[MoE Refactor] Introduce MoERunner abstraction and move execution logic from FusedMoE to DefaultMoERunner (#32344)

Signed-off-by: Bill Nell <bnell@redhat.com>

vllm/model_executor/layers/fused_moe/modular_kernel.py

@@ -1228,7 +1228,7 @@ class FusedMoEModularKernel(torch.nn.Module):
         topk_weights: torch.Tensor,
         topk_ids: torch.Tensor,
         apply_router_weight_on_input: bool,
-        shared_experts_input: torch.Tensor | None = None,
+        shared_experts_input: torch.Tensor | None,
     ) -> torch.Tensor | tuple[torch.Tensor, torch.Tensor]:
         """
         The _finalize method is a wrapper around self.prepare_finalize.finalize