Implicit conversion from float -113.47003154574134 to int loses precision