Nevanlinna_kernel.hpp

#ifndef TRIQS_NEVANLINNA_NEVANLINNA_KERNEL_HPP
#define TRIQS_NEVANLINNA_NEVANLINNA_KERNEL_HPP
 
#include <complex>
 
#include <nda/nda.hpp>
#include <nda/mpi.hpp>
#include <triqs/utility/macros.hpp>
 
#include "kernel.hpp"
#include "Nevanlinna_factorization.hpp"
 
namespace triqs_Nevanlinna {

  class Nevanlinna_kernel : public kernel {
 
    public:
    Nevanlinna_kernel(int precision = mp_digits) : kernel(precision) {
      if (!mpi::communicator().rank())
        std::cerr << "This is Nevanlinna analytical continuation. All off-diagonal elements will be ignored." << std::endl;
    }

    void init(nda::vector_const_view<std::complex<double>> mesh, nda::array_const_view<std::complex<double>, 3> data) override;

    [[nodiscard]] nda::array<std::complex<double>, 3> evaluate(nda::vector_const_view<std::complex<double>> grid) override;

    [[nodiscard]] nda::array<std::complex<double>, 3> evaluate(nda::vector_const_view<std::complex<double>> grid,
                                                               nda::array_const_view<std::complex<double>, 3> theta) override;

    [[nodiscard]] C2PY_PROPERTY_GET(size) size_t size() const override { return _factorizations.size(); }

    [[nodiscard]] C2PY_PROPERTY_GET(Pick_eigenvalues) nda::vector<double> get_Pick_eigenvalues() const override;
 
    private:
    size_t _N_im_freq{};
    std::vector<Nevanlinna_factorization> _factorizations{};
  };
 
} // namespace triqs_Nevanlinna
#endif //TRIQS_NEVANLINNA_NEVANLINNA_KERNEL_HPP