IQTransform.cpp

// 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 +
//----------------------------------------------------------------------
// Includes
//----------------------------------------------------------------------
#include <utility>

#include "MantidAPI/Axis.h"
#include "MantidAPI/IncreasingAxisValidator.h"
#include "MantidAPI/MatrixWorkspace.h"
#include "MantidAPI/RawCountValidator.h"
#include "MantidAPI/WorkspaceUnitValidator.h"
#include "MantidAlgorithms/IQTransform.h"
#include "MantidDataObjects/TableWorkspace.h"
#include "MantidDataObjects/Workspace2D.h"
#include "MantidDataObjects/WorkspaceCreation.h"
#include "MantidHistogramData/Histogram.h"
#include "MantidKernel/ArrayProperty.h"
#include "MantidKernel/BoundedValidator.h"
#include "MantidKernel/CompositeValidator.h"
#include "MantidKernel/ListValidator.h"
#include "MantidKernel/Unit.h"
#include "MantidKernel/VectorHelper.h"

using namespace Mantid::DataObjects;
using namespace Mantid::HistogramData;

namespace Mantid::Algorithms {

// Register the algorithm into the AlgorithmFactory
DECLARE_ALGORITHM(IQTransform)

using namespace Kernel;
using namespace API;

IQTransform::IQTransform() : API::Algorithm(), m_label(std::make_shared<Units::Label>()) {
  /* Just for fun, this is implemented as follows....
   * We fill a map below with the transformation name as the key and
   * a function pointer to the method that does the transformation as
   * the value. The 'TransformType' property is filled with the keys
   * and then we search on that to select the correct function in exec.
   */
  m_transforms["Guinier (spheres)"] = &IQTransform::guinierSpheres;
  m_transforms["Guinier (rods)"] = &IQTransform::guinierRods;
  m_transforms["Guinier (sheets)"] = &IQTransform::guinierSheets;
  m_transforms["Zimm"] = &IQTransform::zimm;
  m_transforms["Debye-Bueche"] = &IQTransform::debyeBueche;
  m_transforms["Holtzer"] = &IQTransform::holtzer;
  m_transforms["Kratky"] = &IQTransform::kratky;
  m_transforms["Porod"] = &IQTransform::porod;
  m_transforms["Log-Log"] = &IQTransform::logLog;
  m_transforms["General"] = &IQTransform::general;
}

void IQTransform::init() {
  auto wsValidator = std::make_shared<CompositeValidator>();
  // Require the input to be in units of Q and to be a distribution
  // (which the result of a SANS reduction in Mantid will be)
  wsValidator->add<WorkspaceUnitValidator>("MomentumTransfer");
  wsValidator->add<RawCountValidator>(false);
  // Require X data to be increasing from left to right
  wsValidator->add<IncreasingAxisValidator>();

  declareProperty(std::make_unique<WorkspaceProperty<>>("InputWorkspace", "", Direction::Input, wsValidator),
                  "The input workspace must be a distribution with units of Q");
  declareProperty(std::make_unique<WorkspaceProperty<>>("OutputWorkspace", "", Direction::Output),
                  "The name of the output workspace");

  // Extract the keys from the transformations map to pass to the property
  std::set<std::string> plottype;
  for (TransformMap::const_iterator it = m_transforms.begin(); it != m_transforms.end(); ++it) {
    plottype.insert(it->first);
  }
  declareProperty("TransformType", "", std::make_shared<StringListValidator>(plottype),
                  "The name of the transformation to be performed on the workspace");

  // A background to be subtracted can be a value or a workspace. Both
  // properties are optional.
  auto mustBePositive = std::make_shared<BoundedValidator<double>>();
  mustBePositive->setLower(0.0);
  declareProperty("BackgroundValue", 0.0, mustBePositive,
                  "A constant value to subtract from the data prior to its transformation");
  declareProperty(
      std::make_unique<WorkspaceProperty<>>("BackgroundWorkspace", "", Direction::Input, PropertyMode::Optional),
      "A workspace to subtract from the input workspace prior to its "
      "transformation."
      "Must be compatible with the input (as for the Minus algorithm).");

  declareProperty(std::make_unique<ArrayProperty<double>>("GeneralFunctionConstants"),
                  "A set of 10 constants to be used (only) with the 'General' "
                  "transformation");
}

void IQTransform::exec() {
  MatrixWorkspace_sptr inputWS = getProperty("InputWorkspace");
  // Print a warning if the input workspace has more than one spectrum
  if (inputWS->getNumberHistograms() > 1) {
    g_log.warning("This algorithm is intended for use on single-spectrum workspaces.\n"
                  "Only the first spectrum will be transformed.");
  }

  // Do background subtraction from a workspace first because it doesn't like
  // potential conversion to point data that follows. Requires a temporary
  // workspace.
  MatrixWorkspace_sptr tmpWS;
  MatrixWorkspace_sptr backgroundWS = getProperty("BackgroundWorkspace");
  if (backgroundWS)
    tmpWS = subtractBackgroundWS(inputWS, backgroundWS);
  else
    tmpWS = inputWS;

