Monica Monica - 24 days ago 10
Python Question

How use boost tuple to return two vectors

I'm rewriting a python code to C++ and had encountered a problem. I created a function in python:

from my_vec import Vector

def param(gamma):

coeff = Vector(6)
exp = Vector(6)

if abs(gamma - 0.3) < 1E-6:
coeff[0] = 33; exp[0] = 0.1;
coeff[1] = -33; exp[1] = 2.2;
coeff[2] = 21; exp[2] = 0.16;
coeff[3] = 23; exp[3] = 3.312;
coeff[4] = 23; exp[4] = 100;
coeff[5] = 32; exp[5] = 59.00;
elif abs(gamma - 0.4) < 1E-6:
coeff[0] = -0.23; exp[0] = 0.02;
coeff[1] = -0.48; exp[1] = 0.18;
coeff[2] = 200; exp[2] = 1.82;
coeff[3] = 200; exp[3] = 3.71;
coeff[4] = 200; exp[4] = 14.28;
coeff[5] = 0.00; exp[5] = 79.11;

return coeff, exp


Subsequently, I made an attempt to write the same use in C++. It seems that using tuple is what I'm looking for.
However, I haven't uses it before and I'm struggling with this task.

I created:

std::vector<float> coef1 = {-0.366770, -0.249990, -0.411230, 0.144690, -0.101790, 0.010510};
std::vector<float> exp1 = {0.02, 1.910630, 0.16492, 3.31721, 10.45634, 59.3438};
std::vector<float> coef2 = {-0.36, -0.24, -0.41, 0.14, -0.10, 0.01};

std::vector<float> exp2 = {0.02, 1.23, 12, 3.31, 11, 22};

typedef boost::tuple<std::vector<float>, std::vector<float>> parameters;
parameters param1{coef1, exp1};
parameters param2{coef2, exp2};


At this point, I'd love to create conditions like
if abs(gamma - 0.3) < 1E-6
etc end return coeff and exp vectors. How can I create a function like this one written in python?

Answer

Here's a quick translate.¹

There's a subtle point where inclusion of <cmath> and using std::abs is essential to get the floating point version of abs.

Live On Coliru

#include <tuple>
#include <vector>
#include <cmath>

using vec6 = std::array<double, 6>;
using params = std::tuple<vec6, vec6>;

params param(double gamma) {
    using std::abs; // enable ADL in case you use e.g. boost::multiprecision as the vector element type
    if (abs(gamma - 0.3) < 1E-6) {
        return std::make_tuple ( 
                vec6{{-0.366770, -0.249990, -0.411230, 0.144690, -0.101790, 0.010510}}, // coeff1
                vec6{{0.02, 1.910630, 0.16492, 3.31721, 10.45634, 59.3438}} ); // exp1
    }
    if (abs(gamma - 0.4) < 1E-6) {
        return std::make_tuple ( 
                vec6{{-0.36, -0.24, -0.41, 0.14, -0.10, 0.01}}, // coeff2
                vec6{{0.02, 1.23, 12, 3.31, 11, 22}} ); // exp2
    }

    throw std::runtime_error("unhandled gamma");
}

int main() {
    auto p = param(0.4);
}

¹ using the values from the tentative C++ version, not the python version