ngsolve 6.2.2501.post47.dev1__cp313-cp313-macosx_10_15_universal2.whl → 6.2.2501.post55.dev1__cp313-cp313-macosx_10_15_universal2.whl

netgen/include/mptools.hpp

@@ -0,0 +1,2187 @@
+#ifndef FILE_MPTOOLS
+#define FILE_MPTOOLS
+
+/*
+  tools for computing with spherical harmonics and multi-poles
+ */
+
+
+#include <bla.hpp>
+#include <coefficient.hpp>
+
+
+namespace ngfem
+{
+
+
+  
+  class SphericalHarmonics
+  {
+    int order;
+    Vector<Complex> coefs;
+
+  public:
+    SphericalHarmonics (int aorder)
+      : order(aorder), coefs(sqr(order+1)) { coefs=0.0; }
+
+    int Order() const { return order; }
+    FlatVector<Complex> Coefs() const { return coefs; }
+
+    Complex & Coef(int n, int m)
+    {
+      return coefs(n*(n+1) + m);  
+    }
+
+    auto CoefsN (int n) const
+    {
+      return FlatVector<Complex> (2*n+1, &coefs(n*(n+1)-n)); 
+    }
+
+    auto Polar (Vec<3> x) const
+    {
+      double phi, theta;
+      if (x(0) == 0 && x(1) == 0)
+        {
+          phi = 0;
+          theta = x(2) > 0 ? 0 : M_PI;
+        }
+      else
+        {
+          phi = atan2(x(1), x(0));
+          theta = acos(x(2)/L2Norm(x));
+        }
+      return tuple{theta, phi};
+    }
+    
+    Complex Eval (Vec<3> x) const
+    {
+      auto [theta, phi] = Polar(x);
+      return Eval(theta, phi);
+    }
+  
+    Complex Eval (double theta, double phi) const
+    {
+      static Timer t("mptool sh evaluate"); RegionTimer rg(t);
+      
+      Matrix legfunc(order+1, order+1);
+      NormalizedLegendreFunctions (order, order, cos(theta), legfunc);
+      Vector<Complex> exp_imphi(order+1);
+      Complex exp_iphi(cos(phi), sin(phi));
+      Complex prod = 1.0;
+      for (int i = 0; i <= order; i++)
+        {
+          exp_imphi(i) = prod;
+          prod *= -exp_iphi;
+        }
+
+      Complex sum = 0.0;
+      int ii = 0;
+      for (int n = 0; n <= order; n++)
+        {
+          for (int m = -n; m < 0; m++, ii++)
+            sum += coefs(ii) * conj(exp_imphi(-m)) * legfunc(-m, n);
+          for (int m = 0; m <= n; m++, ii++)
+            sum += coefs(ii) * exp_imphi(m) * legfunc(m, n);
+        }
+
+      sum /= sqrt(4*M_PI);
+      return sum;
+    }
+
+
+    
+    Complex EvalOrder (int n, Vec<3> x) const
+    {
+      auto [theta, phi] = Polar (x);
+      return EvalOrder(n, theta, phi);
+    }
+  
+    Complex EvalOrder (int n, double theta, double phi) const
+    {
+      static Timer t("mptool sh evalorder");
+
+      RegionTimer rg(t);
+      
+      Matrix legfunc(order+1, order+1);
+      NormalizedLegendreFunctions (order, order, cos(theta), legfunc);
+      Vector<Complex> exp_imphi(order+1);
+      Complex exp_iphi(cos(phi), sin(phi));
+      Complex prod = 1.0;
+      for (int i = 0; i <= order; i++)
+        {
+          exp_imphi(i) = prod;
+          prod *= -exp_iphi;
+        }
+
+      Complex sum = 0.0;
+      auto coefsn = CoefsN(n);
+      int ii = 0;
+      // for (int n = 0; n <= order; n++)
+      {
+        for (int m = -n; m < 0; m++, ii++)
+          sum += coefsn(ii) * conj(exp_imphi(-m)) * legfunc(-m, n);
+        for (int m = 0; m <= n; m++, ii++)
+          sum += coefsn(ii) * exp_imphi(m) * legfunc(m, n);
+      }
+
+      sum /= sqrt(4*M_PI);
+      return sum;
+    }
+
+
+
+    void EvalOrders (Vec<3> x, FlatVector<Complex> vals) const
+    {
+      auto [theta, phi] = Polar(x);
+      return EvalOrders(theta, phi, vals);
+    }
+  
+    void EvalOrders (double theta, double phi, FlatVector<Complex> vals) const
+    {
+      static Timer ts("mptool sh evalorders small");
+      static Timer tl("mptool sh evalorders large");
+      // RegionTimer rg(order < 30 ? ts : tl);
+      // if (order > 30) tl.Start();
+      
+      Matrix legfunc(order+1, order+1);
+      NormalizedLegendreFunctions (order, order, cos(theta), legfunc);
+      Vector<Complex> exp_imphi(order+1);
+      Complex exp_iphi(cos(phi), sin(phi));
+      Complex prod = 1.0;
+      for (int i = 0; i <= order; i++)
+        {
+          exp_imphi(i) = prod;
+          prod *= -exp_iphi;
+        }
+
+
+      for (int n = 0; n <= order; n++)
+        {
+          auto coefsn = CoefsN(n);
+          int ii = 0;
+          
+          Complex sum = 0.0;
+          for (int m = -n; m < 0; m++, ii++)
+            sum += coefsn(ii) * conj(exp_imphi(-m)) * legfunc(-m, n);
+          for (int m = 0; m <= n; m++, ii++)
+            sum += coefsn(ii) * exp_imphi(m) * legfunc(m, n);
+          vals(n) = sum;
+        }
+
+      vals /= sqrt(4*M_PI);
+
+      // if (order > 30) tl.Stop();      
+    }
+
+
+    void Calc (Vec<3> x, FlatVector<Complex> shapes)
+    {
+      auto [theta, phi] = Polar(x);
+
+      Matrix legfunc(order+1, order+1);
+      NormalizedLegendreFunctions (order, order, cos(theta), legfunc);
+      Vector<Complex> exp_imphi(order+1);
+      Complex exp_iphi(cos(phi), sin(phi));
+      Complex prod = 1.0;
+      for (int i = 0; i <= order; i++)
+        {
+          exp_imphi(i) = prod;
+          prod *= -exp_iphi;
+        }
+
+      int ii = 0;
+      for (int n = 0; n <= order; n++)
+        {
+          for (int m = -n; m < 0; m++, ii++)
+            shapes(ii) = conj(exp_imphi(-m)) * legfunc(-m, n);
+          for (int m = 0; m <= n; m++, ii++)
+            shapes(ii) = exp_imphi(m) * legfunc(m, n);
+        }
+
+      shapes /= sqrt(4*M_PI);
+    }
+
+    
+    void RotateZ (double alpha)
+    {
+      // static Timer t("mptool sh RotateZ"); RegionTimer rg(t);
+      if (order < 0) return;
+
+      Vector<Complex> exp_imalpha(order+1);
+      Complex exp_ialpha(cos(alpha), sin(alpha));
+      Complex prod = 1.0;
+      for (int i = 0; i <= order; i++)
+        {
+          exp_imalpha(i) = prod;
+          prod *= exp_ialpha;
+        }
+
+      int ii = 0;
+      for (int n = 0; n <= order; n++)
+        {
+          for (int m = -n; m < 0; m++, ii++)
+            coefs(ii) *= conj(exp_imalpha(-m));
+          for (int m = 0; m <= n; m++, ii++)
+            coefs(ii) *= exp_imalpha(m);
+        }
+    }
+
+
+
+
+    static double CalcAmn (int m, int n)
+    {
+      if (m < 0) m=-m;
+      if (n < m) return 0;
+      return sqrt( (n+1.0+m)*(n+1.0-m) / ( (2*n+1)*(2*n+3) ));
+    }
+  
+    static double CalcBmn (int m, int n)
+    {
+      double sgn = (m >= 0) ? 1 : -1;
+      if ( (m > n) || (-m > n) ) return 0;
+      return sgn * sqrt( (n-m-1.0)*(n-m) / ( (2*n-1.0)*(2*n+1))); 
+    }
+  
+    static double CalcDmn (int m, int n)
+    {
+      double sgn = (m >= 0) ? 1 : -1;
+      return sgn/2 * sqrt((n-m)*(n+m+1));
+    }
+    
+    // Nail A. Gumerov and Ramani Duraiswami book, formula (2.2.12)
+    // add directional derivative divided by kappa to res, both multipoles need same scaling
+    void DirectionalDiffAdd (Vec<3> d, SphericalHarmonics & res, double scale = 1)
+    {
+      double fx = d(0);
+      double fy = d(1);
+      double fz = d(2);
+      double invscale = 1./scale;
+      
+      for (int n = 0; n < order; n++)
+        for (int m = -n; m <= n; m++)
+          {
+            double amn = CalcAmn(m,n);
+            double bmn1 = CalcBmn(-m-1, n+1);
+            double bmn2 = CalcBmn(m-1,n+1);
+            
+            res.Coef(n+1, m-1) += invscale * Complex(0.5*fx,0.5*fy)*bmn2 * Coef(n,m);
+            res.Coef(n+1, m  ) -= invscale * fz*amn * Coef(n,m);                
+            res.Coef(n+1, m+1) += invscale * Complex(0.5*fx,-0.5*fy)*bmn1 * Coef(n,m);
+            
+            res.Coef(n, m) += scale * Complex(-0.5*fx,0.5*fy)*bmn2 * Coef(n+1,m-1);
+            res.Coef(n, m) += scale * fz*amn * Coef(n+1,m);
+            res.Coef(n, m) += scale * Complex(-0.5*fx,-0.5*fy)*bmn1 * Coef(n+1,m+1);
+          }
+    }
+    
+    void RotateY (double alpha)
+    {
+      LocalHeap lh(8*6*sqr(order) + 8*15*order + 500);
+      
+      static Timer t("mptool sh RotateY"); RegionTimer rg(t);
+      /*
+      static std::map<int, unique_ptr<Timer<>>> timers;
+      static std::map<int, unique_ptr<Timer<>>> timerstrafo;      
