[REQ] Add various wp.transform syntax operations for loading and storing #710
Description
Description
We want to bring wp.transform syntax ops up to par with wp.vec and wp.quat.
So we want to support the following in kernels:
lhs = t[idx]
lhs = t.p
lhs = t.q
t[idx] = rhs
t.p = rhs
t.q = rhs
lhs = x[i,j][idx]
lhs = x[i,j].p
lhs = x[i,j].q
x[i,j][idx] = rhs # not differentiable
t[idx] += rhs
t[idx] -= rhs
t.p += rhs
t.p -= rhs
x[i,j] += t
x[i,j] -= tAnd in the python context:
# setters
t.p = rhs
t.q = rhs
wp.ones(N, dtype=wp.transform)