Add RayCluster API with Kuberay support

[hemildesai]

Example usage:

   executor = KubeRayExecutor(
        namespace="your-namespace",
        ray_version="2.43.0",
        image="anyscale/ray:2.43.0-py312-cu125",
        head_cpu="4",
        head_memory="12Gi",
        worker_groups=[
            KubeRayWorkerGroup(
                group_name="worker-group",
                replicas=1,
                gpus_per_worker=8,
            )
        ],
        reuse_volumes_in_worker_groups=True,
        spec_kwargs={"schedulerName": "runai-scheduler"},
    )

    cluster = RayCluster(name="test-cluster", executor=executor)
    cluster.stop()
    cluster.start(wait_until_ready=True)
    
    cluster.port_forward(port=8265, target_port=8265, wait=False)
    cluster.port_forward(port=10001, target_port=10001, wait=True)