is generally larger than its own local box – this is not the shape’s true oriented/rotated corners). If False (default), return the corners in the shape’s own local frame, independent of pose – constant unless the shape’s own parameters (radius, scale, …) change.

Return type:

ndarray(3,8)

extents(world: bool = False) → ndarray

Dimensions (width, depth, height) of this shape’s axis-aligned bounding box.

Parameters:

world – See corners().

Return type:

ndarray(3,)

fk_dict() → dict[str, Any]

fk_dict() outputs shapes pose in dictionary form

Returns:

The shape pose in translation and quternion form

Return type:

dict

property opacity: float

The last channel of color, in [0-1] – 1.0 is fully opaque, 0.0 fully transparent. A convenience for touching just this channel without needing to know or re-specify the current (r, g, b).

Note

“Opacity” here is the same quantity commonly called “alpha” in computer graphics (as in RGBA) – this package uses “opacity” consistently as the public name for it.

This is a read/write property.

Return type:

float

set_alpha(alpha: float | int) → None

Deprecated – use the opacity property instead.

abstractmethod to_dict() → dict[str, Any]

to_dict() returns the shapes information in dictionary form

Returns:

All information about the shape

Return type:

dict

property v: ndarray

Spatial velocity of the shape as a 6-vector: linear velocity v[:3] followed by angular velocity v[3:6]. Used to integrate the shape’s pose between frames, e.g. by Swift.step() when no per-step callback is supplied.

This is a read/write property.

Return type:

ndarray(6)