  // Create the output workspace
  const size_t length = tmpWS->blocksize();
  MatrixWorkspace_sptr outputWS = create<MatrixWorkspace>(*inputWS, 1, Points(length));
  m_label->setLabel("");
  outputWS->setYUnit("");
  // Copy the data over. Assume single spectrum input (output will be).
  outputWS->setPoints(0, tmpWS->points(0));
  outputWS->setSharedY(0, tmpWS->sharedY(0));
  outputWS->setSharedE(0, tmpWS->sharedE(0));

  // Subtract a constant background if requested
  const double background = getProperty("BackgroundValue");
  if (background > 0.0)
    outputWS->mutableY(0) -= background;

  // Select the desired transformation function and call it
  TransformFunc f = m_transforms.find(getProperty("TransformType"))->second;
  (this->*f)(outputWS);

  // Need the generic label unit on this (unless the unit on the X axis hasn't
  // changed)
  if (!m_label->caption().empty())
    outputWS->getAxis(0)->unit() = m_label;
  setProperty("OutputWorkspace", outputWS);
}

/** Uses the Minus algorithm to subtract the background workspace from the given
 * workspace.
 *  If the ChildAlgorithm fails (e.g. if the background workspace is the wrong
 * size), then this
 *  entire algorithm will.
 *  @param ws         The workspace to perform the subtraction on
 *  @param background The workspace containing the background values
 */
API::MatrixWorkspace_sptr IQTransform::subtractBackgroundWS(const API::MatrixWorkspace_sptr &ws,
                                                            const API::MatrixWorkspace_sptr &background) {
  g_log.debug() << "Subtracting the workspace " << background->getName() << " from the input workspace.\n";
  return ws - background;
}

/** @name Available transformation functions */
//@{

/** Performs the Guinier (spheres) transformation: Ln(I) v Q^2
 *  @param ws The workspace to be transformed
 *  @throw std::range_error if an attempt is made to take log of a negative
 * number
 */
void IQTransform::guinierSpheres(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);
  std::transform(X.cbegin(), X.cend(), X.begin(), VectorHelper::Squares<double>());
  std::transform(E.cbegin(), E.cend(), Y.begin(), E.begin(), std::divides<double>());
  std::transform(Y.cbegin(), Y.cend(), Y.begin(), VectorHelper::LogNoThrow<double>());

  ws->setYUnitLabel("Ln(I)");
  m_label->setLabel("Q^2");
}

/** Performs the Guinier (rods) transformation: Ln(IQ) v Q^2
 *  @param ws The workspace to be transformed
 *  @throw std::range_error if an attempt is made to take log of a negative
 * number
 */
void IQTransform::guinierRods(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);

  std::transform(E.cbegin(), E.cend(), Y.begin(), E.begin(), std::divides<double>());
  std::transform(Y.cbegin(), Y.cend(), X.begin(), Y.begin(), std::multiplies<double>());
  std::transform(Y.cbegin(), Y.cend(), Y.begin(), VectorHelper::LogNoThrow<double>());
  std::transform(X.cbegin(), X.cend(), X.begin(), VectorHelper::Squares<double>());

  ws->setYUnitLabel("Ln(I x Q)");
  m_label->setLabel("Q^2");
}

/** Performs the Guinier (sheets) transformation: Ln(IQ^2) v Q^2
 *  @param ws The workspace to be transformed
 *  @throw std::range_error if an attempt is made to take log of a negative
 * number
 */
void IQTransform::guinierSheets(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);

  std::transform(E.cbegin(), E.cend(), Y.begin(), E.begin(), std::divides<double>());
  std::transform(X.cbegin(), X.cend(), X.begin(), VectorHelper::Squares<double>());
  std::transform(Y.cbegin(), Y.cend(), X.begin(), Y.begin(), std::multiplies<double>());
  std::transform(Y.cbegin(), Y.cend(), Y.begin(), VectorHelper::LogNoThrow<double>());

  ws->setYUnitLabel("Ln(I x Q^2)");
  m_label->setLabel("Q^2");
}

/** Performs the Zimm transformation: 1/I v Q^2
 *  The output is set to zero for negative input Y values
 *  @param ws The workspace to be transformed
 */
void IQTransform::zimm(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);
  std::transform(X.cbegin(), X.cend(), X.begin(), VectorHelper::Squares<double>());
  for (size_t i = 0; i < Y.size(); ++i) {
    if (Y[i] > 0.0) {
      Y[i] = 1.0 / Y[i];
      E[i] *= std::pow(Y[i], 2);
    } else {
      Y[i] = 0.0;
      E[i] = 0.0;
    }
  }

  ws->setYUnitLabel("1/I");
  m_label->setLabel("Q^2");
}

/** Performs the Debye-Bueche transformation: 1/sqrt(I) v Q^2
 *  The output is set to zero for negative input Y values
 *  @param ws The workspace to be transformed
 */
void IQTransform::debyeBueche(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);
  std::transform(X.cbegin(), X.cend(), X.begin(), VectorHelper::Squares<double>());
  for (size_t i = 0; i < Y.size(); ++i) {
    if (Y[i] > 0.0) {
      Y[i] = 1.0 / std::sqrt(Y[i]);
      E[i] *= std::pow(Y[i], 3);
    } else {
      Y[i] = 0.0;
      E[i] = 0.0;
    }
  }

