//	Atan2(+Inf, -Inf) = 3Pi/4
//	Atan2(-Inf, -Inf) = -3Pi/4
//	Atan2(y, +Inf) = 0
//	Atan2(y>0, -Inf) = +Pi
//	Atan2(y<0, -Inf) = -Pi
//	Atan2(+Inf, x) = +Pi/2
//	Atan2(-Inf, x) = -Pi/2
func Atan2(y, x float64) float64 {
	if haveArchAtan2 {
		return archAtan2(y, x)
	}
	return atan2(y, x)
}

func atan2(y, x float64) float64 {
	// special cases
	switch {
	case IsNaN(y) || IsNaN(x):
		return NaN()

	-Pi,             // atan2(-0, -Pi)
	-Pi,             // atan2(-0, -0)
	Copysign(0, -1), // atan2(-0, +0)
	Copysign(0, -1), // atan2(-0, +Pi)
	Copysign(0, -1), // atan2(-0, +Inf)
	NaN(),           // atan2(-0, NaN)
	Pi,              // atan2(+0, -Inf)
	Pi,              // atan2(+0, -Pi)
	Pi,              // atan2(+0, -0)
	0,               // atan2(+0, +0)
	0,               // atan2(+0, +Pi)
	0,               // atan2(+0, +Inf)

    // CHECK: tf.Atan2
    // expected-remark@+1 {{lowering requires bounded tensor operands}}
    %0 = "tf.Atan2"(%arg0, %arg0) : (tensor<*xf32>, tensor<*xf32>) -> tensor<*xf32>

    func.return %0 : tensor<*xf32>
  }

  // CHECK-LABEL: dynamic_operand
  func.func @dynamic_operand(%arg0: tensor<?xf32>) -> tensor<?xf32> {
    // CHECK: tf.Atan2

func ExampleAsinh() {
	fmt.Printf("%.2f", math.Asinh(0))
	// Output: 0.00
}

func ExampleAtan() {
	fmt.Printf("%.2f", math.Atan(0))
	// Output: 0.00
}

func ExampleAtan2() {
	fmt.Printf("%.2f", math.Atan2(0, 0))
	// Output: 0.00
}

func ExampleAtanh() {
	fmt.Printf("%.2f", math.Atanh(0))
	// Output: 0.00
}

func ExampleCopysign() {
	fmt.Printf("%.2f", math.Copysign(3.2, -1))
	// Output: -3.20
}

    func.func @main(%arg0: tensor<2xf32>) -> tensor<2xf32> {
      %0 = "tf.Const"() {value = dense<1.42> : tensor<2xf32>} : () -> tensor<2xf32>
      %1 = "tf.Atan2"(%arg0, %0) : (tensor<2xf32>, tensor<2xf32>) -> tensor<2xf32>
      func.return %0 : tensor<2xf32>
    }
  })";

  TF_ASSERT_OK(LegalizeModule(kModuleWithConstParam));
}

#include "tensorflow/core/lib/random/random.h"

namespace tensorflow {
namespace {

using ops::Abs;
using ops::Add;
using ops::AddN;
using ops::AddV2;
using ops::Atan2;
using ops::BatchMatMul;
using ops::BatchMatMulV3;
using ops::Cast;
using ops::ClipByValue;
using ops::Const;
using ops::Cumsum;
using ops::Div;
using ops::DivNoNan;
using ops::MatMul;
using ops::Max;

  RANDOM_UNIFORM = 148,
  MULTINOMIAL = 149,
  GELU = 150,
  DYNAMIC_UPDATE_SLICE = 151,
  RELU_0_TO_1 = 152,
  UNSORTED_SEGMENT_PROD = 153,
  UNSORTED_SEGMENT_MAX = 154,
  UNSORTED_SEGMENT_SUM = 155,
  ATAN2 = 156,
  UNSORTED_SEGMENT_MIN = 157,
  SIGN = 158,
  BITCAST = 159,
  BITWISE_XOR = 160,
  RIGHT_SHIFT = 161,
  // All Operators start with STABLEHLO_ prefixes are subject to change

// CHECK:  return %[[RES0]] : tensor<1xi64>
}

func.func @atan2(%arg0: tensor<8xf32>, %arg1: tensor<8xf32>) -> tensor<8xf32> {
  %0 = "tf.Atan2"(%arg0, %arg1) : (tensor<8xf32>, tensor<8xf32>) -> tensor<8xf32>
  func.return %0 : tensor<8xf32>

// CHECK-LABEL: atan2
// CHECK: %[[RES0:.*]] = "tfl.atan2"(%arg0, %arg1) : (tensor<8xf32>, tensor<8xf32>) -> tensor<8xf32>

  grad_outputs->push_back(Mul(scope, x, grad_inv));
  grad_outputs->push_back(Mul(scope, Neg(scope, y), grad_inv));
  return scope.status();
}
REGISTER_GRADIENT_OP("Atan2", Atan2Grad);

// BinaryGradCommon handles the setup for binary ops that broadcast
// their inputs.
Status BinaryGradCommon(const Scope& scope, const Operation& op,

  );

  let results = (outs TFL_TensorOf<[F32, I32]>:$output);
}

def TFL_Atan2Op: TFL_Op<"atan2", [
  Pure,
  SameOperandsAndResultShape,
  SameOperandsAndResultElementType]> {

  let summary = "Atan2 operation";
  let description = [{
    The "atan2" operation computes the arctangent of y/x element-wise,
    respecting signs of the arguments.
  }];