+      if (timers.find(order)==timers.end())
+        {
+          timers[order] = make_unique<Timer<>> (string("mptools sh RotateY ")+ToString(order));
+          timerstrafo[order] = make_unique<Timer<>> (string("mptools sh RotateY trafo ")+ToString(order));
+        }
+      RegionTimer rg(*timers[order]);
+      */
+      double s = sin(alpha);
+      double c = cos(alpha);
+
+      FlatMatrix<> normalized_leg_func(order+2, order+2, lh);
+      NormalizedLegendreFunctions(order+1, order+1, c, normalized_leg_func);
+
+      if (alpha < 0)
+        for (int i = 1; i <= order+1; i+=2)
+          normalized_leg_func.Row(i) *= -1;
+      
+      // cout << "leg = " << endl << normalized_leg_func << endl;
+      FlatVector<> Dmn(2*order+1, lh);
+
+      for (int n=1; n <= order; n++)
+        {
+          HeapReset hr(lh);
+          
+          FlatMatrix<double> trafo(n+1, 2*n+1, lh); 
+          /*
+            Recursive Computation of Spherical Harmonic Rotation Coefficients of Large Degree
+            Nail A. Gumerov and Ramani Duraiswami
+            within Excursions in Harmonic Analysis, Volume 3
+            
+            page 130 
+          */
+
+          // Step 2
+          // H(0,m)
+          trafo.Col(n) = 1.0/sqrt(2*n+1) * normalized_leg_func.Col(n).Range(n+1);
+          for (int m = 1; m <= n; m += 2) trafo(m,n) *= -1;
+          // Step 3
+          // H(1,m)
+          FlatVector<double> tmp = 1.0/sqrt(2*n+3) * normalized_leg_func.Col(n+1).Range(n+2) | lh;
+          for (int m = 1; m < tmp.Size(); m += 2) tmp(m) *= -1;
+          for (int m = 1; m <= n; m++)
+            trafo.Col(n+1)(m) = 1/CalcBmn(0,n+1) * (  CalcBmn(-m-1, n+1)*(1-c)/2 * tmp(m+1)
+                                                      - CalcBmn(m-1,n+1)*(1+c)/2 * tmp(m-1)
+                                                      - CalcAmn(m,n) * s*tmp(m));
+
+          // Step 4
+          // diamond - recursion
+          for (int mp = -n; mp <= n; mp++)
+            Dmn(order+mp) = CalcDmn(mp, n);
+
+          for (int mp = 1; mp < n; mp++)
+            {
+              double invDmn = 1.0 / Dmn(order+mp);
+              for (int m = mp; m < n; m++)
+                trafo(m, n+mp+1) = invDmn  * ( Dmn(order+mp-1) *trafo(m  ,n+mp-1)
+                                               -Dmn(order+m-1)*trafo(m-1,n+mp)
+                                               +Dmn(order+m)  *trafo(m+1,n+mp));
+              int m = n;
+              trafo(m, n+mp+1) = invDmn * ( Dmn(order+mp-1,n)*trafo(m  ,n+mp-1)
+                                            -Dmn(order+m-1,n)*trafo(m-1,n+mp));
+            }
+      
+          // Step 5
+          // diamond - recursion, negative      
+          for (int mp = 0; mp > -n; mp--)
+            {
+              double invDmn = 1.0 / Dmn(order+mp-1);              
+              for (int m = -mp+1; m < n; m++)
+                trafo(m, n+mp-1) = invDmn * (  Dmn(order+mp,n)*trafo(m  ,n+mp+1)
+                                               +Dmn(order+m-1,n)*trafo(m-1,n+mp)
+                                               -Dmn(order+m  ,n)*trafo(m+1,n+mp));
+              int m = n;
+              trafo(m, n+mp-1) = invDmn * (  Dmn(order+mp,n)*trafo(m  ,n+mp+1)
+                                             +Dmn(order+m-1,n)*trafo(m-1,n+mp));
+            }
+
+          // RegionTimer rgtrafo(*timerstrafo[order]);                    
+          // Step 6
+          // symmetries in m and mp
+          for (int m = 0; m <= n; m++)
+            {
+              auto dst = trafo.Row(m).Range(n+m, n+n+1);
+              auto src = trafo.Col(n+m).Range(m, n+1);
+              dst = src;
+            }
+          for (int m = 0; m <= n; m++)
+            {
+              auto dst = trafo.Row(m).Range(n-n, n-m+1).Reversed();
+              auto src = trafo.Col(n-m).Range(m, n+1);
+              dst = src;
+            }
+          /*
+          double errortho = L2Norm( Matrix(trafo*Trans(trafo) - Identity(n+1)));
+          if (errortho > 1e-10)
+            {
+              *testout << "n = " << n << " order = " << Order() << ", alpha = " << alpha << ", errortho = " << errortho << endl;
+              if (n < 10)
+                *testout << trafo*Trans(trafo) << endl;
+            }
+          */
+
+          FlatVector<Complex> cn = CoefsN(n);
+          FlatVector<Complex> old = cn | lh;
+
+          cn.Slice(0,1) = Trans(trafo) * old.Range(n, 2*n+1);
+          cn.Slice(0,1).Reversed() += Trans(trafo.Rows(1,n+1)) * old.Range(0,n).Reversed();
+
+          for (int m = 1; m <= n; m+=2)
+            {
+              cn(n+m) *= -1;
+              cn(n-m) *= -1;
+            }
+        }
+    }
+    
+  };
+
+
+  // https://fortran-lang.discourse.group/t/looking-for-spherical-bessel-and-hankel-functions-of-first-and-second-kind-and-arbitrary-order/2308/2
+
+  // adapted from fmm3d
+template <typename Tz> 
+void besseljs3d (int nterms, Tz z, double scale,
+                 FlatVector<Tz> fjs, FlatVector<Tz> fjder)
+{
+  /*
+c**********************************************************************
+c
+c PURPOSE:
+c
+c	This subroutine evaluates the first NTERMS spherical Bessel 
+c	functions and if required, their derivatives.
+c	It incorporates a scaling parameter SCALE so that
+c       
+c		fjs_n(z)=j_n(z)/SCALE^n
+c		fjder_n(z)=\frac{\partial fjs_n(z)}{\partial z}
+c
+c INPUT:
+c
+c    nterms (integer): order of expansion of output array fjs 
+c    z     (complex *16): argument of the spherical Bessel functions
+c    scale    (real *8) : scaling factor (discussed above)
+c    ifder  (integer): flag indicating whether to calculate "fjder"
+c		          0	NO
+c		          1	YES
+c OUTPUT:
+c    fjs   (complex *16): array of scaled Bessel functions.
+c    fjder (complex *16): array of derivs of scaled Bessel functions.
+c
+c
+  */
+
+  
+  // c ... Initializing ...
+
+  // set to asymptotic values if argument is sufficiently small
+  if (abs(z) < 1e-200)
+    {
+      fjs(0) = 1;
+      for (int i = 1; i <= nterms; i++)
+        fjs(i) = 0.0;
+
+      if (fjder.Size())
+        {
+          fjder = 0.0;
+          fjder(1) = 1.0/(3*scale);
+        }
+      return;
+    }
+
+
+  //  ... Step 1: recursion up to find ntop, starting from nterms
+
+  Tz zinv=1.0/z;
+
+  Tz fjm1 = 0.0;
+  Tz fj0 = 1.0;
+
+  /*
+    c
+    cc     note max point for upward recurrence is
+    c      hard coded to nterms + 1000,
+    c      this might cause loss of accuracy for some
+    c      arguments in the complex plane for large 
+    c      nterms. For example, it is a terrible idea
+    c      to use this code for z>>nterms^2
+  */
+  
+  // int lwfjs = nterms + 100000;
+  int ntop = nterms+1000;
+
+  for (int i = nterms; ; i++)
+    {
+      double dcoef = 2*i+1.0;
+      Tz fj1 = dcoef*zinv*fj0-fjm1;
+      double dd = sqr(abs(fj1));
+      if (dd > 1e40)
+        {
+          ntop=i+1;
+          break;
+        }
+      fjm1 = fj0;
+      fj0 = fj1;
+      if (i > nterms+100000)
+        throw Exception("bessel failed 1");
+    }
+
+  Array<bool> iscale(ntop+1);
+  Vector<Tz> fjtmp(ntop+1);
+
+  /*
+    c ... Step 2: Recursion back down to generate the unscaled jfuns:
+    c             if magnitude exceeds UPBOUND2, rescale and continue the 
+    c	      recursion (saving the order at which rescaling occurred 
+    c	      in array iscale.
+  */
+
+  iscale = false;
+
+  fjtmp(ntop) = 0.0;
+  fjtmp(ntop-1) = 1.0;
+  for (int i = ntop-1; i>=1; i--)
+    {
+      double dcoef = 2*i+1.0;
+      fjtmp(i-1) = dcoef*zinv*fjtmp(i)-fjtmp(i+1);
+      double dd = sqr(abs(fjtmp(i-1)));
+      if (dd > 1e40)
+        {
+          fjtmp(i) *= 1e-40;
+          fjtmp(i-1) *= 1e-40;
+          iscale[i] = true;
+        }
+    }
+
+  /*
+    c
+    c ...  Step 3: go back up to the top and make sure that all
+    c              Bessel functions are scaled by the same factor
+    c              (i.e. the net total of times rescaling was invoked
+    c              on the way down in the previous loop).
