`cudaq.from_state` overload should be a member on the kernel builder class

Required prerequisites

• Search the issue tracker to check if your feature has already been mentioned or rejected in other issues.

Describe the feature

We have 2 overloads of the cudaq::from_state function. The first, which I believe should remain the same, returns a kernel given a state vector. The second accepts a kernel, a set of qubits, the state, and returns null (cudaq::from_state(Kernel kernel, QuakeValue qubits, ndarray state) -> None). It would make much more sense if this second overload was a member function on the Kernel/kernel_builder type. See:

kernel = cudaq.make_kernel()
qubits = kernel.qalloc()
state = np.array(...)

# It doesn't make sense to have to use a function from the outer cudaq 
# namespace to generate operations within the kernel function body .
cudaq.from_state(kernel, qubits, state)

# Instead:
# The from_state function now takes the kernel as self, and only needs the qubits and state.
kernel.from_state(qubits, state)