Front-end renaming of schedule to linearization

Replaces Rename schedule to outline merge request.

Rename+deprecate get_one_scheduled_kernel() -> get_one_linearized_kernel(), but don't change most calls to get_one_scheduled_kernel() until later (to avoid merge conflicts with other branches).

Rename+deprecate KernelState.SCHEDULE -> KernelState.LINEARIZATION, but don't change accesses to KernelState.SCHEDULE until later (to avoid merge conflicts with other branches).

Allow linearization argument to be passed to LoopKernel.__init__(), but store it in LoopKernel.schedule for now since we're not changing internal representation yet.