+    c              At the same time, add scaling to fjs array.
+    c
+  */
+  
+  double scalinv = 1.0/scale;
+  double sctot = 1.0;
+  for (int i = 1; i <= ntop; i++)
+    {
+      sctot *= scalinv;
+      if (iscale[i-1])
+        sctot *= 1e-40;
+      fjtmp(i) *= sctot;
+    }
+  
+  //  Determine the normalization parameter:
+  
+  fj0=sin(z)*zinv;
+  Tz fj1=fj0*zinv-cos(z)*zinv;
+  
+  double d0=abs(fj0);
+  double d1=abs(fj1);
+  Tz zscale;
+  if (d1 > d0) 
+    zscale=fj1/(fjtmp(1)*scale);
+  else
+    zscale=fj0/fjtmp(0);
+
+  // Scale the jfuns by zscale:
+      
+  Tz ztmp=zscale;
+  for (int i = 0; i <= nterms; i++)
+    fjs(i)=fjtmp(i)*ztmp;
+
+
+  // Finally, calculate the derivatives if desired:
+
+  if (fjder.Size())
+    {
+      fjder(0) = -fjs(1)*scale;
+      for (int i = 1; i <= nterms; i++)
+        {
+          double dc1=i/(2*i+1.0);
+          double dc2=1.0-dc1;
+          dc1=dc1*scalinv;
+          dc2=dc2*scale;
+          fjder(i)=(dc1*fjtmp(i-1)-dc2*fjtmp(i+1))*ztmp;
+        }
+    }
+}
+
+
+  
+
+  /*
+  // from A. Barnett 
+  // http://www.fresco.org.uk/functions/barnett/index.htm
+
+  spherical bessel functions of first (the j_n) and second (the y_n) kind.
+
+  j0(r) = sin(r)/r
+  j1(r) = (sin(r)-r cos(r)) / r**2
+
+  y0(r) = -cos(r)/r
+  y1(r) = (-cos(r)-r*sin(r)) / r**2
+  */
+  void SBESJY (double x, int lmax,
+               FlatVector<double> j,
+               FlatVector<double> y,
+               FlatVector<double> jp,
+               FlatVector<double> yp)
+  {
+    if (x < 1e-8)
+      {
+        cout << "TODO: special treatment for small x" << endl;
+        return;
+      }
+
+    double xinv = 1/x;
+
+    if (lmax > 0)
+      {
+        double twoxi = 2*xinv;
+        double sl = lmax*xinv;
+        double tk = 2*sl+3*xinv;
+        double cf1 = sl;
+        double den = 1;
+        if (abs(cf1) < 1e-8) cf1 = 1e-8;
+        double c = cf1;
+        double d = 0;
+        for (int l = 1; l < 10000; l++)
+          {
+            c = tk-1/c;
+            d = tk-d;
+            if (abs(c) < 1e-8) c = 1e-8;
+            if (abs(d) < 1e-8) d = 1e-8;
+            d = 1/d;
+            double dcf1 = d*c;
+            cf1 *= dcf1;
+            if (d < 0) den = -den;
+            if (abs(dcf1-1) < 1e-10)
+              break;
+            tk += twoxi;
+            // nfp = l;
+          }
+
+        j(lmax) = den;
+        jp(lmax) = cf1*den;
+        for (int l = lmax; l >= 1; l--)
+          {
+            j(l-1) = (sl+xinv)*j(l) + jp(l);
+            sl = sl-xinv;
+            jp(l-1) = sl*j(l-1) - j(l);
+          }
+      }
+
+    double j0 = j(0);
+    double jp0 = jp(0);
+
+    // C------ CALCULATE THE L=0 SPHERICAL BESSEL FUNCTIONS DIRECTLY
+    j(0) = xinv * sin(x);
+    y(0) = -xinv * cos(x);
+    jp(0) = -y(0)-xinv*j(0);
+    yp(0) = j(0)-xinv*y(0);
+
+    double omega = (abs(j0)>abs(jp0)) ? j(0)/j0 : jp(0) / jp0;    // fix for x \approx 2 pi
+    double sl = 0;
+    for (int l = 1; l <=lmax; l++)
+      {
+        j(l) *= omega;
+        jp(l) *= omega;
+        y(l) = sl*y(l-1) - yp(l-1);
+        sl += xinv;
+        yp(l) = y(l-1) - (sl+xinv)*y(l);
+      }
+  }
+
+
+  
+  template <typename T>
+  void SphericalBessel (int n, double rho, double scale, T && values)
+  {
+    Vector<double> j(n+1), jp(n+1);
+    besseljs3d<double> (n, rho, scale,  j, jp);
+    values = j;
+  }
+
+
+  template <typename T>
+  void SphericalHankel1 (int n, double rho, double scale, T && values)
+  {
+    // Complex imag(0,1);
+    /*
+    if (n >= 0)
+      values(0) = exp(imag*rho) / (imag*rho);
+    if (n >= 1)
+      values(1) = -imag*values(0)*(1.0-1.0/(imag*rho));
+    
+    for (int i = 2; i <= n; i++)
+      values(i) = (2*i-1)/rho * values(i-1) - values(i-2);
+    */
+    
+    if (rho < 1e-8)
+      {
+        values = Complex(0);
+        return;
+      }
+    Vector j(n+1), y(n+1), jp(n+1), yp(n+1);
+    SBESJY (rho, n, j, y, jp, yp);
+
+    values = j + Complex(0,1) * y;
+    if (scale != 1.0)
+      {
+        double prod = 1.0;
+        for (int i = 0; i <= n; i++)
+          {
+            values(i) *= prod;
+            prod *= scale;
+          }
+      }
+  }
+
+
+
+
+  
+  // hn1 = jn+ i*yn
+  class MPSingular
+  {
+  public:
+    template <typename T>
+    static void Eval (int order, double r, double scale, T && values)
+    {
+      SphericalHankel1(order, r, scale,  values);
+    }
+  };
+  
+  // jn
+  class MPRegular
+  {
+  public:    
+    template <typename T>
+    static void Eval (int order, double r, double scale, T && values)
+    {
+      SphericalBessel (order, r, 1.0/scale, values);
+    }
+  };
+  
+  
+
+
+  template <typename RADIAL>
+  class MultiPole
+  {
+    SphericalHarmonics sh;
+    double kappa;
+    double scale;
+  public:
+    MultiPole (int aorder, double akappa, double ascale = 1) 
+      : sh(aorder), kappa(akappa), scale(ascale) { }
+
+    Complex & Coef(int n, int m) { return sh.Coef(n,m); }
+    auto & SH() { return sh; }
+    const auto & SH() const { return sh; }
+    double Kappa() const { return kappa; }
+    double Scale() const { return scale; }    
+    int Order() const { return sh.Order(); }
+    
+    MultiPole<RADIAL> Truncate(int neworder) const
+    {
+      if (neworder > sh.Order()) neworder=sh.Order();
+      MultiPole nmp(neworder, kappa);
+      nmp.sh.Coefs() = sh.Coefs().Range(sqr(neworder+1));
+      return nmp;
+    }
+
+    MultiPole & operator+= (const MultiPole & mp2)
+    {
+      size_t commonsize = min(SH().Coefs().Size(), mp2.SH().Coefs().Size());
+      SH().Coefs().Range(commonsize) += mp2.SH().Coefs().Range(commonsize);
+      return *this;
+    }
+    
+    Complex Eval (Vec<3> x) const
+    {
+      if (sh.Order() < 0) return 0;
+
+      Vector<Complex> radial(sh.Order()+1);
+      Vector<Complex> shvals(sh.Order()+1);
+      
+      RADIAL::Eval(sh.Order(), kappa*L2Norm(x), scale, radial);
+      sh.EvalOrders (x, shvals);
+      
+      Complex sum = 0;
+      for (int i = 0; i <= sh.Order(); i++)
+        sum +=  shvals(i) * radial(i);
+
+      return sum;
+    }
+
+    void AddCharge (Vec<3> x, Complex c)
+    {      
+      if constexpr (!std::is_same<RADIAL,MPSingular>())
+        throw Exception("AddCharge assumes singular MP");
+      
+      // static Timer t("mptool AddCharge"); RegionTimer rg(t);      
+      
+      if (L2Norm(x) < 1e-10)
+        {
+          sh.Coef(0,0) += c * Complex(0,1)*kappa/sqrt(4*M_PI);
+          return;
+        }
+
+      // cout << "add charge, kappa rho = " << kappa*L2Norm(x) << ", order = " << sh.Order() << endl;
+      
+      Vector<Complex> radial(sh.Order()+1);
+      Vector<Complex> sh_shapes(sqr (sh.Order()+1));
+
+      RADIAL::Eval(sh.Order(), kappa*L2Norm(x), 1.0/scale, radial);
+      // cout << "radial = " << radial << endl;
+      sh.Calc(x, sh_shapes);
+
+      for (int i = 0; i <= sh.Order(); i++)
+        {
+          IntRange r(sqr(i), sqr(i+1));
+          sh.Coefs().Range(r) += c * Complex(0,1)*kappa * radial(i).real()*Conj(sh_shapes.Range(r));
+        }
+    }
+
+    
+    void AddDipole (Vec<3> x, Vec<3> d, Complex c)
+    {
+      // static Timer t("mptool AddDipole"); RegionTimer rg(t);      
+      /*
+      double eps = 1e-4;
+      AddCharge(x+eps*d, -c/(2*eps));
+      AddCharge(x-eps*d, c/(2*eps));
+      return;
+      */
+      
+      if constexpr (!std::is_same<RADIAL,MPSingular>())
+        throw Exception("AddCharge assumes singular MP");
+
+      /*
+      // book, formula (2.2.20)
+      // dipole in origin:
+      MultiPole<MPSingular> tmp(1, kappa);
+      tmp.SH().Coef(1,1)  += Complex(0,1)*sqr(kappa)*sh.CalcBmn(-1,1)/(2*sqrt(4*M_PI)) * d(0)*c;
+      tmp.SH().Coef(1,-1) += Complex(0,1)*sqr(kappa)*sh.CalcBmn(-1,1)/(2*sqrt(4*M_PI)) * d(0)*c;
+
+      tmp.SH().Coef(1,1)  += Complex(1,0)*sqr(kappa)*sh.CalcBmn(-1,1)/(2*sqrt(4*M_PI)) * d(1)*c;
+      tmp.SH().Coef(1,-1) -= Complex(1,0)*sqr(kappa)*sh.CalcBmn(-1,1)/(2*sqrt(4*M_PI)) * d(1)*c;
+      
+      tmp.SH().Coef(1,0) += -Complex(0,1)*kappa*kappa*sh.CalcAmn(0,0)/sqrt(4*M_PI) *d(2)*c;
