InelasticDiffSphere.cpp

Go to the documentation of this file.

// Mantid Repository : https://github.com/mantidproject/mantid
//
// Copyright © 2018 ISIS Rutherford Appleton Laboratory UKRI,
//   NScD Oak Ridge National Laboratory, European Spallation Source,
//   Institut Laue - Langevin & CSNS, Institute of High Energy Physics, CAS
// SPDX - License - Identifier: GPL - 3.0 +
#include "MantidCurveFitting/Functions/InelasticDiffSphere.h"
 
#include "MantidAPI/FunctionFactory.h"
#include "MantidAPI/IFunction.h"
#include "MantidAPI/MatrixWorkspace.h"
#include "MantidAPI/SpectrumInfo.h"
#include "MantidGeometry/IDetector.h"
#include "MantidGeometry/Instrument/DetectorInfo.h"
#include "MantidKernel/Exception.h"
#include "MantidKernel/UnitConversion.h"
#include "MantidTypes/SpectrumDefinition.h"
 
#include <boost/math/special_functions/bessel.hpp>
 
#include <cmath>
#include <limits>
 
namespace {
Mantid::Kernel::Logger g_log("InelasticDiffSphere");
}
 
namespace Mantid::CurveFitting::Functions {
 
DECLARE_FUNCTION(InelasticDiffSphere)
 
 
InelasticDiffSphere::InelasticDiffSphere() : m_lmax(24), m_divZone(0.1), m_hbar(0.658211626) {
  this->declareParameter("Intensity", 1.0, "scaling factor");
  this->declareParameter("Radius", 2.0, "Sphere radius, in Angstroms");
  this->declareParameter("Diffusion", 0.05,
                         "Diffusion coefficient, in units of "
                         "A^2*THz, if energy in meV, or A^2*PHz "
                         "if energy in ueV");
  this->declareParameter("Shift", 0.0, "Shift in domain");
 
  declareAttributes();
}
 
void InelasticDiffSphere::initXnlCoeff() {
  /* List of 98 coefficients sorted by increasing value (F.Volino,Mol. Phys.
   * 41,271-279,1980)
   * For each coefficient, the triad (coeff, l, n) is defined
   */
  size_t ncoeff = 98;
 
  double xvalues[] = {2.081576,  3.342094,  4.493409,  4.514100,  5.646704,  5.940370,  6.756456,  7.289932,  7.725252,
                      7.851078,  8.583755,  8.934839,  9.205840,  9.840446,  10.010371, 10.613855, 10.904122, 11.070207,
                      11.079418, 11.972730, 12.143204, 12.279334, 12.404445, 13.202620, 13.295564, 13.472030, 13.846112,
                      14.066194, 14.258341, 14.590552, 14.651263, 15.244514, 15.310887, 15.579236, 15.819216, 15.863222,
                      16.360674, 16.609346, 16.977550, 17.042902, 17.117506, 17.220755, 17.408034, 17.947180, 18.127564,
                      18.356318, 18.453241, 18.468148, 18.742646, 19.262710, 19.270294, 19.496524, 19.581889, 19.862424,
                      20.221857, 20.371303, 20.406581, 20.538074, 20.559428, 20.795967, 21.231068, 21.537120, 21.578053,
                      21.666607, 21.840012, 21.899697, 21.999955, 22.578058, 22.616601, 22.662493, 23.082796, 23.106568,
                      23.194996, 23.390490, 23.519453, 23.653839, 23.783192, 23.906450, 24.360789, 24.382038, 24.474825,
                      24.689873, 24.850085, 24.899636, 25.052825, 25.218652, 25.561873, 25.604057, 25.724794, 25.846084,
                      26.012188, 26.283265, 26.516603, 26.552589, 26.666054, 26.735177, 26.758685, 26.837518};
 
