#include "tensorflow/c/eager/c_api_unified_experimental.h"
#include "tensorflow/c/eager/unified_api_testutil.h"
#include "tensorflow/c/experimental/ops/math_ops.h"
#include "tensorflow/c/tf_status_helper.h"
#include "tensorflow/c/tf_tensor.h"
#include "tensorflow/core/platform/tensor_float_32_utils.h"
#include "tensorflow/core/platform/test.h"

namespace tensorflow {
namespace gradients {
namespace internal {
namespace {

}

namespace tensorflow {
Status TF_TensorToTensor(const TF_Tensor* src, Tensor* dst);

// Helpers for loadding a TensorFlow PluggableDevice plugin (a .so file).
Status LoadPluggableDeviceLibrary(const char* library_filename, void** result);
}  // namespace tensorflow

void TFE_InferShapes(TFE_Op* tfe_op, TF_ShapeAndTypeList* input_shapes,
                     TF_Tensor** input_tensors,

        /*in_flight_nodes_limit=*/0);
    TFE_ContextSetExecutorForThread(ctx, executor);
    CHECK_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);

    TF_Tensor* t =
        TF_AllocateTensor(TF_INT32, nullptr, 0, 1 * sizeof(tensorflow::int32));
    *reinterpret_cast<tensorflow::int32*>(TF_TensorData(t)) = 42;
    TFE_TensorHandle* h = TFE_NewTensorHandle(t, status);

// empty, this call is blocked.
//
// Tensors are enqueued via the corresponding TF enqueue op.
// TODO(hongm): Add support for `timeout_ms`.
TF_CAPI_EXPORT extern TF_Tensor* TF_DequeueNamedTensor(TF_Session* session,
                                                       int tensor_id,
                                                       TF_Status* status);

class TFTensorType
    : public mlir::Type::TypeBase<TFTensorType, mlir::Type, mlir::TypeStorage> {
 public:
  using Base::Base;
  static constexpr mlir::StringLiteral name = "tensorflow.tf_mlrt.tf_tensor";
};

// The MLIR type represents a tensorflow::Device*
class TFDeviceType
    : public mlir::Type::TypeBase<TFDeviceType, mlir::Type, mlir::TypeStorage> {
 public:
  using Base::Base;

    CPred<"$_self.isa<mlir::TFR::" # name # "Type>()">,
    "TFR " # name #" type">,
    BuildableType<"$_builder.getType<mlir::TFR::" # name # "Type>()">;
def TFR_TensorType : TFR_Type<"TFRTensor">;
def TFR_TensorListType : TFR_Type<"TFRTensorList">;
def TFR_AllTensorTypes : Type<Or<[
    TFR_TensorType.predicate,
    TFR_TensorListType.predicate]>, "all tensor related types">;

// attribute argument types

  TF_DeleteStatus(s);
  return NULL;
}

// A compute function. This will never actually get called in this test, it's
// just nice to know that it compiles.
void compute(void* kernel, TF_OpKernelContext* ctx) {
  TF_Tensor* input;
  TF_Status* s = TF_NewStatus();
  TF_GetInput(ctx, 0, &input, s);
  TF_DeleteTensor(input);
  TF_DeleteStatus(s);
}

// Exercises tensorflow's C API.
int main(int argc, char** argv) {

  );

  TF_DerivedOperandTypeListAttr Tin = TF_DerivedOperandTypeListAttr<0>;
}

def TF_DumpTensorOp : TF_Op<"DumpTensor", []> {
  let summary = "Dump tensor proto.";

  let arguments = (ins
    Arg<TF_Tensor>:$input,

    StrAttr:$log_dir_path,
    StrAttr:$file_name,
    BoolAttr:$enabled,
    StrAttr:$func_name,
    StrAttr:$node_name
  );

  TF_DerivedOperandTypeAttr T = TF_DerivedOperandTypeAttr<0>;

                      absl::MakeSpan(outputs),
                      /*use_function=*/!std::get<2>(GetParam()));
  ASSERT_EQ(errors::OK, s.code()) << s.message();

  TF_Tensor* result_tensor;
  s = GetValue(outputs[0], &result_tensor);
  ASSERT_EQ(errors::OK, s.code()) << s.message();
  auto result_value = static_cast<float*>(TF_TensorData(result_tensor));
  EXPECT_EQ(*result_value, 1.0);

  ~CApiFunctionTest() override {
    TF_DeleteFunction(func_);
    TF_DeleteGraph(host_graph_);
    TF_DeleteGraph(func_graph_);
    TF_DeleteStatus(s_);
  }

  void Run(const std::vector<std::pair<TF_Operation*, TF_Tensor*>>& inputs,
           TF_Operation* output, int32_t expected_result) {
    Run(inputs, {{output, 0}}, {expected_result});
  }

  // Run the host graph, which now contains a function and check that