// RUN: tf-quant-opt %s -split-input-file -quant-merge-duplicate-resource-ops | FileCheck %s

func.func @merge_duplicate_variable(%arg0: tensor<1x20xf32>, %arg1: tensor<!tf_type.string>) -> (tensor<20x4096xf32>) {
  %0 = tf_executor.graph {
    %outputs_5, %control_6 = tf_executor.island wraps "tf.Const"() {value = dense<""> : tensor<1x!tf_type.string>} : () -> tensor<1x!tf_type.string>

  }

 private:
  OpBuilder op_builder_;
  EmptyPatternRewriter empty_rewriter_;
  Tf2XlaRewriter tf2xla_rewriter_;
};

// This should only have unit tests. End to end tests should be done with
// FILECHECK and MLIR tests.
class Tf2XlaRewriterTest : public ::testing::Test {
 public:
  void SetUp() override {
    tensorflow::XlaOpRegistry::RegisterCompilationKernels();
  }

// RUN: tf-opt %s -tfl-raise-custom-ops="test-raise-tf-targets=tf.FakeQuantWithMinMaxVarsPerChannel,tf.FakeQuantWithMinMaxVars" -tfl-prepare-tf=use-fake-quant-num-bits=true | FileCheck --check-prefix LOBIT --dump-input=always %s

// RUN: tf-tfrt-opt -split-input-file -tfrt-optimize-tf-control-flow-side-effect %s | FileCheck %s

func.func @no_side_effect_cond(%arg: tensor<i32>) -> tensor<i1> {
  %0 = "tf.Const"() {value = dense<16> : tensor<i32> } : () -> tensor<i32>
  %1 = "tf.Less"(%arg, %0) : (tensor<i32>, tensor<i32>) -> tensor<i1>
  func.return %1 : tensor<i1>
}

func.func @no_side_effect_body(%arg: tensor<i32>) -> tensor<i32> {

// RUN: tf-quant-opt %s -split-input-file -quant-lift-quantizable-spots-as-functions -quant-quantize='target-opset=XLA' -verify-each=false | FileCheck %s

func.func private @conv(%input: tensor<1x3x4x3xf32> {tf._user_specified_name = "input_tensor"}) -> tensor<*xf32> attributes {tf._construction_context = "kEagerRuntime", tf._input_shapes = [#tf_type.shape<1x3x4x3>]} {
  %weight = arith.constant dense_resource<__elided__> : tensor<2x3x3x2xf32>

// RUN: tf-quant-opt %s -split-input-file -verify-diagnostics \
// RUN:   -quant-duplicate-shape-determining-constants | FileCheck %s

// CHECK-LABEL: @duplicate_const_for_shape_determining_operand_at_idx_1
// CHECK-SAME: (%[[ARG_0:.*]]: tensor<?x2xi32>)
func.func private @duplicate_const_for_shape_determining_operand_at_idx_1(%arg0: tensor<?x2xi32>) -> tensor<?x2x1xi32> {
  %cst = "tf.Const"() {device = "", value = dense<2> : tensor<i32>} : () -> tensor<i32>

// RUN: flatbuffer_translate -mlir-to-tflite-flatbuffer %s -o - | flatbuffer_translate --tflite-flatbuffer-to-mlir - -o - | FileCheck %s
// Ensure constants roundtrip exactly

func.func @bool() -> tensor<4xi1> {
  // CHECK-LABEL: @bool
  // CHECK: value = dense<[false, true, true, false]> : tensor<4xi1>
  %0 = "tfl.pseudo_const"() { value = dense<[false, true, true, false]> : tensor<4xi1> } : () -> tensor<4xi1>
  func.return %0 : tensor<4xi1>
}

// RUN: stablehlo-quant-opt -optimize-int-graph -split-input-file %s -verify-diagnostics | FileCheck %s

// CHECK-LABEL: func @convolution_add_add
func.func @convolution_add_add(
    %lhs: tensor<?x3x2x1xi8>, %rhs: tensor<2x1x1x1xi8>,
    %zp_offset: tensor<?x2x2x1xi32>, %bias: tensor<1xi32>
  ) -> tensor<?x2x2x1xi32> {
  // CHECK-DAG: %[[conv:.*]] = mhlo.convolution
  // CHECK-DAG: %[[combined:.*]] = chlo.broadcast_add %[[zp_offset:.*]], %[[bias:.*]]

// RUN: tf-opt -tf-saved-model-optimize-global-tensors -split-input-file %s | FileCheck %s
//===----------------------------------------------------------------------===//
// Immutability.
//===----------------------------------------------------------------------===//

// CHECK-LABEL: module attributes {tf_saved_model.semantics}
module attributes {tf_saved_model.semantics} {

  // Test case: This test exercises marking a global tensor as immutable after it propagates

// RUN: flatbuffer_translate -mlir-to-tflite-flatbuffer %s -o - | flatbuffer_to_string - | FileCheck %s