self, dependency: Optional[Callable[..., Any]] = None, *, use_cache: bool = True
    ):
        self.dependency = dependency
        self.use_cache = use_cache

    def __repr__(self) -> str:
        attr = getattr(self.dependency, "__name__", type(self.dependency).__name__)
        cache = "" if self.use_cache else ", use_cache=False"
        return f"{self.__class__.__name__}({attr}{cache})"


class Security(Depends):

  // 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,

  ElementsAttr attr;
  if (!matchPattern(value, m_Constant(&attr))) return {};
  if (attr.getNumElements() != 1) return {};
  IntegerAttr int_attr = *attr.getValues<IntegerAttr>().begin();
  return int_attr.getValue().getSExtValue();
}

// Returns a RankedTensorType which is similar to `input_type` but replaces the
// dimension size of `dim` with `dim_size`.  For example,

                        AttrValue* value) {
  TF_RETURN_IF_ERROR(ConvertAttribute(
      mlir::cast<mlir::FlatSymbolRefAttr>(attr.getName()), value));
  TF_RETURN_IF_ERROR(ConvertAttributes(attr.getAttrs().getValue(),
                                       /*attrs_to_ignore=*/{}, remove_ref_type,
                                       value->mutable_func()->mutable_attr()));
  return absl::OkStatus();
}

        }

        rewriter.setInsertionPoint(quantizing_op);
        Operation* float_op = rewriter.clone(*quantizing_op);
        quantizing_op->setAttr(kDebugModeOpQuantAttrName,
                               rewriter.getUnitAttr());
        float_op->setAttr(kDebugModeOpFloatAttrName, rewriter.getUnitAttr());
        RewireFloatModelBackbone(quantizing_op, float_op);
        return success();
      }

        encoding.getBounds()[i] == ShapedType::kDynamic) {
      return false;
    }
  }
  return true;
}

bool HasSymbolRefAttr(Operation* op) {
  for (const auto& attr : op->getAttrs()) {
    Attribute attr_value = attr.getValue();
    if (mlir::isa<SymbolRefAttr>(attr_value)) {
      return true;
    } else if (auto array_attr = mlir::dyn_cast<ArrayAttr>(attr_value)) {

                                                TF_Status* status) {
  ::tensorflow::TensorShape shape;
  auto* cc_ctx = reinterpret_cast<::tensorflow::OpKernelConstruction*>(ctx);
  ::tensorflow::Status s = cc_ctx->GetAttr(attr_name, &shape);
  ::tensorflow::Set_TF_Status_from_Status(status, s);
  size_t rank = static_cast<size_t>(shape.dims());

  if (!status->status.ok()) return;

  if (num_dims != rank) {

      for (const auto& attr_name : attr_names) {
        if (!global_iter_id_op->hasAttr(attr_name)) continue;
        offset->setAttr(attr_name, global_iter_id_op->getAttr(attr_name));
        new_global_iter_id->setAttr(attr_name,
                                    global_iter_id_op->getAttr(attr_name));
      }
      // Make the next function to get inlined use a different offset.
      ++offset_value;
    }

  }
  xla::ConvolutionDimensionNumbers dnums =
      ConvertConvolutionDimensionNumbers(dnums_input);

  SmallVector<NamedAttribute> converted_attrs;
  for (auto attr : op->getAttrs()) {
    if (attr.getName() == op.getFeatureGroupCountAttrName() ||
        attr.getName() == op.getBatchGroupCountAttrName()) {
      converted_attrs.push_back(attr);

    assert(funcs_for_cluster != tpu_funcs.end());
    assert(!funcs_for_cluster->second.empty());
    if (funcs_for_cluster->second.size() == 1) return false;
    for (NamedAttribute attr : op->getAttrs()) {
      auto symbol_ref = mlir::dyn_cast<FlatSymbolRefAttr>(attr.getValue());
      if (!symbol_ref) continue;
      func::FuncOp callee =
          symbol_table.lookup<func::FuncOp>(symbol_ref.getValue());