  size_t lvalues[] = {1,  2,  0, 3,  4,  1, 5,  2,  0,  6,  3, 7,  1,  4,  8, 2,  0,  5,  9,  3,  10, 6,  1,  11, 4,
                      7,  2,  0, 12, 5,  8, 3,  13, 1,  9,  6, 14, 4,  10, 2, 7,  0,  15, 5,  11, 8,  16, 3,  1,  6,
                      12, 17, 9, 4,  2,  0, 13, 18, 7,  10, 5, 14, 19, 3,  8, 1,  11, 6,  20, 15, 4,  9,  12, 2,  0,
                      21, 16, 7, 10, 13, 5, 22, 3,  17, 1,  8, 14, 11, 23, 6, 18, 4,  9,  2,  0,  15, 24, 12};
 
  size_t nvalues[] = {0, 0, 1, 0, 0, 1, 0, 1, 2, 0, 1, 0, 2, 1, 0, 2, 3, 1, 0, 2, 0, 1, 3, 0, 2, 1, 3, 4, 0, 2, 1, 3, 0,
                      4, 1, 2, 0, 3, 1, 4, 2, 5, 0, 3, 1, 2, 0, 4, 5, 3, 1, 0, 2, 4, 5, 6, 1, 0, 3, 2, 4, 1, 0, 5, 3, 6,
                      2, 4, 0, 1, 5, 3, 2, 6, 7, 0, 1, 4, 3, 2, 5, 0, 6, 1, 7, 4, 2, 3, 0, 5, 1, 6, 4, 7, 8, 2, 0, 3};
 
  for (size_t i = 0; i < ncoeff; i++) {
    xnlc coeff;
    coeff.x = xvalues[i];
    coeff.l = lvalues[i];
    coeff.n = nvalues[i];
    m_xnl.emplace_back(coeff);
  }
}
 
void InelasticDiffSphere::initAlphaCoeff() {
  for (std::vector<xnlc>::const_iterator it = m_xnl.begin(); it != m_xnl.end(); ++it) {
    double x = it->x; // eigenvalue for a (n, l) pair
    auto l = static_cast<double>(it->l);
    m_alpha.emplace_back((2.0 * l + 1) * 6.0 * x * x / (x * x - l * (l + 1)));
  }
}
 
void InelasticDiffSphere::initLinJlist() {
  for (auto &coeff : m_xnl) {
    linearJ abJ;
    auto x = coeff.x; // eigenvalue for a (n, l) pair
    auto l = static_cast<unsigned int>(coeff.l);
    double Qa = x - m_divZone; // left of the numerical divergence point
    double J0 = (Qa * boost::math::sph_bessel(l + 1, Qa) - l * boost::math::sph_bessel(l, Qa)) / (Qa * Qa - x * x);
    Qa = x + m_divZone; // right of the numerical divergence point
    double J1 = (Qa * boost::math::sph_bessel(l + 1, Qa) - l * boost::math::sph_bessel(l, Qa)) / (Qa * Qa - x * x);
    abJ.slope = (J1 - J0) / (2 * m_divZone);          // slope of the linear
                                                      // interpolation
    abJ.intercept = J0 - abJ.slope * (x - m_divZone); // intercept of the linear interpolation
    m_linearJlist.emplace_back(abJ);                  // store the parameters of the linear interpolation for this it->x
  }
}
 
void InelasticDiffSphere::init() {
  this->initXnlCoeff();   // initialize m_xnl with the list of coefficients xnlist
  this->initAlphaCoeff(); // initialize m_alpha, certain factors constant over
                          // the fit
  this->initLinJlist();   // initialize m_linearJlist, linear interpolation around
                          // numerical divergence
}
 
std::vector<double> InelasticDiffSphere::LorentzianCoefficients(double a) const {
  // precompute the 2+m_lmax spherical bessel functions (26 in total)
  std::vector<double> jl(2 + m_lmax);
  for (size_t l = 0; l < 2 + m_lmax; l++) {
    jl[l] = boost::math::sph_bessel(static_cast<unsigned int>(l), a);
  }
 