+      tmp.TransformAdd (*this, -x);
+      */
+
+      MultiPole<MPSingular> tmp(Order(), kappa, Scale());
+      tmp.AddCharge(x, c);
+      tmp.SH().DirectionalDiffAdd (kappa*d,  this->SH(), Scale());
+    }
+
+
+    void ChangeScaleTo (double newscale)
+    {
+      double fac = scale/newscale;
+      double prod = 1;
+      for (int n = 0; n <= sh.Order(); n++, prod*= fac)
+        sh.CoefsN(n) *= prod;
+      scale = newscale;
+    }
+
+    template <typename TARGET>
+    void Transform (MultiPole<TARGET> & target, Vec<3> dist) const
+    {
+      if (target.SH().Order() < 0) return;
+      if (SH().Order() < 0)
+        {
+          target.SH().Coefs() = 0.0;
+          return;
+        }
+      
+      static Timer t("mptool Transform "+ToString(typeid(RADIAL).name())+ToString(typeid(TARGET).name()));      
+      RegionTimer reg(t);
+      
+      double len = L2Norm(dist);
+      double theta, phi;
+
+      if (len < 1e-30)
+        theta = 0;
+      else
+        theta = acos (dist(2) / len);
+
+      if (sqr(dist(0))+sqr(dist(1)) < 1e-30)
+        phi = 0;
+      else
+        phi = atan2(dist(1), dist(0));
+        
+      
+      MultiPole<RADIAL> tmp(*this);
+      tmp.SH().RotateZ(phi);
+      tmp.SH().RotateY(theta);
+
+      tmp.ShiftZ(-len, target);
+      
+      target.SH().RotateY(-theta);
+      target.SH().RotateZ(-phi);
+    }
+    
+    template <typename TARGET>
+    void TransformAdd (MultiPole<TARGET> & target, Vec<3> dist) const
+    {
+      if (SH().Order() < 0) return;
+      if (target.SH().Order() < 0) return;      
+      
+      MultiPole<TARGET> tmp{target};
+      Transform(tmp, dist);
+      target.SH().Coefs() += tmp.SH().Coefs();
+    }
+
+#ifdef VER1
+    template <typename TARGET>
+    void ShiftZ (double z, MultiPole<TARGET> & target)
+    {
+      static Timer t("mptool ShiftZ"+ToString(typeid(RADIAL).name())+ToString(typeid(TARGET).name()));
+      
+      RegionTimer rg(t);
+      
+      int os = sh.Order();
+      int ot = target.SH().Order();
+
+      if (os > 100 && ot > 100 && abs(z)*kappa > 0.3*min(os,ot) && is_same<RADIAL,TARGET>())
+        {
+          MultiPole<TARGET> tmp {target};
+          ShiftZ(z/2, tmp);
+          tmp.ShiftZ(z/2, target);
+          return;
+        }
+
+      
+      target.SH().Coefs()=0.0;
+
+      LocalHeap lh( 16*( (os+ot+1)*(os+1) + (os+1 + ot+1) ) + 8*2*(os+ot+1) + 500);
+
+      FlatMatrix<Complex> trafo(os+ot+1, os+1, lh); 
+      FlatVector<Complex> hv1(os+1, lh), hv2(ot+1, lh);
+      FlatVector<double> amn(os+ot+1, lh);
+      FlatVector<double> inv_amn(os+ot+1, lh);
+      
+      // trafo = Complex(0.0);
+
+
+      double tscale = target.Scale();
+      double inv_tscale = 1.0/tscale;
+
+      
+      // (185) from paper 'fast, exact, stable, Gumerov+Duraiswami
+      // RADIAL::Eval(os+ot, kappa*abs(z), trafo.Col(0));
+      if (typeid(RADIAL) == typeid(TARGET))
+        SphericalBessel (os+ot, kappa*abs(z), tscale, trafo.Col(0));
+      else
+        SphericalHankel1 (os+ot, kappa*abs(z), inv_tscale, trafo.Col(0));
+
+      /*
+      if (L2Norm(trafo.Col(0)) > 1e5 || std::isnan(L2Norm(trafo.Col(0))))
+        {
+          *testout << "large Hankel: " << L2Norm(trafo.Col(0)) << endl;
+          *testout << "kappa z = " << kappa*z << ", os = " << os << ", ot = " << ot << endl;
+        }
+      */
+      // if (L2Norm(trafo.Col(0)) > 1e5)
+      // throw Exception ("z-shift - coefs large");
+      
+      if (z < 0)
+        for (int l = 1; l < trafo.Height(); l+=2) trafo(l,0) *= -1;
+      
+      for (int l = 0; l <= os+ot; l++)
+        trafo(l,0) *= sqrt(2*l+1);
+
+      if (os > 0)
+        {
+          for (int l = 1; l < os+ot; l++)   
+            trafo(l,1) = -scale/sh.CalcAmn(0,0) * (sh.CalcAmn(0,l)*tscale*trafo(l+1,0)
+                                                   -sh.CalcAmn(0,l-1)*inv_tscale*trafo(l-1,0));
+          trafo(0,1) = -scale*tscale*trafo(1,0);
+        }
+      
+      for (int n = 1; n < os; n++)
+        {
+          for (int l = 1; l < os+ot-n; l++)
+            trafo(l,n+1) = -scale/sh.CalcAmn(0,n) * (sh.CalcAmn(0,l)*tscale*trafo(l+1,n)
+                                                     -sh.CalcAmn(0,l-1)*inv_tscale*trafo(l-1,n)
+                                                     -sh.CalcAmn(0,n-1)*scale*trafo(l,n-1));
+          trafo(0,n+1) = pow(-scale*tscale,n+1)*trafo(n+1,0);
+        }
+
+      cout << "m = " << 0 << endl
+           << trafo.Rows(0,ot+1) << endl;
+
+      
+      for (int n = 0; n <= os; n++)
+        hv1(n) = sh.Coef(n,0);
+      hv2 = trafo.Rows(ot+1) * hv1;
+      for (int n = 0; n <= ot; n++)
+        target.SH().Coef(n,0) = hv2(n);
+
+      
+      for (int m = 1; m <= min(os,ot); m++)
+        {
+          // fill recursive formula (187)
+          for (int l = m; l <= os+ot-m; l++)
+            trafo(l,m) = scale/sh.CalcBmn(-m, m) * (sh.CalcBmn(-m, l)*inv_tscale*trafo(l-1, m-1)
+                                                    -sh.CalcBmn(m-1,l+1)*tscale*trafo(l+1,m-1));
+
+
+          /*
+          cout << "m = " << m << endl;
+          cout << "             norm col0 = " << L2Norm(trafo.Col(m).Range(m,os+ot-m+1)) << endl;
+          */
+          
+          for (int l = m-1; l < os+ot-m; l++)
+            {
+              amn(l) = sh.CalcAmn(m,l);
+              inv_amn(l) = scale/amn(l);
+            }
+
+          double prod = 1; 
+          for (int n = m; n < os; n++)
+            {
+              for (int l = m+1; l < os+ot-n; l++)
+                trafo(l,n+1) = -inv_amn(n) * (amn(l)*tscale*trafo(l+1,n)
+                                              -amn(l-1)*inv_tscale*trafo(l-1,n)
+                                              -amn(n-1)*scale*trafo(l,n-1));
+
+              prod *= -scale*tscale;
+              trafo(m,n+1) = prod*trafo(n+1,m);              
+            }
+
+          /*
+          cout << "                            norm trafo = "
+               << L2Norm(trafo.Rows(m,ot+1).Cols(m,os+1)) 
+               << " ortho " << L2Norm( Trans(trafo.Rows(m,ot+1).Cols(m,os+1))*trafo.Rows(m,ot+1).Cols(m,os+1)
+                                       - Identity(os+1-m)) << endl;
+          */
+          /*
+          *testout << "norm trafo = " << L2Norm(trafo.Rows(m, ot+1).Cols(m,os+1)) << endl;
+          if ( L2Norm(trafo.Rows(m, ot+1).Cols(m,os+1)) > 1e30)
+            {
+              *testout << trafo.Rows(m, ot+1).Cols(m,os+1) << endl;
+              for (int i = m; i < os+1; i++)
+                {
+                  *testout << "norm col " << i << " = " << L2Norm(trafo.Col(i).Range(m,os+ot-i)) << endl;
+                  *testout << "col " << i << " = " << trafo.Col(i).Range(m,os+ot-i) << endl;
+                }
+              throw Exception("large mat");
+            }
+          */
+
+          cout << "m = " << m << endl
+               << trafo.Rows(m,ot+1).Cols(m,os+1) << endl;
+          
+          for (int n = m; n <= os; n++)
+            hv1(n) = sh.Coef(n,m);
+          hv2.Range(m,ot+1) = trafo.Rows(m,ot+1).Cols(m,os+1) * hv1.Range(m,os+1);
+          for (int n = m; n <= ot; n++)
+            target.SH().Coef(n,m) = hv2(n);
+          
+          for (int n = m; n <= os; n++)
+            hv1(n) = sh.Coef(n,-m);
+          hv2.Range(m,ot+1) = trafo.Rows(m,ot+1).Cols(m,os+1) * hv1.Range(m,os+1);
+          for (int n = m; n <= ot; n++)
+            target.SH().Coef(n,-m) = hv2(n);
+        }
+    }
+#endif
+
+
+#ifdef VER2
+    
+    template <typename TARGET>
+    void ShiftZ (double z, MultiPole<TARGET> & target)
+    {
+      static Timer t("mptool ShiftZ"+ToString(typeid(RADIAL).name())+ToString(typeid(TARGET).name()));
+      RegionTimer rg(t);
+      
+      int os = sh.Order();
+      int ot = target.SH().Order();
+      
+      target.SH().Coefs()=0.0;
+
+      LocalHeap lh( 16*( (os+ot+1)*(os+1) + (os+1 + ot+1) ) + 8*2*(os+ot+1) + 500);
+
+      FlatMatrix<Complex> trafo(os+ot+1, os+1, lh); 
+      FlatVector<Complex> hv1(os+1, lh), hv2(ot+1, lh);
