[PWGCF] Add spherical harmonics decomposition to unified Femto framework

Author: IgorPtak

PWGCF/Femto/Core/femtoSpherHarMath.h

@@ -0,0 +1,225 @@
+// Copyright 2019-2025 CERN and copyright holders of ALICE O2.
+// See https://alice-o2.web.cern.ch/copyright for details of the copyright holders.
+// All rights not expressly granted are reserved.
+//
+// This software is distributed under the terms of the GNU General Public
+// License v3 (GPL Version 3), copied verbatim in the file "COPYING".
+//
+// In applying this license CERN does not waive the privileges and immunities
+// granted to it by virtue of its status as an Intergovernmental Organization
+// or submit itself to any jurisdiction.
+
+/// \file femtoSpherHarMath.h
+/// \brief Container for the calculation of spherical harmonics components
+/// \author Pritam Chakraborty, WUT Warsaw, pritam.chakraborty@pw.edu.pl
+
+#ifndef PWGCF_FEMTO_CORE_FEMTOSPHERHARMATH_H_
+#define PWGCF_FEMTO_CORE_FEMTOSPHERHARMATH_H_
+
+#include <CommonConstants/MathConstants.h>
+
+#include <array>
+#include <cmath>
+#include <complex>
+
+namespace o2::analysis::femto
+{
+
+/// \class SpherHarMath
+/// \brief Container for math calculations of quantities related to pairs
+class SpherHarMath
+{
+ public:
+  static constexpr int MaxSupportedL = 5;                 // analytic Ylm implemented up to l=5
+  static constexpr int TrigCacheSize = MaxSupportedL + 1; // sin/cos powers cache size (indices 0..5)
+  static constexpr double SmallLength = 1e-10;            // numerical guard for |r|, |z|
+
+  /// Values of various coefficients
+  void initializeYlms()
+  {
+    double oneoversqrtpi = 1.0 / std::sqrt(o2::constants::math::PI);
+
+    // l=0 prefactors
+    fgPrefactors[0] = 0.5 * oneoversqrtpi;
+
+    // l=1 prefactors
+    fgPrefactors[1] = 0.5 * std::sqrt(3.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[2] = 0.5 * std::sqrt(3.0) * oneoversqrtpi;
+    fgPrefactors[3] = -fgPrefactors[1];
+
+    // l=2 prefactors
+    fgPrefactors[4] = 0.25 * std::sqrt(15.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[5] = 0.5 * std::sqrt(15.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[6] = 0.25 * std::sqrt(5.0) * oneoversqrtpi;
+    fgPrefactors[7] = -fgPrefactors[5];
+    fgPrefactors[8] = fgPrefactors[4];
+
+    // l=3 prefactors
+    fgPrefactors[9] = 0.125 * std::sqrt(35.0) * oneoversqrtpi;
+    fgPrefactors[10] = 0.25 * std::sqrt(105.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[11] = 0.125 * std::sqrt(21.0) * oneoversqrtpi;
+    fgPrefactors[12] = 0.25 * std::sqrt(7.0) * oneoversqrtpi;
+    fgPrefactors[13] = -fgPrefactors[11];
+    fgPrefactors[14] = fgPrefactors[10];
+    fgPrefactors[15] = -fgPrefactors[9];
+
+    // l=4 prefactors
+    fgPrefactors[16] = 3.0 / 16.0 * std::sqrt(35.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[17] = 3.0 / 8.0 * std::sqrt(35.0) * oneoversqrtpi;
+    fgPrefactors[18] = 3.0 / 8.0 * std::sqrt(5.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[19] = 3.0 / 8.0 * std::sqrt(5.0) * oneoversqrtpi;
+    fgPrefactors[20] = 3.0 / 16.0 * oneoversqrtpi;
+    fgPrefactors[21] = -fgPrefactors[19];
+    fgPrefactors[22] = fgPrefactors[18];
+    fgPrefactors[23] = -fgPrefactors[17];
+    fgPrefactors[24] = fgPrefactors[16];
+
+    // l=5 prefactors
+    fgPrefactors[25] = 3.0 / 32.0 * std::sqrt(77.0) * oneoversqrtpi;
+    fgPrefactors[26] = 3.0 / 16.0 * std::sqrt(385.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[27] = 1.0 / 32.0 * std::sqrt(385.0) * oneoversqrtpi;
+    fgPrefactors[28] = 1.0 / 8.0 * std::sqrt(1155.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[29] = 1.0 / 16.0 * std::sqrt(165.0 / 2.0) * oneoversqrtpi;
+    fgPrefactors[30] = 1.0 / 16.0 * std::sqrt(11.0) * oneoversqrtpi;
+    fgPrefactors[31] = -fgPrefactors[29];
+    fgPrefactors[32] = fgPrefactors[28];
+    fgPrefactors[33] = -fgPrefactors[27];
+    fgPrefactors[34] = fgPrefactors[26];
+    fgPrefactors[35] = -fgPrefactors[25];
+
+    fgPrefshift[0] = 0;
+    fgPrefshift[1] = 2;
+    fgPrefshift[2] = 6;
+    fgPrefshift[3] = 12;
+    fgPrefshift[4] = 20;
+    fgPrefshift[5] = 30;
+