  ws->setYUnitLabel("1/sqrt(I)");
  m_label->setLabel("Q^2");
}

/** Performs the Holtzer transformation: IQ v Q
 *  @param ws The workspace to be transformed
 */
void IQTransform::holtzer(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);
  std::transform(Y.cbegin(), Y.cend(), X.begin(), Y.begin(), std::multiplies<double>());
  std::transform(E.cbegin(), E.cend(), X.begin(), E.begin(), std::multiplies<double>());

  ws->setYUnitLabel("I x Q");
}

/** Performs the Kratky transformation: IQ^2 v Q
 *  @param ws The workspace to be transformed
 */
void IQTransform::kratky(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);
  MantidVec Q2(X.size());
  std::transform(X.cbegin(), X.cend(), Q2.begin(), VectorHelper::Squares<double>());
  std::transform(Y.cbegin(), Y.cend(), Q2.begin(), Y.begin(), std::multiplies<double>());
  std::transform(E.cbegin(), E.cend(), Q2.begin(), E.begin(), std::multiplies<double>());

  ws->setYUnitLabel("I x Q^2");
}

/** Performs the Porod transformation: IQ^4 v Q
 *  @param ws The workspace to be transformed
 */
void IQTransform::porod(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);
  MantidVec Q4(X.size());
  std::transform(X.cbegin(), X.cend(), X.cbegin(), Q4.begin(), VectorHelper::TimesSquares<double>());
  std::transform(Y.cbegin(), Y.cend(), Q4.begin(), Y.begin(), std::multiplies<double>());
  std::transform(E.cbegin(), E.cend(), Q4.begin(), E.begin(), std::multiplies<double>());

  ws->setYUnitLabel("I x Q^4");
}

/** Performs a log-log transformation: Ln(I) v Ln(Q)
 *  @param ws The workspace to be transformed
 *  @throw std::range_error if an attempt is made to take log of a negative
 * number
 */
void IQTransform::logLog(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);

  std::transform(X.cbegin(), X.cend(), X.begin(), VectorHelper::Log<double>());
  std::transform(E.cbegin(), E.cend(), Y.begin(), E.begin(), std::divides<double>());
  std::transform(Y.cbegin(), Y.cend(), Y.begin(), VectorHelper::LogNoThrow<double>());

  ws->setYUnitLabel("Ln(I)");
  m_label->setLabel("Ln(Q)");
}

/** Performs a transformation of the form: Q^A x I^B x Ln(Q^C x I^D x E) v Q^F x
 * I^G x Ln(Q^H x I^I x J).
 *  Uses the 'GeneralFunctionConstants' property where A-J are the 10 (ordered)
 * input constants.
 *  @param ws The workspace to be transformed
 *  @throw std::range_error if an attempt is made to take log of a negative
 * number
 */
void IQTransform::general(const API::MatrixWorkspace_sptr &ws) {
  auto &X = ws->mutableX(0);
  auto &Y = ws->mutableY(0);
  auto &E = ws->mutableE(0);
  const std::vector<double> C = getProperty("GeneralFunctionConstants");
  // Check for the correct number of elements
  if (C.size() != 10) {
    std::string mess("The General transformation requires 10 values to be provided.");
    g_log.error(mess);
    throw std::invalid_argument(mess);
  }

  for (size_t i = 0; i < Y.size(); ++i) {
    double tmpX = std::pow(X[i], C[7]) * std::pow(Y[i], C[8]) * C[9];
    if (tmpX <= 0.0)
      throw std::range_error("Attempt to take log of a zero or negative number.");
    tmpX = std::pow(X[i], C[5]) * std::pow(Y[i], C[6]) * std::log(tmpX);
    const double tmpY = std::pow(X[i], C[2]) * std::pow(Y[i], C[3]) * C[4];
    if (tmpY <= 0.0)
      throw std::range_error("Attempt to take log of a zero or negative number.");
    const double newY = std::pow(X[i], C[0]) * std::pow(Y[i], C[1]) * std::log(tmpY);

    E[i] *= std::pow(X[i], C[0]) *
            (C[1] * std::pow(Y[i], C[1] - 1) * std::log(tmpY) +
             ((std::pow(Y[i], C[1]) * std::pow(X[i], C[2]) * C[4] * C[3] * std::pow(Y[i], C[3] - 1)) / tmpY));
    X[i] = tmpX;
    Y[i] = newY;
  }

  std::stringstream ylabel;
  ylabel << "Q^" << C[0] << " x I^" << C[1] << " x Ln( Q^" << C[2] << " x I^" << C[3] << " x " << C[4] << ")";
  ws->setYUnitLabel(ylabel.str());
  std::stringstream xlabel;
  xlabel << "Q^" << C[5] << " x I^" << C[6] << " x Ln( Q^" << C[7] << " x I^" << C[8] << " x " << C[9] << ")";
  m_label->setLabel(xlabel.str());
}

//@}

} // namespace Mantid::Algorithms