+      FlatVector<double> amn(os+ot+1, lh);
+      FlatVector<double> inv_amn(os+ot+1, lh);
+      
+      // trafo = Complex(0.0);
+
+      double tscale = target.Scale();
+      double inv_tscale = 1.0/tscale;
+
+      
+      // (185) from paper 'fast, exact, stable, Gumerov+Duraiswami
+      // RADIAL::Eval(os+ot, kappa*abs(z), trafo.Col(0));
+      if (typeid(RADIAL) == typeid(TARGET))
+        SphericalBessel (os+ot, kappa*abs(z), tscale, trafo.Col(0));
+      else
+        SphericalHankel1 (os+ot, kappa*abs(z), inv_tscale, trafo.Col(0));
+      
+      if (z < 0)
+        for (int l = 1; l < trafo.Height(); l+=2) trafo(l,0) *= -1;
+      
+      // for (int l = 0; l <= os+ot; l++)
+      // trafo(l,0) *= sqrt(2*l+1);
+
+      if (os > 0)
+        {
+          for (int l = 1; l < os+ot; l++)
+            {
+              /*
+              trafo(l,1) = -scale/sh.CalcAmn(0,0) *
+                (sh.CalcAmn(0,l)*tscale*trafo(l+1,0)
+                 -sh.CalcAmn(0,l-1)*inv_tscale*trafo(l-1,0));
+              */
+
+              int m = 0, n = 0;
+              double fac = ((2*l+1.0)/(2*n+1.0));
+              trafo(l,n+1) = -scale/ ( sqrt((n+1+m)*(n+1-m)) * fac)   *
+                (sqrt( (l+1+m)*(l+1-m)) * tscale * trafo(l+1,n)
+                 -sqrt( (l+m)*(l-m) )  * inv_tscale * trafo(l-1,n)
+                 -sqrt( (n+m)*(n-m) ) * fac * scale * trafo(l,n-1));
+            }
+          trafo(0,1) = -scale*tscale*trafo(1,0);
+        }
+      
+      for (int n = 1; n < os; n++)
+        {
+          for (int l = 1; l < os+ot-n; l++)
+            {
+              /*
+                trafo(l,n+1) = -scale/sh.CalcAmn(0,n) * (sh.CalcAmn(0,l)*tscale*trafo(l+1,n)
+                -sh.CalcAmn(0,l-1)*inv_tscale*trafo(l-1,n)
+                -sh.CalcAmn(0,n-1)*scale*trafo(l,n-1));
+              */
+              
+              int m = 0;
+              double fac = ((2*l+1.0)/(2*n+1.0));
+              trafo(l,n+1) = -scale / ( sqrt((n+1+m)*(n+1-m)) * fac)   *
+                (sqrt( (l+1+m)*(l+1-m)) * tscale * trafo(l+1,n)
+                 -sqrt( (l+m)*(l-m) )  * inv_tscale * trafo(l-1,n)
+                 -sqrt( (n+m)*(n-m) ) * fac * scale * trafo(l,n-1));
+            }
+          trafo(0,n+1) = pow(-scale*tscale,n+1)*trafo(n+1,0);
+        }
+
+      
+      Matrix<Complex> scaledtrafo(os+ot+1, os+1);
+      for (int l = 0; l <= os+ot; l++)
+        for (int n = 0; n <= os; n++)
+          scaledtrafo(l,n) = trafo(l,n) * sqrt( (2*l+1)*(2*n+1) );
+
+      // cout << "m = " << 0 << endl
+      // << scaledtrafo.Rows(0,ot+1) << endl;
+      
+      for (int n = 0; n <= os; n++)
+        hv1(n) = sh.Coef(n,0) * sqrt(2*n+1);
+      hv2 = trafo.Rows(ot+1) * hv1;
+      for (int n = 0; n <= ot; n++)
+        target.SH().Coef(n,0) = hv2(n) * sqrt(2*n+1);
+
+      
+      for (int m = 1; m <= min(os,ot); m++)
+        {
+          // fill recursive formula (187)
+          for (int l = m; l <= os+ot-m; l++)
+            {
+              trafo(l,m) = scale/sh.CalcBmn(-m, m) *
+                (sh.CalcBmn(-m, l)*inv_tscale * trafo(l-1, m-1) * sqrt( (2*l-1)*(2*m-1) )
+                 -sh.CalcBmn(m-1,l+1)*tscale * trafo(l+1,m-1) * sqrt( (2*l+3)*(2*m-1)) );
+              trafo(l,m) /= sqrt( (2*l+1)*(2*m+1) );
+            }
+
+          cout << "m = " << m << endl;
+          cout << "             norm col0 = " << L2Norm(trafo.Col(m).Range(m,os+ot-m+1)) << endl;
+          
+          for (int l = m-1; l < os+ot-m; l++)
+            {
+              amn(l) = sh.CalcAmn(m,l);
+              inv_amn(l) = scale/amn(l);
+            }
+
+          double prod = 1; 
+          for (int n = m; n < os; n++)
+            {
+              for (int l = m+1; l < os+ot-n; l++)
+                {
+                  /*
+                    trafo(l,n+1) = -inv_amn(n) * (amn(l)*tscale*trafo(l+1,n)
+                    -amn(l-1)*inv_tscale*trafo(l-1,n)
+                    -amn(n-1)*scale*trafo(l,n-1));
+                  */
+                  
+                
+                  double fac = ((2*l+1.0)/(2*n+1.0));
+                  trafo(l,n+1) = -scale / ( sqrt((n+1+m)*(n+1-m)) * fac)   *
+                    (sqrt( (l+1+m)*(l+1-m)) * tscale * trafo(l+1,n)
+                     -sqrt( (l+m)*(l-m) )  * inv_tscale * trafo(l-1,n)
+                     -sqrt( (n+m)*(n-m) ) * fac * scale * trafo(l,n-1));
+                }
+              prod *= -scale*tscale;
+              trafo(m,n+1) = prod*trafo(n+1,m);              
+            }
+
+          double normleft = 0;
+          for (int l = m; l <= ot; l++)
+            for (int n = m; n <= min(l,os); n++)
+              normleft += sqr(abs(trafo(l,n)));
+          normleft = sqrt(normleft);
+          cout << "                            norm trafo = "
+               << L2Norm(trafo.Rows(m,ot+1).Cols(m,os+1)) << ",  normleft = " << normleft << endl;
+          
+          // << " ortho " << L2Norm( Trans(trafo.Rows(m,ot+1).Cols(m,os+1))*trafo.Rows(m,ot+1).Cols(m,os+1)
+          // - Identity(os+1-m)) << endl;
+          /*
+          *testout << "norm trafo = " << L2Norm(trafo.Rows(m, ot+1).Cols(m,os+1)) << endl;
+          if ( L2Norm(trafo.Rows(m, ot+1).Cols(m,os+1)) > 1e30)
+            {
+              *testout << trafo.Rows(m, ot+1).Cols(m,os+1) << endl;
+              for (int i = m; i < os+1; i++)
+                {
+                  *testout << "norm col " << i << " = " << L2Norm(trafo.Col(i).Range(m,os+ot-i)) << endl;
+                  *testout << "col " << i << " = " << trafo.Col(i).Range(m,os+ot-i) << endl;
+                }
+              throw Exception("large mat");
+            }
+          */
+
+
+          /*
+          Matrix<Complex> scaledtrafo(os+ot+1, os+1);
+          for (int l = 0; l <= os+ot; l++)
+            for (int n = 0; n <= os; n++)
+              scaledtrafo(l,n) = trafo(l,n) * sqrt( (2*l+1)*(2*n+1) );
+          
+          cout << "m = " << m << endl
+               << scaledtrafo.Rows(m,ot+1).Cols(m,os+1) << endl;
+          */
+          
+          for (int n = m; n <= os; n++)
+            hv1(n) = sh.Coef(n,m) * sqrt(2*n+1);
+          hv2.Range(m,ot+1) = trafo.Rows(m,ot+1).Cols(m,os+1) * hv1.Range(m,os+1);
+          for (int n = m; n <= ot; n++)
+            target.SH().Coef(n,m) = hv2(n)*sqrt(2*n+1);
+          
+          for (int n = m; n <= os; n++)
+            hv1(n) = sh.Coef(n,-m) * sqrt(2*n+1);
+          hv2.Range(m,ot+1) = trafo.Rows(m,ot+1).Cols(m,os+1) * hv1.Range(m,os+1);
+          for (int n = m; n <= ot; n++)
+            target.SH().Coef(n,-m) = hv2(n) * sqrt(2*n+1);
+        }
+    }
+#endif
+
+
+
+
+    template <typename TARGET>
+    void ShiftZ (double z, MultiPole<TARGET> & target)
+    {
+      static Timer t("mptool ShiftZ"+ToString(typeid(RADIAL).name())+ToString(typeid(TARGET).name()));
+      RegionTimer rg(t);
+      
+      int os = sh.Order();
+      int ot = target.SH().Order();
+      
+      target.SH().Coefs()=0.0;
+
+      LocalHeap lh( 32*( (os+ot+1)*(os+ot+1) + (os+1 + ot+1) ) + 8*3*(os+ot+1) + 500);
+
+      FlatMatrix<Complex> trafo(os+ot+1, max(os,ot)+1, lh);
+      FlatMatrix<Complex> oldtrafo(os+ot+1, max(os,ot)+1, lh);     
+      FlatVector<Complex> hv1(os+1, lh), hv2(ot+1, lh);
+      
+      // trafo = Complex(0.0);
+
+      double tscale = target.Scale();
+      double inv_tscale = 1.0/tscale;
+
+      FlatVector<double> amn(os+ot+1, lh);
+      FlatVector<double> inv_amn(os+ot+1, lh);
+
+      FlatVector<double> powscale(os+1, lh);
+      double prod = 1;
+      for (int i = 0; i <= os; i++)
+        {
+          powscale(i) = prod;
+          prod *= -scale*tscale;
+        }
+      
+      // (185) from paper 'fast, exact, stable, Gumerov+Duraiswami
+      // RADIAL::Eval(os+ot, kappa*abs(z), trafo.Col(0));
+      if (typeid(RADIAL) == typeid(TARGET))
+        SphericalBessel (os+ot, kappa*abs(z), tscale, trafo.Col(0));
+      else
+        SphericalHankel1 (os+ot, kappa*abs(z), inv_tscale, trafo.Col(0));
+      
+      if (z < 0)
+        for (int l = 1; l < trafo.Height(); l+=2) trafo(l,0) *= -1;