+    fgPlmshift[0] = 0;
+    fgPlmshift[1] = 2;
+    fgPlmshift[2] = 5;
+    fgPlmshift[3] = 9;
+    fgPlmshift[4] = 14;
+    fgPlmshift[5] = 20;
+  }
+
+  /// Function to calculate Legendre Polynomials
+  ///  \param lmax Maximum value of L component
+  ///  \param ctheta Value of theta
+  ///  \param lbuf values of coefficients
+  void legendreUpToYlm(int lmax, double ctheta, std::array<double, 36>& lbuf)
+  {
+    // Calculate a set of legendre polynomials up to a given l
+    // with spherical input
+    std::array<double, 6> sins{};
+    std::array<double, 6> coss{};
+    sins[0] = 0.0;
+    coss[0] = 1.0;
+    sins[1] = std::sqrt(1 - ctheta * ctheta);
+    coss[1] = ctheta;
+    for (int iter = 2; iter < TrigCacheSize; iter++) {
+      sins[iter] = sins[iter - 1] * sins[1];
+      coss[iter] = coss[iter - 1] * coss[1];
+    }
+
+    // Legendre polynomials l=0
+    lbuf[0] = 1.0;
+
+    // Legendre polynomials l=1
+    if (lmax > 0) {
+      lbuf[1] = sins[1];
+      lbuf[2] = coss[1];
+    }
+
+    // Legendre polynomials l=2
+    if (lmax > 1) {
+      lbuf[3] = sins[2];
+      lbuf[4] = sins[1] * coss[1];
+      lbuf[5] = 3 * coss[2] - 1;
+    }
+
+    // Legendre polynomials l=3
+    if (lmax > 2) { // o2-linter: disable=magic-number (l index, mirrors lmax>0/1 above)
+      lbuf[6] = sins[3];
+      lbuf[7] = sins[2] * coss[1];
+      lbuf[8] = (5 * coss[2] - 1) * sins[1];
+      lbuf[9] = 5 * coss[3] - 3 * coss[1];
+    }
+
+    // Legendre polynomials l=4
+    if (lmax > 3) { // o2-linter: disable=magic-number (l index)
+      lbuf[10] = sins[4];
+      lbuf[11] = sins[3] * coss[1];
+      lbuf[12] = (7 * coss[2] - 1) * sins[2];
+      lbuf[13] = (7 * coss[3] - 3 * coss[1]) * sins[1];
+      lbuf[14] = 35 * coss[4] - 30 * coss[2] + 3;
+    }
+
+    // Legendre polynomials l=5
+    if (lmax > 4) { // o2-linter: disable=magic-number (l index)
+      lbuf[15] = sins[5];
+      lbuf[16] = sins[4] * coss[1];
+      lbuf[17] = (9 * coss[2] - 1) * sins[3];
+      lbuf[18] = (3 * coss[3] - 1 * coss[1]) * sins[2];
+      lbuf[19] = (21 * coss[4] - 14 * coss[2] + 1) * sins[1];
+      lbuf[20] = 63 * coss[5] - 70 * coss[3] + 15 * coss[1];
+    }
+  }
+
+  /// Function to calculate a set of Ylms up to a given l with cartesian input
+  void doYlmUpToL(int lmax, double x, double y, double z, std::complex<double>* ylms)
+  {
+    double ctheta = 0.0;
+    double phi = 0.0;
+
+    double r = std::sqrt(x * x + y * y + z * z);
+    if (r < SmallLength || std::fabs(z) < SmallLength) {
+      ctheta = 0.0;
+    } else {
+      ctheta = z / r;
+    }
+    phi = std::atan2(y, x);
+    doYlmUpToL(lmax, ctheta, phi, ylms);
+  }
+
+  /// Function to calculate a set of Ylms up to a given l with spherical input
+  void doYlmUpToL(int lmax, double ctheta, double phi, std::complex<double>* ylms)
+  {
+    int lcur = 0;
+    double lpol = 0.0;
+
+    std::array<double, 6> coss{};
+    std::array<double, 6> sins{};
+
+    std::array<double, 36> lbuf{};
+    legendreUpToYlm(lmax, ctheta, lbuf);
+    initializeYlms();
+
+    for (int iter = 1; iter <= lmax; iter++) {
+      coss[iter - 1] = std::cos(iter * phi);
+      sins[iter - 1] = std::sin(iter * phi);
+    }
+
+    ylms[lcur++] = fgPrefactors[0] * lbuf[0] * std::complex<double>(1, 0);
+
+    for (int il = 1; il <= lmax; il++) {
+      // First im = 0
+      ylms[lcur + il] = fgPrefactors[fgPrefshift[il]] * lbuf[static_cast<int>(fgPlmshift[il])] * std::complex<double>(1.0, 0.0);
+      // Im != 0
+      for (int im = 1; im <= il; im++) {
+        lpol = lbuf[static_cast<int>(fgPlmshift[il]) - im];
+        ylms[lcur + il - im] = fgPrefactors[fgPrefshift[il] - im] * lpol * std::complex<double>(coss[im - 1], -sins[im - 1]);
+        ylms[lcur + il + im] = fgPrefactors[fgPrefshift[il] + im] * lpol * std::complex<double>(coss[im - 1], sins[im - 1]);
+      }
+      lcur += 2 * il + 1;
+    }
+  }
+
+ private:
+  std::array<float, 36> fgPrefactors{};
+  std::array<int, 10> fgPrefshift{};
+  std::array<int, 10> fgPlmshift{};
+};
+
+} // namespace o2::analysis::femto
+
+#endif // PWGCF_FEMTO_CORE_FEMTOSPHERHARMATH_H_