  // store the coefficient of each Lorentzian in vector YJ(a,w)
  size_t ncoeff = m_xnl.size();
  std::vector<double> YJ(ncoeff);
 
  for (size_t i = 0; i < ncoeff; i++) {
    auto x = m_xnl[i].x;
    auto l = static_cast<unsigned int>(m_xnl[i].l);
 
    double J;
    if (fabs(a - x) > m_divZone) {
      J = (a * jl[l + 1] - l * jl[l]) / (a * a - x * x);
    } else {
      J = m_linearJlist[i].slope * a + m_linearJlist[i].intercept; // linear interpolation instead
    }
 
    YJ[i] = m_alpha[i] * (J * J);
  }
 
  return YJ;
} // end of LorentzianCoefficients
 
void InelasticDiffSphere::function1D(double *out, const double *xValues, const size_t nData) const {
  auto I = this->getParameter("Intensity");
  auto R = this->getParameter("Radius");
  auto D = this->getParameter("Diffusion");
  auto S = this->getParameter("Shift");
 
  double Q;
  if (this->getAttribute("Q").asDouble() == EMPTY_DBL()) {
    if (m_qValueCache.empty()) {
      throw std::runtime_error("No Q attribute provided and cannot retrieve from worksapce.");
    }
 
    auto specIdx = this->getAttribute("WorkspaceIndex").asInt();
    Q = m_qValueCache[specIdx];
 
    g_log.debug() << "Get Q value for workspace index " << specIdx << ": " << Q << '\n';
  } else {
    Q = this->getAttribute("Q").asDouble();
 
    g_log.debug() << "Using Q attribute: " << Q << '\n';
  }
 
  // // Penalize negative parameters
  if (I < std::numeric_limits<double>::epsilon() || R < std::numeric_limits<double>::epsilon() ||
      D < std::numeric_limits<double>::epsilon()) {
    for (size_t i = 0; i < nData; i++) {
      out[i] = std::numeric_limits<double>::infinity();
    }
    return;
  }
 
  // List of Lorentzian HWHM
  std::vector<double> HWHM;
  size_t ncoeff = m_xnl.size();
  for (size_t n = 0; n < ncoeff; n++) {
    auto x = m_xnl[n].x; // eigenvalue
    HWHM.emplace_back(m_hbar * x * x * D / (R * R));
  }
 
  std::vector<double> YJ;
  YJ = this->LorentzianCoefficients(Q * R); // The (2l+1)*A_{n,l}
  for (size_t i = 0; i < nData; i++) {
    double energy = xValues[i] - S; // from meV to THz (or from micro-eV to PHz)
    out[i] = 0.0;
    for (size_t n = 0; n < ncoeff; n++) {
      double L = (1.0 / M_PI) * HWHM[n] / (HWHM[n] * HWHM[n] + energy * energy); // Lorentzian
      out[i] += I * YJ[n] * L;
    }
  }
}
 
void InelasticDiffSphere::setWorkspace(std::shared_ptr<const API::Workspace> ws) {
  m_qValueCache.clear();
 
  auto workspace = std::dynamic_pointer_cast<const API::MatrixWorkspace>(ws);
  if (!workspace)
    return;
 
  const auto &spectrumInfo = workspace->spectrumInfo();
  const auto &detectorIDs = workspace->detectorInfo().detectorIDs();
  for (size_t idx = 0; idx < spectrumInfo.size(); idx++) {
    if (!spectrumInfo.hasDetectors(idx)) {
      m_qValueCache.clear();
      g_log.information("Cannot populate Q values from workspace - no detectors set.");
      break;
    }
 
    const auto detectorIndex = spectrumInfo.spectrumDefinition(idx)[0].first;
 
    try {
      double efixed = workspace->getEFixed(detectorIDs[detectorIndex]);
      double usingTheta = 0.5 * spectrumInfo.twoTheta(idx);
 
      double q = Mantid::Kernel::UnitConversion::convertToElasticQ(usingTheta, efixed);
 
      m_qValueCache.emplace_back(q);
    } catch (std::runtime_error &) {
      m_qValueCache.clear();
      g_log.information("Cannot populate Q values from workspace - could not "
                        "find EFixed value.");
      return;
    }
  }
}
 
} // namespace Mantid::CurveFitting::Functions