+
+      for (int l = 0; l <= os+ot; l++)
+        trafo(l,0) *= sqrt(2*l+1);
+      
+      // for (int l = 0; l <= os+ot; l++)
+      // trafo(l,0) *= sqrt(2*l+1);
+
+      if (os > 0)
+        {
+          for (int l = 1; l < os+ot; l++)   
+            trafo(l,1) = -scale/sh.CalcAmn(0,0) * (sh.CalcAmn(0,l)*tscale*trafo(l+1,0)
+                                                   -sh.CalcAmn(0,l-1)*inv_tscale*trafo(l-1,0));
+          trafo(0,1) = -scale*tscale*trafo(1,0);
+        }
+
+      for (int n = 1; n < trafo.Width()-1; n++)
+        {
+          for (int l = 1; l < os+ot-n; l++)
+            trafo(l,n+1) = -scale/sh.CalcAmn(0,n) * (sh.CalcAmn(0,l)*tscale*trafo(l+1,n)
+                                                     -sh.CalcAmn(0,l-1)*inv_tscale*trafo(l-1,n)
+                                                     -sh.CalcAmn(0,n-1)*scale*trafo(l,n-1));
+          trafo(0,n+1) = pow(-scale*tscale,n+1)*trafo(n+1,0);
+        }
+      
+      for (int n = 0; n <= os; n++)
+        hv1(n) = sh.Coef(n,0);
+      hv2 = trafo.Rows(ot+1).Cols(os+1) * hv1;
+      for (int n = 0; n <= ot; n++)
+        target.SH().Coef(n,0) = hv2(n);
+
+      
+      for (int m = 1; m <= min(os,ot); m++)
+        {
+
+
+          for (int l = m-1; l < os+ot-m; l++)
+            {
+              amn(l) = sh.CalcAmn(m,l);
+              inv_amn(l) = scale/amn(l);
+            }
+
+          if (typeid(RADIAL) != typeid(TARGET))
+            
+            {
+              // fill recursive formula (187)
+              for (int l = m; l <= os+ot-m; l++)
+                trafo(l,m) = scale/sh.CalcBmn(-m, m) * (sh.CalcBmn(-m, l)*inv_tscale*trafo(l-1, m-1)  
+                                                        -sh.CalcBmn(m-1,l+1)*tscale*trafo(l+1,m-1));  
+              
+              for (int n = m; n < os; n++)
+                {
+                  for (int l = n+1; l < os+ot-n; l++)
+                    trafo(l,n+1) = -inv_amn(n) * (amn(l)*tscale*trafo(l+1,n)
+                                                  -amn(l-1)*inv_tscale*trafo(l-1,n)
+                                                  -amn(n-1)*scale*trafo(l,n-1));
+                }
+            }
+
+          else
+
+            {
+              /*
+              for (int n = m; n < trafo.Width()-1; n++)
+                {
+                  for (int l = n+1; l < os+ot-n; l++)
+                    trafo(l,n+1) = -inv_amn(n) * (amn(l)*tscale*trafo(l+1,n)
+                                                  -amn(l-1)*inv_tscale*trafo(l-1,n)
+                                                  -amn(n-1)*scale*trafo(l,n-1));
+                }
+              */
+
+              trafo.Swap (oldtrafo);
+              trafo = 0.0;
+              
+              // fill recursive formula (187)
+              for (int l = m; l <= os+ot-m; l++)
+                trafo(l,m) = scale/sh.CalcBmn(-m, m) * (sh.CalcBmn(-m, l)*inv_tscale*oldtrafo(l-1, m-1)  
+                                                        -sh.CalcBmn(m-1,l+1)*tscale*oldtrafo(l+1,m-1));  
+              
+              for (int n = m; n < trafo.Width()-1; n++)
+                {
+                  // int l = 2*order-n-1;
+                  int l = trafo.Height()-n-2;
+                  
+                  trafo(l,n+1) = scale/sh.CalcBmn(-m,n+1)* (sh.CalcBmn(m-1,n) * scale*trafo(l,n-1)
+                                                            - sh.CalcBmn(m-1,l+1)*tscale*oldtrafo(l+1,n)     
+                                                            + sh.CalcBmn(-m,l)  * 1/tscale*oldtrafo(l-1,n) );
+                  
+                  trafo(l-1,n) = tscale/amn(l-1) * (amn(l)  *   tscale*trafo(l+1,n)
+                                                    - amn(n-1)* scale*trafo(l,n-1)
+                                                    + amn(n)* 1/scale*trafo(l,n+1));
+                }
+              
+              // the same thing 1 row up
+              for (int n = m; n < trafo.Width()-2; n++)
+                {
+                  // int l = 2*order-n-2;
+                  int l = trafo.Height()-n-3;
+                  
+                  trafo(l,n+1) = scale/sh.CalcBmn(-m,n+1)* (sh.CalcBmn(m-1,n)     * scale*trafo(l,n-1)
+                                                            - sh.CalcBmn(m-1,l+1) * tscale* oldtrafo(l+1,n)   
+                                                            + sh.CalcBmn(-m,l)    * 1/tscale* oldtrafo(l-1,n) ); 
+                  
+                  trafo(l-1,n) = tscale/amn(l-1) * (amn(l)   * tscale*trafo(l+1,n)
+                                                    -amn(n-1)* scale*trafo(l,n-1) +
+                                                    amn(n)   * 1/scale*trafo(l,n+1)) ;
+                }
+              
+              
+              // for (int l = 2*order; l >= m; l--)
+              //   for (int n = m+1; n < min(2*order-l,l); n++)
+              for (int l = trafo.Height()-1; l >= m; l--)
+                for (int n = m+1; n < min<int>(trafo.Height()-1-l,l); n++)
+                  {
+                    trafo(l-1,n) = tscale/amn(l-1)* ( amn(l)  * tscale*trafo(l+1,n)
+                                                      -amn(n-1)* scale*trafo(l,n-1)
+                                                      +amn(n)  * 1/scale*trafo(l,n+1)) ;
+                  }
+            }
+          
+          
+          /*
+            cout << "m = " << m << endl
+            << trafo << endl;
+          */
+          for (int n = m; n < os; n++)
+            for (int l = n+1; l <= os; l++)
+              trafo(n,l) = powscale(l-n) * trafo(l,n);              
+              // trafo(n,l) = pow(-scale*tscale, l-n) * trafo(l,n);
+
+          /*
+          cout << "                            norm trafo = "
+               << L2Norm(trafo.Rows(m,ot+1).Cols(m,os+1)) << endl;
+          */
+          
+          for (int n = m; n <= os; n++)
+            hv1(n) = sh.Coef(n,m);
+          hv2.Range(m,ot+1) = trafo.Rows(m,ot+1).Cols(m,os+1) * hv1.Range(m,os+1);
+          for (int n = m; n <= ot; n++)
+            target.SH().Coef(n,m) = hv2(n);
+          
+          for (int n = m; n <= os; n++)
+            hv1(n) = sh.Coef(n,-m);
+          hv2.Range(m,ot+1) = trafo.Rows(m,ot+1).Cols(m,os+1) * hv1.Range(m,os+1);
+          for (int n = m; n <= ot; n++)
+            target.SH().Coef(n,-m) = hv2(n);
+        }
+    }
+    
+
+    
+
+
+    
+  };
+  
+  
+
+  // ***************** parameters ****************
+
+  static int MPOrder (double rho_kappa)
+  {
+    return max (20, int(2*rho_kappa));
+  }
+  static constexpr int maxdirect = 100;
+
+  class SingularMLMultiPole
+  {
+    static Array<size_t> nodes_on_level;    
+    
+    struct Node
+    {
+      Vec<3> center;
+      double r;
+      int level;
+      std::array<unique_ptr<Node>,8> childs;
+      MultiPole<MPSingular> mp;
+
+      Array<tuple<Vec<3>, Complex>> charges;
+      Array<tuple<Vec<3>, Vec<3>, Complex>> dipoles;
+      int total_sources;
+      
+      Node (Vec<3> acenter, double ar, int alevel, int order, double kappa)
+        : center(acenter), r(ar), level(alevel), mp(MPOrder(ar*kappa), kappa, min(1.0, 1*r*kappa))
+          // : center(acenter), r(ar), level(alevel), mp(MPOrder(ar*kappa), kappa, 1.0)
+      {
+        // cout << "singml, add node, level = " << level << endl;
+        if (level < nodes_on_level.Size())
+          nodes_on_level[level]++;
+      }
+
+
+      void CreateChilds()
+      {
+        if (childs[0]) throw Exception("have already childs");
+        for (int i = 0; i < 8; i++)
+          {
+            Vec<3> cc = center;
+            cc(0) += (i&1) ? r/2 : -r/2;
+            cc(1) += (i&2) ? r/2 : -r/2;
+            cc(2) += (i&4) ? r/2 : -r/2;
+            childs[i] = make_unique<Node> (cc, r/2, level+1, max(mp.SH().Order()/2, 8), mp.Kappa());
+          }
+      }
+      
+
+      void AddCharge (Vec<3> x, Complex c)
+      {
+        if (childs[0])
+          {
+            // directly send to childs:
+            int childnum  = 0;
+            if (x(0) > center(0)) childnum += 1;
