support batched axpy of 1D

[yasahi-hpc]

This PR aims at improving the implementation and testing of serial/team/teamvector axpy.

• Moving implementation details into Impl namespace.
• If alpha is a View, we assume X and Y are 2D Views including the batch dimension. If alpha is a scalar, we assume X and Y are 1D Views without batch dimension. This aligns with other batched kernels.

Kokkos::parallel_for(
        "axpy", policy, KOKKOS_LAMBDA(int ib) {
          auto alpha = m_alpha(k);
          auto sub_x = Kokkos::subview(m_X, k, Kokkos::ALL);
          auto sub_y = Kokkos::subview(m_Y, k, Kokkos::ALL);
          // y = y + alpha * x
          KokkosBatched::SerialAxpy::invoke(alpha, sub_x, sub_y);
        });

• Adding unit-tests for the scalar cases

[yasahi-hpc]

Should I need to update docs in this PR?
I am planning to first add an example and add docs.
Would that be fine with you?