warp.transform_multiply

warp.transform_multiply(

a: Transformation[Float],

b: Transformation[Float],

) → Transformation[Float]

• Kernel: true
• Python: true
• Differentiable: true

Return the composition of transformations a and b, applying b before a.

With unit quaternions, the operation is associative but not commutative, and transform_point(a * b, x) == transform_point(a, transform_point(b, x)). Non-unit quaternions may distort and invalidate these identities.

Parameters:

• a – Outer transformation, applied second.

• b – Inner transformation, applied first.

Returns:

The composed transformation, with translation a.p + quat_rotate(a.q, b.p) and rotation a.q * b.q.

Example

@wp.kernel
def compose(
    a: wp.array[wp.transform], b: wp.array[wp.transform], out: wp.array[wp.transform]
):
    i = wp.tid()
    out[i] = wp.transform_multiply(a[i], b[i])

# a turns a quarter turn about the z axis and shifts along x, b only shifts along x
a = wp.array([wp.transform(wp.vec3(1.0, 0.0, 0.0), wp.quat_rpy(0.0, 0.0, wp.pi / 2.0))],
             dtype=wp.transform)
b = wp.array([wp.transform(wp.vec3(2.0, 0.0, 0.0), wp.quat_identity())], dtype=wp.transform)
out = wp.empty(1, dtype=wp.transform)
wp.launch(compose, dim=1, inputs=[a, b], outputs=[out])
print(np.round(out.numpy(), 3))

[[1.    2.    0.    0.    0.    0.707 0.707]]