EXPECT_EQ(xla_device_assignment->computation_count(), 2);
  ASSERT_EQ(xla_device_assignment->computation_devices_size(), 2);
  const auto& computation_device_0 =
      xla_device_assignment->computation_devices(0);
  ASSERT_EQ(computation_device_0.replica_device_ids_size(), 4);
  const auto& computation_device_1 =
      xla_device_assignment->computation_devices(1);
  ASSERT_EQ(computation_device_1.replica_device_ids_size(), 4);

  ASSERT_EQ(unclustered_consumer_inputs.size(), 2);
  EXPECT_EQ(unclustered_consumer_inputs[0]->name(),
            "ClusteredProducer1/declustered");
  EXPECT_EQ(unclustered_consumer_inputs[1]->name(), "Input");

  ASSERT_TRUE(GetInputsForNode(*graph, "ClusteredProducer1/declustered",
                               &declustered_producer_1_inputs));
  ASSERT_EQ(declustered_producer_1_inputs.size(), 2);

    @Override
    void resetExpectations() {
        try {
            if (failure != null) {
                throw failure
            }
            for (ServerExpectation e in expectations) {
                e.assertMet()
            }
        } finally {
            failure = null
            expectations.clear()
        }
    }

    @Override
    protected void after() {
        stop()
        resetExpectations()

    TF_EXPECT_OK(status);
    ASSERT_NE(nullptr, kernel.get());

    p.op_kernel = kernel.get();
    OpKernelContext ctx(&p);
    kernel->Compute(&ctx);

    ASSERT_EQ(2, num_inputs);
    ASSERT_EQ(1, num_outputs);
    ASSERT_EQ(123, ctx.mutable_output(0)->scalar<tensorflow::uint8>()());
  }
}

TEST(TestKernel, DeleteKernelBuilderIsOkOnNull) {
  TF_DeleteKernelBuilder(nullptr);
}

  TF_OperationDescription* desc = TF_NewOperation(graph, "Const", name);
  TF_SetAttrTensor(desc, "value", t, s);
  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);
  TF_SetAttrType(desc, "dtype", TF_TensorType(t));
  *op = TF_FinishOperation(desc, s);
  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);
  ASSERT_NE(*op, nullptr);
}

TF_Operation* Const(TF_Tensor* t, TF_Graph* graph, TF_Status* s,

#include "tensorflow/cc/experimental/libtf/impl/scalars.h"

#include "tensorflow/core/platform/test.h"

namespace tf {
namespace libtf {
namespace impl {

TEST(ScalarsTest, TestHeterogeneousAddition) {
  ASSERT_EQ((Int64(1) + Float32(0.375)).get(), 1.375);
}

}  // namespace impl
}  // namespace libtf

      << "], strides=[" << absl::StrJoin(strides, ",") << "])";

  auto* dlm_out =
      static_cast<DLManagedTensor*>(TFE_HandleToDLPack(handle, status));
  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);
  const DLTensor* dltensor_out = &dlm_out->dl_tensor;
  EXPECT_EQ(dltensor_out->device.device_type, dltensor_in->device.device_type);

      TFE_NewOp(context.get(), "Identity", status.get()), TFE_DeleteOp);
  TFE_OpReset(reused_op.get(), "Identity", custom_device_name, status.get());
  ASSERT_TRUE(TF_GetCode(status.get()) == TF_OK) << TF_Message(status.get());
  ASSERT_EQ(tensorflow::string(TFE_OpGetDevice(reused_op.get(), status.get())),
            tensorflow::string(custom_device_name));
  ASSERT_TRUE(TF_GetCode(status.get()) == TF_OK) << TF_Message(status.get());

    std::string first_device =
        TFE_TensorHandleBackingDeviceName(components[0].get(), status.get());
    ASSERT_EQ(underlying_devices[0], first_device);
    std::string second_device =
        TFE_TensorHandleBackingDeviceName(components[1].get(), status.get());
    ASSERT_EQ(underlying_devices[1], second_device);
  }

  // Add a parallel tensor with different values on each device to the variable.
  {

// Given an unmarshalled CNI config JSON map, return the plugin list asserted as a []interface{}
func GetPlugins(cniConfigMap map[string]any) (plugins []any, err error) {
	plugins, ok := cniConfigMap["plugins"].([]any)
	if !ok {
		err = fmt.Errorf("error reading plugin list from CNI config")
		return
	}
	return
}

// Given the raw plugin interface, return the plugin asserted as a map[string]interface{}