name: "function"
      input_arg {
        name: "inputs"
        type: DT_FLOAT
      }
      input_arg {
        name: "statefulpartitionedcall_args_1"
        type: DT_RESOURCE
      }
      output_arg {
        name: "identity"
        type: DT_FLOAT
      }
      is_stateful: true
    }
    node_def {
      name: "Identity"
      op: "Identity"
      input: "inputs"
      attr {

// CHECK:        function
// CHECK:          signature
// CHECK:            input_arg
// CHECK:              type: DT_RESOURCE
// CHECK:            input_arg
// CHECK:              type: DT_RESOURCE
// CHECK:            output_arg
// CHECK:              type: DT_RESOURCE
// CHECK:          ret

// Check _resource_arg_unique_id for each argument. Since they alias each other,
// both values are 0.
// CHECK:          arg_attr

    }
  }
}
library {
  function {
    signature {
      name: "__inference_Dataset_map_<lambda>_8"
      input_arg {
        name: "args_0"
        type: DT_INT32
      }
      output_arg {
        name: "identity"
        type: DT_INT32
      }
    }
    node_def {
      name: "mul/y"
      op: "Const"
      attr {
        key: "dtype"
        value {

// CHECK-NEXT:       name: "function0"
// CHECK-NEXT:       input_arg {
// CHECK-NEXT:         name: "function0"
// CHECK:            input_arg {
// CHECK-NEXT:         name: "function01"
// CHECK:            output_arg {
// CHECK-NEXT:         name: "function02"
// CHECK:          node_def {
// CHECK-NEXT:       name: "[[NAME:[^"]*]]"
// CHECK-NEXT:       op: "Identity"
// CHECK-NEXT:       input: "function0"
// CHECK:          ret {

      name: "function"
      input_arg {
        name: "inputs"
        type: DT_FLOAT
      }
      input_arg {
        name: "statefulpartitionedcall_args_1"
        type: DT_RESOURCE
      }
      output_arg {
        name: "identity"
        type: DT_FLOAT
      }
      is_stateful: true
    }
    node_def {
      name: "Identity"
      op: "Identity"
      input: "inputs"
      attr {

      op_name = name.qn[1]
      op_def, _ = self._op_defs.lookup(op_name)
      if len(op_def.output_arg) == 1:
        return {_get_type_from_proto(op_def.output_arg[0])}, None
      return ({tuple(_get_type_from_proto(arg) for arg in op_def.output_arg)},
              None)

    elif f_type == (TFRTypes.PY_BUILTIN_FUNC,):
      assert name.is_simple()
      if name == QN('range'):

op {
  name: "Foo"
  input_arg {
    name: "images"
    description: "Images to process."
  }
  input_arg {
    name: "dim"
    description: "Description for dim."
    type: DT_FLOAT
  }
  output_arg {
    name: "output"
    description: "Description for output."
    type: DT_FLOAT
  }
  attr {
    name: "T"
    type: "type"
    description: "Type for images"
    allowed_values {
      list {

  in.mutable_composite_tensor()->add_components()->set_name("input1:0");
  in.mutable_composite_tensor()->add_components()->set_name("input2:0");

  TensorInfo& out = (*signature_def.mutable_outputs())["output_arg"];
  out.mutable_composite_tensor()->add_components()->set_name("output2:0");
  out.mutable_composite_tensor()->add_components()->set_name("output1:0");

  AddSignatureDefToSavedModelBundle(signature_def, "signature_def",

  EXPECT_EQ(f1_rewritten->signature().input_arg(1).type(), DT_RESOURCE);
  ASSERT_EQ(f1_rewritten->signature().output_arg_size(), 1);
  EXPECT_EQ(f1_rewritten->signature().output_arg(0).type(), DT_BOOL);

  // Check node "if" input and output edges.
  auto node_name_index = g->BuildNodeNameIndex();
  const Node *if_node = node_name_index.at("if");
  ASSERT_NE(if_node, nullptr);

      "    signature {"
      "      name: 'AddVariablesFunction'"
      "      input_arg {"
      "        name: 'var'"
      "        type: DT_RESOURCE"
      "      }"
      "      output_arg {"
      "        name: 'sum'"
      "        type: DT_FLOAT"
      "      }"
      "    }"
      "    node_def {"
      "      name: 'read0'"
      "      op: 'ReadVariableOp'"
      "      input: 'var'"