auto cluster = builder.create<mlir::tf_device::ClusterOp>(
      mlir::UnknownLoc::get(&context), result_types);
  cluster->setAttr(kNumCoresPerReplicaAttr,
                   builder.getIntegerAttr(builder.getIntegerType(64), 1));
  cluster->setAttr(kTopologyAttr, builder.getStringAttr(""));
  cluster->setAttr(kDeviceAssignmentAttr, builder.getArrayAttr({}));

  EXPECT_FALSE(HasModelParallelism(cluster));
}

  added_func_op->setAttr(kInterfaceNameAttr, interface_name);
  added_call_op->setAttr(kInterfaceNameAttr, interface_name);

  StringAttr device = mlir::cast<StringAttr>(
      added_func_op->getRegion(0).getBlocks().front().front().getAttr(kDevice));
  StringAttr inference_type = mlir::cast<StringAttr>(
      added_func_op->getRegion(0).getBlocks().front().front().getAttr(
          kInferenceType));

  mlir::OwningOpRef<mlir::ModuleOp> module_ref =
      mlir::ModuleOp::create(mlir::UnknownLoc::get(&context));

  AddDevicesToOp(*module_ref, /*device_set=*/nullptr);
  EXPECT_EQ((*module_ref)->getAttr("tf.devices"), nullptr);
}

TEST(DeviceUtilTest, GetDevicesFromOpNoDevicesAttribute) {
  mlir::MLIRContext context;
  mlir::OwningOpRef<mlir::ModuleOp> module_ref =

    if (op->getDialect() != tf_dialect) return WalkResult::advance();

    if (parallel_group_attr) {
      op->setAttr(TF::kParallelExecAnnotation, parallel_group_attr);
    }
    auto device_attr = op->getAttr(kDeviceAttr);
    if (!device_attr) {
      op->setAttr(kDeviceAttr, launch.getDeviceAttr());
      return WalkResult::advance();
    }

  // function name can be changed by those passes.
  call_op->setAttr(TF::kStablehloEntryFunctionAttrName,
                   FlatSymbolRefAttr::get(builder.getStringAttr(func_name)));
  // Store the custom attribute to restore the function name when loading it
  // back in the post calibration stage. As mentioned above, the above entry
  // function attribute is not reliable.
  call_op->setAttr(kOriginalStablehloEntryFunctionAttrName,

                                          dq.getArg());
      dq.getResult().replaceAllUsesWith(dcast);
      if (auto extra_attr = op->getAttr(mlir::quant::kVolatileOpAttrName)) {
        dcast->setAttr(mlir::quant::kVolatileOpAttrName, extra_attr);
      }
      dq.erase();
    } else if (auto q = llvm::dyn_cast<quantfork::QuantizeCastOp>(op)) {
      auto out_type = q.getResult().getType();

        return func.emitError()
               << "failed to rename StableHLO function " << func.getSymName();
      }
    }
    if (is_main_func) {
      main_func_name = func.getSymNameAttr();
    }
    func->setAttr(kFromXlaCallModuleAttrName, builder.getUnitAttr());
  }
  if (!main_func_name) {
    return stablehlo_module.emitError()
           << "StableHLO module does not have an entry function";
  }
  return main_func_name;

  // Name of the main function for the eventual executable needs to be set.
  main_func->setAttr(kTfSavedModelExportedNamesAttr,
                     builder.getStrArrayAttr({func_name}));
  main_func->setAttr(
      kTfSavedModelInitializerTypeAttr,
      builder.getStringAttr(kTfSavedModelInitializerRestoreType));
  module->setAttr("tf_saved_model.semantics",
                  UnitAttr::get(builder.getContext()));

      return;
    }

    const auto [min_value, max_value] = *min_max_values;
    mlir::OpBuilder builder(aggregator_op);
    aggregator_op->setAttr("min", builder.getF32FloatAttr(min_value));
    aggregator_op->setAttr("max", builder.getF32FloatAttr(max_value));
  });
  return status;
}

bool IsCalibrationRequired(mlir::ModuleOp module_op) {
  bool calibration_required = false;

      auto new_key = StringAttr::get(ctx, old_key.str() + "_dtoh_0");
      op->setAttr("key", new_key);
    } else if (isa<TF::XlaRecvFromHostOp>(op)) {
      StringRef old_key = op->getAttrOfType<StringAttr>("key").getValue();
      auto new_key = StringAttr::get(ctx, old_key.str() + "_htod_0");
      op->setAttr("key", new_key);
    }
  });
  return success();
}