namespace mlir {
namespace TFL {
namespace {
#define GEN_PASS_DEF_LEGALIZEVARIABLESPASS
#include "tensorflow/compiler/mlir/lite/transforms/passes.h.inc"

// Attribute name to identify whether variables should be legalized to TFLite or
// not.
const char kLegalizeTflVariables[] = "tfl._legalize_tfl_variables";

bool HasSupportedElementType(Operation* op) {
  return utils::IsSupportedVariableType(op);
}

  target.addLegalDialect<func::FuncDialect>();
  target.addLegalDialect<tensor::TensorDialect>();
  target.addLegalDialect<shape::ShapeDialect>();
  target.addLegalOp<func::CallOp>();

  // These ops are legalized in LegalizeTFCommunication after this and that pass
  // only operates on MHLO control flow ops.
  target.addLegalOp<TF::_XlaHostComputeMlirOp, TF::XlaSendToHostOp,
                    TF::XlaRecvFromHostOp>();

      index_lt.getLhs() != body.getArgument(1) ||
      index_lt.getRhs() != body.getArgument(3))
    return failure();

  return success();
}

// Returns true if the given reduce op can be legalized to ArgMax/ArgMin ops.
std::optional<bool> IsReduceOpLegal(mhlo::ReduceOp reduce_op) {
  if (succeeded(MatchReduceToArgMinMaxType1(reduce_op, true, true)) ||
      succeeded(MatchReduceToArgMinMaxType1(reduce_op, false, true)) ||

  EXPECT_FALSE(IsDynamicPadderOp(TypeID::get<TF::ConstOp>()));
}

// This test is kind of odd. We go through all the Tensorflow types and check
// whether they are legalized with MLIR, TF2XLA, or both. Ideally the sets are
// disjoint, but until that happens, this tests ensures that the set doesn't
// grow.
TEST_F(LegalizationOpConfigTest, CountLoweringsSet) {
  int mlir_lowering_count = 0;

    }
    return success();
  }

  virtual bool IsValueInitValue(const DenseElementsAttr& attr) const = 0;
};

// Returns true if the given reduce op can be legalized to ArgMax/ArgMin ops.
std::optional<bool> IsReduceOpLegal(mhlo::ReduceOp reduce_op);

}  // namespace odml
}  // namespace mlir

    if (inserted_value) continue;
    // We can only clone the constant op or const->dequantize combo. The latter
    // case is useful for float16 quantization. Since all ops have been
    // legalized to tflite ops, so we only care about ConstOp or QConstOp or
    // mlir constant op.
    Operation* input_op = operand.getDefiningOp();
    if (input_op == nullptr) return failure();

    Attribute attr;

    Value output_lhs =
        bmm_op.getAdjX() ? create_z_x_transpose_op(input_lhs) : input_lhs;

    // The rhs need to be transposed if adj_y == false AND this matmul will be
    // legalized to tfl.fully_connected
    Value output_rhs =
        !bmm_op.getAdjY() ? create_z_x_transpose_op(input_rhs) : input_rhs;

    Type output_type = bmm_op.getResult().getType();

// RUN: tf-opt --tfl-legalize-tf-while %s -o - | FileCheck %s
// RUN: tf-opt --tfl-legalize-tf-while %s -o - --tfl-legalize-tf-while --inline='default-pipeline=''' | FileCheck %s --check-prefix=INLINE
// RUN: tf-opt --tfl-legalize-tf-while %s -o - --tfl-legalize-tf-while --inline | FileCheck %s --check-prefix=CANON

// RUN: tf-opt %s -split-input-file -tfl-legalize-variables-tf --cse | FileCheck %s

// Test for case with no session initialize op.
module attributes {tf_saved_model.semantics} {
  // CHECK-LABEL: serving_default
  func.func @serving_default(%arg0: tensor<1x10xf32> {tf_saved_model.index_path = ["x"]}) ->

// RUN: odml-to-stablehlo-opt %s --stablehlo-legalize-vhlo -split-input-file | FileCheck %s
// RUN: odml-to-stablehlo-opt --stablehlo-legalize-vhlo %s | odml-to-stablehlo-opt --vhlo-legalize-stablehlo > %t.0
// RUN: odml-to-stablehlo-opt %s > %t.1
// RUN: diff %t.0 %t.1

// CHECK-LABEL: op_tfl
func.func @op_tfl(%arg0 : tensor<f32>) -> (tensor<f32>) {
  // CHECK: %0 = tfl.add %arg0, %arg0 {fused_activation_function = "NONE"} : tensor<f32>