add missing methods for the RegularGridInterpolator

[flabowski]

Hei,
consider the following example (based on scipys RegularGridInterpolator)

// Standard
#include <iostream>
#include <numeric>
#include <vector>
#include <math.h>
// btwxt
#include "btwxt.h"
#include "Linspace.h"

using namespace Btwxt;

std::vector<double> f(std::vector<double> x, std::vector<double> y, std::vector<double> z) {
	int d1 = x.size();
	int d2 = y.size();
	int d3 = z.size();
	size_t flat_idx = 0;
	std::vector<double> result(d1*d2*d3);
	for (int i = 0; i < d1; i++) {
		for (int j = 0; j < d2; j++) {
			for (int k = 0; k < d3; k++) {
				flat_idx = i * d2 * d3 + j * d3 + k;
				result[flat_idx] = 2 * std::pow(x[i], 3) + 3 * std::pow(y[j], 2) - z[k];
			}
		}
	}
	return result;
}

void scipy_rgi_example() {
	std::vector<double> x = linspace(1.0, 4.0, 11);
	std::vector<double> y = linspace(4.0, 7.0, 22);
	std::vector<double> z = linspace(7.0, 9.0, 33);
	std::vector<std::vector<double>> data = { f(x, y, z) };

	std::vector<std::vector<double>> vector_grid = {x, y, z};
	RegularGridInterpolator my_interpolating_function(vector_grid, data);

	my_interpolating_function.set_axis_interp_method(0, Method::LINEAR);
	my_interpolating_function.set_axis_interp_method(1, Method::LINEAR);
	my_interpolating_function.set_axis_interp_method(2, Method::LINEAR);

	my_interpolating_function.set_axis_extrap_method(0, Method::LINEAR); // not implemented yet
	my_interpolating_function.set_axis_extrap_limits(1, { 2, 10 });  // not implemented yet

	print_vector(my_interpolating_function({ 2.1, 6.2, 8.3 }));  // 125.805
	print_vector(my_interpolating_function({ 3.3, 5.2, 7.1 }));  // 146.301
}

with linspace.h given as:

#pragma once
#include <iostream>
#include <vector>

//https://gist.github.com/lorenzoriano/5414671
template <typename T>
std::vector<T> linspace(T a, T b, size_t N) {
	T h = (b - a) / static_cast<T>(N - 1);
	std::vector<T> xs(N);
	typename std::vector<T>::iterator x;
	T val;
	for (x = xs.begin(), val = a; x != xs.end(); ++x, val += h)
		*x = val;
	return xs;
}


void print_vector(std::vector<double> vec)
{
	std::cout << "size: " << vec.size() << std::endl;
	for (double d : vec)
		std::cout << d << " ";
	std::cout << std::endl;
}

Since GriddedData grid_data; is private, I need to set the extrapolation properties via the RegularGridInterpolator.

[tanaya-mankad]

This is quite reasonable; would you like to add the implementations as well?

[flabowski]

Not sure what you mean, maybe i was a bit unclear. The methods set_axis_extrap_limits and set_axis_extrap_method are already implemented in the GriddedData class. It's quite convinient to work with the RegularGridInterpolator class only as shown in my first comment, but since the GriddedData instance is private, you can't modify it. The changes i proposed in btwxt.h allow the user to do that.