warp.transform_point#

warp.transform_point(
xform: Transformation[Float],
point: Vector[Float, Literal[3]],
) → Vector[Float, Literal[3]]#
  • Kernel
  • Python
  • Differentiable

Return point transformed by xform, with rotation applied before translation.

xform.q must have unit length; otherwise the result may distort. Use transform_vector() to transform directions.

Parameters:

  • xform – Transformation to apply.

  • point – Point to transform.

Returns:

quat_rotate(xform.q, point) + xform.p, equivalent to using homogeneous coordinate w = 1.

Example

@wp.kernel
def apply(
    xform: wp.transform,
    points: wp.array[wp.vec3],
    out_points: wp.array[wp.vec3],
    out_vectors: wp.array[wp.vec3],
):
    i = wp.tid()
    out_points[i] = wp.transform_point(xform, points[i])
    out_vectors[i] = wp.transform_vector(xform, points[i])

xform = wp.transform(wp.vec3(0.0, 0.0, 5.0), wp.quat_rpy(0.0, 0.0, wp.pi / 2.0))
points = wp.array([wp.vec3(1.0, 2.0, 0.0)], dtype=wp.vec3)
out_points = wp.empty(1, dtype=wp.vec3)
out_vectors = wp.empty(1, dtype=wp.vec3)
wp.launch(apply, dim=1, inputs=[xform, points], outputs=[out_points, out_vectors])
print(np.round(out_points.numpy(), 3))  # rotated and translated
print(np.round(out_vectors.numpy(), 3))  # rotated only

[[-2.  1.  5.]]
[[-2.  1.  0.]]
warp.transform_point(
mat: Matrix[Float, Literal[4], Literal[4]],
point: Vector[Float, Literal[3]],
) → Vector[Float, Literal[3]]
  • Kernel
  • Python
  • Differentiable

Return point transformed by the 4x4 matrix mat, using homogeneous coordinate w = 1.

The fourth component is discarded without a perspective divide. Matrices that use row-vector conventions, such as those from USD, must be transposed. Use transform_vector() to transform directions.

Parameters:

  • mat – Transformation matrix, applied to a column vector.

  • point – Point to transform.

Returns:

The first three components of mat * (point.x, point.y, point.z, 1).

Example

@wp.kernel
def apply(
    mat: wp.mat44,
    points: wp.array[wp.vec3],
    out_points: wp.array[wp.vec3],
    out_vectors: wp.array[wp.vec3],
):
    i = wp.tid()
    out_points[i] = wp.transform_point(mat, points[i])
    out_vectors[i] = wp.transform_vector(mat, points[i])

# scale by 2 along x and translate by 5 along z
mat = wp.mat44(2.0, 0.0, 0.0, 0.0,
               0.0, 1.0, 0.0, 0.0,
               0.0, 0.0, 1.0, 5.0,
               0.0, 0.0, 0.0, 1.0)
points = wp.array([wp.vec3(1.0, 2.0, 3.0)], dtype=wp.vec3)
out_points = wp.empty(1, dtype=wp.vec3)
out_vectors = wp.empty(1, dtype=wp.vec3)
wp.launch(apply, dim=1, inputs=[mat, points], outputs=[out_points, out_vectors])
print(out_points.numpy())  # translation included
print(out_vectors.numpy())  # translation ignored

[[2. 2. 8.]]
[[2. 2. 3.]]