func.func @serving_default(%arg0: tensor<!tf_type.resource<tensor<100x50xf32>>> {tf.resource_name = "dense/kernel"}, %arg1: tensor<!tf_type.resource<tensor<50xf32>>> {tf.resource_name = "dense/bias"}) -> (tensor<100x50xf32> {tf_saved_model.index_path = ["dense_2"]})
  attributes {tf.entry_function = {control_outputs = "", inputs = "", outputs = "dense_2/Add:0"}, tf_saved_model.exported_names = ["serving_default"]} {

# CHECK:      func {{@[a-zA-Z_0-9]+}}(
# CHECK-SAME:   [[ARG0:%.*]]: tensor<3x1xf32> {tf_saved_model.index_path = ["x"]},
# CHECK-SAME:   [[ARG1:%.*]]: tensor<!tf_type.resource<tensor<1x3xf32>>> {tf_saved_model.bound_input = @[[VAR]]})
# CHECK-SAME:             -> (tensor<3x3xf32> {tf_saved_model.index_path = ["r"]})
# CHECK-SAME: attributes {{.*}} tf_saved_model.exported_names = ["key"]

# CHECK:      func {{@[a-zA-Z_0-9]+}}(

module attributes {tf_saved_model.semantics} {
  // CHECK-LABEL: func @lowers_to_stablehlo
  func.func @lowers_to_stablehlo(%arg0: tensor<i32> {tf_saved_model.index_path = []}) -> (tensor<i32> {tf_saved_model.index_path = []})
    attributes {tf_saved_model.exported_names = ["lowers_to_stablehlo"]}
  {
    // CHECK-DAG: [[one:%.*]] = stablehlo.constant dense<1>

module attributes {tf_saved_model.semantics} {
  // CHECK-LABEL: serving_default
  func.func @serving_default(%arg0: tensor<1x10xf32> {tf_saved_model.index_path = ["x"]}) ->

  // expected-error@+1 {{Incompatible code paths}}
  func.func @f(%arg0: tensor<?xf32> {tf_saved_model.index_path = [0]}, %v: tensor<!tf_type.resource<tensor<?xf32>>> {tf_saved_model.bound_input = @v}, %v1: tensor<!tf_type.resource<tensor<?xf32>>> {tf_saved_model.bound_input = @v1}) -> (tensor<?xf32> {tf_saved_model.index_path = [0]}) attributes {tf_saved_model.exported_names = ["f"]} {
    %pred = arith.constant false

  func.func @__concrete_function_run_computation(
    %arg0: tensor<f32> {tf_saved_model.index_path = [0, "foo"]},
    %arg1: tensor<!tf_type.resource<tensor<1x64xf32>>> {tf_saved_model.bound_input = @some_constant},
    %arg2: tensor<!tf_type.resource<tensor<?x64xf32>>> {tf_saved_model.bound_input = @some_variable}
  ) -> (
    tensor<f32> {tf_saved_model.index_path = [0, "bar"]}
  ) attributes { tf_saved_model.exported_names = ["some_func"] }
  {

  func.func @f(%arg0: tensor<!tf_type.resource<tensor<f32>>> {tf_saved_model.bound_input = @v}) -> (tensor<f32> {tf_saved_model.index_path = []})
  attributes {tf_saved_model.exported_names = ["f"]} {
    %val = "tf.PartitionedCall"(%arg0) {config = "", config_proto = "", executor_type = "", f = @f_callee} : (tensor<!tf_type.resource<tensor<f32>>>) -> (tensor<f32>)

// CHECK: "tf.LookupTableSizeV2"(%arg1)
// CHECK: "tf.LookupTableFindV2"(%arg1
  func.func @main(%arg0: tensor<?x!tf_type.string> {tf_saved_model.index_path = ["input_vocabs:0"]} ) -> (tensor<*xi64>  {tf_saved_model.index_path = ["FakeQuantWithMinMaxArgs_2:0"]}) attributes {tf.entry_function = {inputs = "input_vocabs:0", outputs = "FakeQuantWithMinMaxArgs_2:0"}, tf_saved_model.exported_names = ["main"]} {

namespace {

using ::mlir::tf_saved_model::kTfSavedModelIndexPathAttr;

llvm::StringRef ProcessIndexPath(mlir::ArrayAttr index_path) {
  if (index_path.size() == 1 && mlir::isa<mlir::StringAttr>(index_path[0])) {
    // TODO(chky): Support cases where index_path is not a single string.
    return mlir::cast<mlir::StringAttr>(index_path[0]).getValue();
  }
  return "";
}

# CHECK:      func {{@[a-zA-Z_0-9]+}}(
# CHECK-SAME:   [[ARG0:%.*]]: tensor<3x1xf32> {tf_saved_model.index_path = ["x"]},
# CHECK-SAME:   [[ARG1:%.*]]: tensor<!tf_type.resource<tensor<1x3xf32>>> {tf_saved_model.bound_input = @[[VAR]]})
# CHECK-SAME:             -> (tensor<3x3xf32> {tf_saved_model.index_path = ["r"]})
# CHECK-SAME: attributes {{.*}} tf_saved_model.exported_names = ["key"]