.Attr("SomeDataTypeAttr: type");

  static int num_inputs = 0;
  static int num_outputs = 0;

  // A kernel whose Compute function has a side-effect of updating num_inputs
  // and num_outputs. Various functions on TF_OpKernelContext are also
  // exercised.
  auto my_compute_func = [](void* kernel, TF_OpKernelContext* ctx) {
    num_inputs = TF_NumInputs(ctx);
    num_outputs = TF_NumOutputs(ctx);

        absl::AbortedError(
            "The `ids` and `calibration_methods` must have the same size."));

    // Check the number and type of inputs.
    OP_REQUIRES(context, context->num_inputs() == ids_.size() * 3,
                absl::AbortedError("The number of inputs must be  three times "
                                   "the size of the `ids` list."));
    for (int i = 0; i < ids_.size(); ++i) {

  }
  return absl::OkStatus();
}

std::vector<int> GetResourceVariableIndicesFromContext(OpKernelContext* ctx) {
  std::vector<int> out;
  for (int64 i = 0; i < ctx->num_inputs(); i++) {
    if (ctx->input(i).dtype() == DT_RESOURCE) {
      out.push_back(i);
    }
  }
  return out;
}

Status CreateVariableInfoLookup(
    absl::Span<VariableInfo const> variable_args,

  if (num_constant_inputs < 0 || num_resource_inputs < 0 ||
      num_constant_inputs + num_resource_inputs > n->num_inputs()) {
    return errors::InvalidArgument(
        "Invalid number of constant/resource arguments to XLA kernel.");
  }

  int num_non_constant_inputs =
      n->num_inputs() - num_constant_inputs - num_resource_inputs;

  std::vector<const Edge*> input_edges_vector;

    auto input = pos_and_input.value();
    bool is_packed = input.getIsPacked();
    const int num_operands = input->getNumOperands();
    int num_inputs = is_packed ? 1 : num_replicas;
    if (num_operands != num_inputs)
      return input->emitOpError() << "requires " << num_inputs << " operands";
    if (is_packed) {
      packed_inputs.push_back(input->getOperand(0));
      packed_ops.push_back(input);
    } else {

// if they are all fully defined; std::nullopt otherwise.
std::optional<std::vector<PartialTensorShape>> GetInferredInputShapes(
    int num_inputs, Node* send_from_host_node) {
  std::vector<PartialTensorShape> results(num_inputs);
  for (int i = 0; i < num_inputs; i++) {
    const Edge* e;
    if (!send_from_host_node->input_edge(i, &e).ok()) {
      return std::nullopt;
    }

                                 switch_input == n;
        if (is_loop_invariant) {
          shape_inference::InferenceContext* context =
              shape_refiner->GetContext(n);
          for (int i = 0; i < n->num_inputs(); i++) {
            const Node* input_node;
            if (n->input_node(i, &input_node).ok()) {
              auto shapes_and_types = context->input_handle_shapes_and_types(i);
              if (shapes_and_types) {

}

void TF_AddInputList(TF_OperationDescription* desc, const TF_Output* inputs,
                     int num_inputs) {
  std::vector<NodeBuilder::NodeOut> input_list;
  input_list.reserve(num_inputs);
  for (int i = 0; i < num_inputs; ++i) {
    input_list.emplace_back(&inputs[i].oper->node, inputs[i].index);
  }
  desc->node_builder.Input(input_list);
}

    FunctionType func_type = func.getFunctionType();
    int num_inputs = func_type.getNumInputs();
    int num_results = func_type.getNumResults();

    // For each argument type in function's arguments, change it to uranked
    // tensor type if it's a variant type.
    SmallVector<Type, 8> updated_argument_types;
    updated_argument_types.reserve(num_inputs);
    UpdateTensorListTypes<mlir::OperandRange>(

}

void TFE_OpAddInputList(TFE_Op* op, TFE_TensorHandle** inputs, int num_inputs,
                        TF_Status* status) {
  status->status = tensorflow::unwrap(op)->AddInputList(
      {reinterpret_cast<tensorflow::AbstractTensorHandle**>(
           tensorflow::unwrap(inputs)),
       static_cast<size_t>(num_inputs)});
}

extern int TFE_OpGetFlatInputCount(const TFE_Op* op, TF_Status* status) {