}
}
versions {
  producer: 29
  min_consumer: 12
}

# Verify that functions from the library are properly imported.

# CHECK-LABEL:  func @main() {
# CHECK:    "tf.Placeholder"

# CHECK: tf.Const
# CHECK-SAME: value = dense<""> : tensor<3x!tf_type.string>

node {
  name: "save/SaveV2/shape_and_slices"
  op: "Const"
  attr {
    key: "_output_shapes"
    value {
      list {
        shape {
          dim {
            size: 3
          }
        }
      }
    }
  }
  attr {
    key: "dtype"
    value {
      type: DT_STRING

    %arg1: tensor<1x3xf32> {tf_saved_model.index_path = [0]})
      -> (tensor<3x3xf32> {tf_saved_model.index_path = []}) {
  // CHECK: {{%.*}} = corert.get_op_handler %arg0 "/device:GPU:0"
  %2 = "tf.MatMul"(%arg0, %arg1) {T = f32, _output_shapes = ["tfshape$dim { size: 3 } dim { size: 3 }"], device = "/device:GPU:0", transpose_a = false, transpose_b = false} : (tensor<3x1xf32>, tensor<1x3xf32>) -> tensor<3x3xf32>
  func.return %2 : tensor<3x3xf32>

    key: "index"
    value {
      i: 0
    }
  }
}
node {
  name: "args_1"
  op: "_Arg"
  attr {
    key: "T"
    value {
      type: DT_INT32
    }
  }
  attr {
    key: "_output_shapes"
    value {
      list {
        shape {
          dim {
            size: 2
          }
          dim {
            size: 4
          }
          dim {
            size: 6
          }

    tf_executor.fetch
  }
  func.return
}

// CHECK:       name: "RecvTPUEmbedding"
// CHECK-NEXT:  op: "RecvTPUEmbeddingActivations"
// CHECK-NEXT:  attr {
// CHECK-NEXT:    key: "_output_shapes"
// CHECK-NEXT:    value {
// CHECK-NEXT:      list {
// CHECK-NEXT:        shape {
// CHECK-NEXT:          dim {
// CHECK-NEXT:            size: 512
// CHECK-NEXT:          }
// CHECK-NEXT:          dim {

    tf_executor.fetch %arg0 : tensor<*xi32>
  }
  func.return %0 : tensor<*xi32>
}

// CHECK: node {
// CHECK:  name: "tf.LegacyCall"
// CHECK-NEXT:  op: "foo0"
// CHECK:  attr {
// CHECK-NEXT:  key: "_output_shapes"
// CHECK-NEXT:     value {
// CHECK-NEXT:       list {
// CHECK-NEXT:         shape {
// CHECK:  attr {
// CHECK-NEXT:  key: "_tpu_replicate"
// CHECK-NEXT:    value {
// CHECK-NEXT:      s: "cluster"

// general.
struct MLIRImportOptions {
  // If true, functionalize the input graph before importing it into MLIR.
  bool upgrade_legacy = false;

  // Whether to unconditionally use the shape set via _output_shapes on import.
  bool unconditionally_use_set_output_shapes = false;

  // Apply default attributes from the op definition to the loaded op.
  bool add_default_attributes = true;

// `ignore_unregistered_attrs` argument is set to true, the attributes which are
// not in the op registry will be ignored. If the `ignore_unregistered_attrs`
// argument is not set to true, _output_shapes attribute is added to nodes with
// ShapedType for the leading values with ShapedType in the results of the
// nodes. Set it to true if the returned NodeDef will be executed by the linked
// TF Eager runtime.

      end++;
    if (begin != end) {
      mlir::TF::ResultShapeRange output_shapes = {
          mlir::TF::ResultShapeIterator(begin),
          mlir::TF::ResultShapeIterator(end)};
      SetShapeAttribute("_output_shapes", output_shapes, attributes);
    }
  }

  return absl::OkStatus();
}

  %4 = "tf.Add"(%2, %arg2) : (tensor<?x8x10xf32>, tensor<8x10xf32>) -> tensor<?x8x10xf32>
  %5 = "tf.Add"(%arg1, %arg2) : (tensor<8x10xf32>, tensor<8x10xf32>) -> tensor<8x10xf32>
  %6 = "tf.Const"() {_output_shapes = ["tfshape$"], device = "/device:CPU:0", dtype = f32, value = dense<1.000000e+00> : tensor<f32>} : () -> tensor<f32>