+            if (x(1) > center(1)) childnum += 2;
+            if (x(2) > center(2)) childnum += 4;
+            childs[childnum] -> AddCharge(x, c);
+            return;
+          }
+
+        charges.Append( tuple{x,c} );
+
+        if (r*mp.Kappa() < 1e-8) return;
+        if (charges.Size() < maxdirect && r*mp.Kappa() < 1)
+          return;
+
+        CreateChilds();
+
+        for (auto [x,c] : charges)
+          AddCharge (x,c);
+        for (auto [x,d,c] : dipoles)
+          AddDipole (x,d,c);
+        
+        charges.SetSize0();
+        dipoles.SetSize0();        
+      }
+
+
+      void AddDipole (Vec<3> x, Vec<3> d, Complex c)
+      {
+        if (childs[0])
+          {
+            // directly send to childs:
+
+            int childnum  = 0;
+            if (x(0) > center(0)) childnum += 1;
+            if (x(1) > center(1)) childnum += 2;
+            if (x(2) > center(2)) childnum += 4;
+            childs[childnum] -> AddDipole(x, d, c);
+            return;
+          }
+
+        dipoles.Append (tuple{x,d,c});
+
+        if (dipoles.Size() < maxdirect || r < 1e-8)
+          return;
+        
+        CreateChilds();
+
+        for (auto [x,c] : charges)
+          AddCharge (x,c);
+        for (auto [x,d,c] : dipoles)
+          AddDipole (x,d,c);
+
+        charges.SetSize0();
+        dipoles.SetSize0();        
+      }
+
+      
+      Complex Evaluate(Vec<3> p) const
+      {
+        Complex sum = 0;
+        if (childs[0])
+          {
+            for (auto & child : childs)
+              sum += child->Evaluate(p);
+            return sum;
+          }
+
+        for (auto [x,c] : charges)
+          if (double rho = L2Norm(p-x); rho > 0)
+            sum += c*(1/(4*M_PI))*exp(Complex(0,rho*mp.Kappa())) / rho;
+        
+        for (auto [x,d,c] : dipoles)
+          if (double rho = L2Norm(p-x); rho > 0)
+            {
+              Vec<3> drhodp = 1.0/rho * (p-x);
+              Complex dGdrho = c*(1/(4*M_PI))*exp(Complex(0,rho*mp.Kappa())) *
+                (Complex(0, mp.Kappa())/rho - 1.0/sqr(rho));
+              sum += dGdrho * InnerProduct(drhodp, d);
+            }
+
+        return sum;
+      }
+
+      void CalcTotalSources()
+      {
+        total_sources = charges.Size() + dipoles.Size();
+        for (auto & child : childs)
+          if (child)
+            {
+              child->CalcTotalSources();
+              total_sources += child->total_sources;
+            }
+      }
+      
+      void CalcMP()
+      {
+        mp.SH().Coefs() = 0.0;
+        if (childs[0])
+          {
+            if (total_sources < 1000)
+              for (auto & child : childs)
+                child->CalcMP();
+            else
+              ParallelFor (8, [&] (int nr)
+                           {
+                             childs[nr] -> CalcMP();
+                           });
+
+            
+            for (auto & child : childs)            
+              child->mp.TransformAdd(mp, center-child->center);
+          }
+        else
+          {
+            if (charges.Size()+dipoles.Size() == 0)
+              {
+                mp = MultiPole<MPSingular> (-1, mp.Kappa());
+                return;
+              }
+
+            for (auto [x,c] : charges)
+              mp.AddCharge (x-center,c);
+            
+            for (auto [x,d,c] : dipoles)
+              mp.AddDipole (x-center, d, c);
+          }
+      }
+      
+      Complex EvaluateMP(Vec<3> p) const
+      {
+        if (charges.Size() || dipoles.Size())
+          return Evaluate(p);
+        
+        if (L2Norm(p-center) > 3*r)
+          return mp.Eval(p-center);
+        
+        if (!childs[0]) //  || level==1)
+          return Evaluate(p);
+          
+        Complex sum = 0.0;
+        for (auto & child : childs)
+          sum += child->EvaluateMP(p);
+        return sum;
+      }
+      
+      
+      void Print (ostream & ost) const
+      {
+        ost << "c = " << center << ", r = " << r << endl;
+        // for (int i = 0; i < loc_pnts.Size(); i++)
+        for (auto [x,c] : charges)
+          ost << "xi = " << x << ", ci = " << c << endl;
+
+        for (int i = 0; i < 8; i++)
+          if (childs[i]) childs[i] -> Print (ost);
+      }
+
+      double Norm () const
+      {
+        double norm = L2Norm(mp.SH().Coefs());
+        if (childs[0])
+          for (auto & ch : childs)
+            norm += ch->Norm();
+        return norm;
+      }
+      
+      size_t NumCoefficients() const
+      {
+        size_t num = sqr(mp.SH().Order()+1);
+        if (childs[0])
+          for (auto & ch : childs)
+            num += ch->NumCoefficients();
+        return num;
+      }
+    };
+    
+    Node root;
+    bool havemp = false;
+    
+  public:
+    SingularMLMultiPole (Vec<3> center, double r, int order, double kappa)
+      : root(center, r, 0, order, kappa)
+    {
+      nodes_on_level = 0;
+      nodes_on_level[0] = 1;
+    }
+
+    double Kappa() const { return root.mp.Kappa(); }
+    
+    void AddCharge(Vec<3> x, Complex c)
+    {
+      root.AddCharge(x, c);
+    }
+
+    void AddDipole(Vec<3> x, Vec<3> d, Complex c)
+    {
+      root.AddDipole(x, d, c);
+    }
+
+    void Print (ostream & ost) const
+    {
+      root.Print(ost);
+    }
+
+    double Norm() const
+    {
+      return root.Norm();
+    }
+
+    size_t NumCoefficients() const
+    {
+      return root.NumCoefficients();
+    }
+    
+    void CalcMP()
+    {
+      static Timer t("mptool compute singular MLMP"); RegionTimer rg(t);
+
+      /*
+      int maxlevel = 0;
+      for (auto [i,num] : Enumerate(nodes_on_level))
+        if (num > 0) maxlevel = i;
+
+      for (int i = 0; i <= maxlevel; i++)
+        cout << "sing " <<  i << ": " << nodes_on_level[i] << endl;
+      */
+      root.CalcTotalSources();
+      root.CalcMP();
+      
+      havemp = true;
+    }
+
+    Complex Evaluate (Vec<3> p) const
+    {
+      if (havemp)
+        return root.EvaluateMP(p);
+      else
+        return root.Evaluate(p);
+    }
+
+    friend class RegularMLMultiPole;
+  };
+
+
+  inline ostream & operator<< (ostream & ost, const SingularMLMultiPole & mlmp)
+  {
+    mlmp.Print(ost);
+    return ost;
+  }
+
+
+  class RegularMLMultiPole
+  {
+    static Array<size_t> nodes_on_level;
+    
+    struct Node
+    {
+      Vec<3> center;
+      double r;
+      int level;
+      std::array<unique_ptr<Node>,8> childs;
+      MultiPole<MPRegular> mp;
+      Array<Vec<3>> targets;
+      int total_targets;
+
+      Array<const SingularMLMultiPole::Node*> singnodes;
+
+      Node (Vec<3> acenter, double ar, int alevel, int order, double kappa)
+        : center(acenter), r(ar), level(alevel), mp(MPOrder(ar*kappa), kappa, 1.0/min(1.0, 0.25*r*kappa))
+          // : center(acenter), r(ar), level(alevel), mp(MPOrder(ar*kappa), kappa, 1.0)
+      {
+        if (level < nodes_on_level.Size())
+          nodes_on_level[level]++;
+      }
+
+
+      void CreateChilds()
+      {
+        if (childs[0]) throw Exception("have already childs");
+        // create children nodes:
+        for (int i = 0; i < 8; i++)
+          {
+            Vec<3> cc = center;
+            cc(0) += (i&1) ? r/2 : -r/2;
+            cc(1) += (i&2) ? r/2 : -r/2;
+            cc(2) += (i&4) ? r/2 : -r/2;
+            childs[i] = make_unique<Node> (cc, r/2, level+1, max(mp.SH().Order()/2, 8), mp.Kappa());
+          }
+      }
+
+      void AddSingularNode (const SingularMLMultiPole::Node & singnode, bool allow_refine)
+      {
+        if (mp.SH().Order() < 0) return;
+        if (singnode.mp.SH().Order() < 0) return;
+        if (L2Norm(singnode.mp.SH().Coefs()) == 0) return;
+        if (level > 20)
+          {
+            singnodes.Append(&singnode);            
+            return;
+          }
+        
+        // static Timer t("AddSingularNode"); RegionTimer reg(t);
+        
+        Vec<3> dist = center-singnode.center;
+
+        // if (L2Norm(dist)*mp.Kappa() > (mp.Order()+singnode.mp.Order()))
