See the License for the specific language governing permissions and
limitations under the License.
==============================================================================*/
#include "tensorflow/c/experimental/gradients/tape/tape_operation.h"

#include "tensorflow/c/eager/abstract_context.h"
#include "tensorflow/c/eager/gradients.h"

namespace tensorflow {
namespace gradients {
TapeOperation::TapeOperation(AbstractOperation* parent_op, Tape* tape,

limitations under the License.
==============================================================================*/
#ifndef TENSORFLOW_C_EAGER_ABSTRACT_OPERATION_H_
#define TENSORFLOW_C_EAGER_ABSTRACT_OPERATION_H_

#include <memory>

#include "absl/types/span.h"
#include "tensorflow/c/eager/abstract_tensor_handle.h"
#include "tensorflow/c/tensor_interface.h"

    TF_DeleteOutputList(mm_outputs);
  }

  // Finalize the function by providing the returned values.
  TF_AbstractFunction* func;
  {
    // We want to return the output of both add operations and MatMul operation,
    // create a new list and populate it.
    TF_OutputList* func_outputs = TF_NewOutputList();
    TF_OutputListPushBack(func_outputs, add_output1, s);
    ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);

  TF_Graph* graph = TF_NewGraph();

  // Make a placeholder operation.
  TF_Operation* feed = Placeholder(graph, s);
  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);

  // Make a constant operation with the scalar "2".
  TF_Operation* two = ScalarConst(2, graph, s);
  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);

  // Add operation.
  TF_Operation* add = Add(feed, two, graph, s);

    hdrs = [
        "array_ops.h",
    ],
    visibility = [
        "//tensorflow:internal",
    ],
    deps = [
        "//tensorflow/c/eager:abstract_context",
        "//tensorflow/c/eager:abstract_operation",
        "//tensorflow/c/eager:abstract_tensor_handle",
        "//tensorflow/c/eager:c_api_unified_internal",
        "//tensorflow/c/eager:tracing_utils",
        "//tensorflow/core:framework",

  tensorflow::NodeBuilder node_builder;
  TF_Graph* graph;
  std::set<tensorflow::string> colocation_constraints;
};

struct TF_Operation {
  tensorflow::Node node;

 private:
  ~TF_Operation() = default;
};

struct TF_Session {
  TF_Session(tensorflow::Session* s, TF_Graph* g);

  tensorflow::Session* session;
  TF_Graph* const graph;

    TF_Status* status);

// Controls how to act when we try to run an operation on a given device but
// some input tensors are not on that device.
// LINT.IfChange
// Note: Keep in sync with internal copy of enum in eager/context.h.
typedef enum TFE_ContextDevicePlacementPolicy {
  // Running operations with input tensors on the wrong device will fail.
  TFE_DEVICE_PLACEMENT_EXPLICIT = 0,

  // case placement information is lost when resolving the handle.
  //
  // If false, a caller might implement pretty-printing by resolving and
  // iterating over the resulting tensor. This may still be viable if resolving
  // the handle loses information, but `SummarizeValue` would be more precise.
  virtual bool PreferCustomSummarizer() const { return false; }

  TF_Status* status = TF_NewStatus();
  TF_Operation* arg = TF_FinishOperation(arg_descr, status);
  ASSERT_TRUE(TF_GetCode(status) == TF_OK) << TF_Message(status);
  TF_OperationDescription* id_descr =
      TF_NewOperation(function_graph, "Identity", "id");
  TF_SetAttrType(id_descr, "T", TF_INT32);
  TF_AddInput(id_descr, {arg, 0});
  TF_Operation* id = TF_FinishOperation(id_descr, status);

  TFE_DeleteContext(ctx);
}

tensorflow::AttrValueMap ExtractAttrs(TFE_Op* op) {
  tensorflow::AttrValueMap attr_values;
  tensorflow::EagerOperation* operation =
      tensorflow::OperationFromInterface(tensorflow::unwrap(op));
  operation->Attrs().FillAttrValueMap(&attr_values);
  return attr_values;
}

TEST(CAPI, TestTFE_OpInferSingleInputAttrs) {
  TF_Status* status = TF_NewStatus();