+        if (L2Norm(dist) > 2*(r + singnode.r))
+          {
+            if (singnode.mp.Order() > 2 * mp.Order() &&
+                singnode.childs[0] &&
+                singnode.childs[0]->mp.Order() < singnode.mp.Order())
+              {
+                for (auto & child : singnode.childs)
+                  AddSingularNode (*child, allow_refine);
+                return;
+              }
+
+            // static Timer t("mptool transform Helmholtz-criterion"); RegionTimer r(t);
+            singnode.mp.TransformAdd(mp, dist);
+            return;
+          }
+
+
+        if ( singnode.childs[0]==nullptr )
+          {
+            singnodes.Append(&singnode);
+            return;
+          }
+        
+        if (r > singnode.r)
+          {
+            if (allow_refine)
+              {
+                if (!childs[0])
+                  CreateChilds();
+                
+                for (auto & ch : childs)
+                  ch -> AddSingularNode (singnode, allow_refine);
+              }
+            else
+              {
+                if (total_targets < 1000)
+                  {
+                    for (auto & ch : childs)
+                      if (ch)
+                        ch -> AddSingularNode (singnode, allow_refine);
+                  }
+                else
+                  ParallelFor (8, [&] (int nr)
+                               {
+                                 if (childs[nr])
+                                   childs[nr] -> AddSingularNode (singnode, allow_refine);
+                               });
+                
+                if (targets.Size())
+                  singnodes.Append(&singnode);
+              }
+          }
+        else
+          {
+            for (auto & childsing : singnode.childs)
+              AddSingularNode (*childsing, allow_refine);
+          }
+      }
+
+      void LocalizeExpansion(bool allow_refine)
+      {
+        if (allow_refine)
+          if (mp.Order() > 20 && !childs[0])
+            CreateChilds();
+
+        if (childs[0])
+          {
+            for (auto & ch : childs)
+              {
+                if (L2Norm(mp.SH().Coefs()) > 0)
+                  mp.TransformAdd (ch->mp, ch->center-center);
+                ch->LocalizeExpansion(allow_refine);
+              }
+            mp = MultiPole<MPRegular>(-1, mp.Kappa());
+            //mp.SH().Coefs()=0.0;
+          }
+      }
+      
+      Complex Evaluate (Vec<3> p) const
+      {
+        // *testout << "eval p = " << p << ", level = " << level << ", center = " << center <<  ", r = " << r << endl;
+        Complex sum = 0.0;
+        /*
+        if (childs[0])
+          {
+            int childnum = 0;
+            if (p(0) > center(0)) childnum += 1;
+            if (p(1) > center(1)) childnum += 2;
+            if (p(2) > center(2)) childnum += 4;
+            sum = childs[childnum]->Evaluate(p);
+          }
+        */
+        int childnum = 0;
+        if (p(0) > center(0)) childnum += 1;
+        if (p(1) > center(1)) childnum += 2;
+        if (p(2) > center(2)) childnum += 4;
+        if (childs[childnum])
+          sum = childs[childnum]->Evaluate(p);
+        else
+          sum = mp.Eval(p-center);
+
+
+        static Timer t("mptool direct evaluate"); RegionTimer r(t);
+        for (auto sn : singnodes)
+          sum += sn->EvaluateMP(p);
+
+        return sum;
+      }
+
+      double Norm() const
+      {
+        double norm = L2Norm(mp.SH().Coefs());
+        if (childs[0])
+          for (auto & ch : childs)
+            norm += ch->Norm();
+        return norm;
+      }
+
+      size_t NumCoefficients() const
+      {
+        size_t num = sqr(mp.SH().Order()+1);
+        if (childs[0])
+          for (auto & ch : childs)
+            num += ch->NumCoefficients();
+        return num;
+      }
+      
+      void AddTarget (Vec<3> x)
+      {
+        if (childs[0])
+          {
+            // directly send to childs:
+            int childnum  = 0;
+            if (x(0) > center(0)) childnum += 1;
+            if (x(1) > center(1)) childnum += 2;
+            if (x(2) > center(2)) childnum += 4;
+            childs[childnum] -> AddTarget( x );
+            return;
+          }
+
+        targets.Append( x );
+
+        if (r*mp.Kappa() < 1e-8) return;
+        if (targets.Size() < maxdirect && r*mp.Kappa() < 1)
+          return;
+
+        CreateChilds();
+
+        for (auto t : targets)
+          AddTarget (t);
+        targets.SetSize0();
+      }
+
+      void CalcTotalTargets()
+      {
+        total_targets = targets.Size();
+        for (auto & child : childs)
+          if (child)
+            {
+              child->CalcTotalTargets();
+              total_targets += child->total_targets;
+            }
+      }
+
+      void RemoveEmptyTrees()
+      {
+        for (auto & child : childs)
+          if (child)
+            {
+              child->RemoveEmptyTrees();
+              // if (child->total_targets == 0)
+              // child = nullptr;
+            }
+
+        if (total_targets == 0)
+          mp = MultiPole<MPRegular>(-1, mp.Kappa());
+      }
+      
+    };
+    
+    Node root;
+    shared_ptr<SingularMLMultiPole> singmp;
+    
+  public:
+    RegularMLMultiPole (shared_ptr<SingularMLMultiPole> asingmp, Vec<3> center, double r, int order)
+      : root(center, r, 0, order, asingmp->Kappa()), singmp(asingmp)
+    {
+      if (!singmp->havemp) throw Exception("first call Calc for singular MP");
+
+      nodes_on_level = 0;
+      nodes_on_level[0] = 1;
+      {
+        static Timer t("mptool compute regular MLMP"); RegionTimer rg(t);            
+        root.AddSingularNode(singmp->root, true);
+        // cout << "norm after S->R conversion: " << root.Norm() << endl;
+      }
+
+
+      /*
+      int maxlevel = 0;
+      for (auto [i,num] : Enumerate(nodes_on_level))
+        if (num > 0) maxlevel = i;
+
+      for (int i = 0; i <= maxlevel; i++)
+        cout << "reg " << i << ": " << nodes_on_level[i] << endl;
+      */
+      
+      {
+        static Timer t("mptool expand regular MLMP"); RegionTimer rg(t);                  
+        root.LocalizeExpansion(true);
+        // cout << "norm after local expansion: " << root.Norm() << endl;        
+      }
+    }
+
+    RegularMLMultiPole (Vec<3> center, double r, int order, double kappa)
+      : root(center, r, 0, order, kappa)
+    {
+      nodes_on_level = 0;
+      nodes_on_level[0] = 1;
+    }
+    
+    void AddTarget (Vec<3> t)
+    {
+      root.AddTarget (t);
+    }
+
+    void CalcMP(shared_ptr<SingularMLMultiPole> asingmp)
+    {
+      singmp = asingmp;
+
+      root.CalcTotalTargets();
+      root.RemoveEmptyTrees();
+        
+      root.AddSingularNode(singmp->root, false);
+
+      /*
+      int maxlevel = 0;
+      for (auto [i,num] : Enumerate(RegularMLMultiPole::nodes_on_level))
+        if (num > 0) maxlevel = i;
+
+      for (int i = 0; i <= maxlevel; i++)
+        cout << "reg " << i << ": " << RegularMLMultiPole::nodes_on_level[i] << endl;
+      */
+
+      root.LocalizeExpansion(false);
+    }
+
+    double Norm() const
+    {
+      return root.Norm();
+    }
+
+    size_t NumCoefficients() const
+    {
+      return root.NumCoefficients();
+    }
+
+    Complex Evaluate (Vec<3> p) const
+    {
+      // static Timer t("mptool Eval MLMP regular"); RegionTimer r(t);
+      if (L2Norm(p-root.center) > root.r) return 0.0;
+      return root.Evaluate(p);
+    }
+
+    
+  };
+  
+  Array<size_t> RegularMLMultiPole::nodes_on_level(100);
+  Array<size_t> SingularMLMultiPole::nodes_on_level(100);
+
+  // ******************** Coefficient Functions *********************
+
+  class SphericalHarmonicsCF : public CoefficientFunction
+  {
+    SphericalHarmonics sh;
+  public:
+    SphericalHarmonicsCF (int order)
+      : CoefficientFunction(1, true), sh(order) { }
+    Complex & Coef(int n, int m) { return sh.Coef(n,m); } 
+    
+    virtual double Evaluate (const BaseMappedIntegrationPoint & ip) const override
+    { throw Exception("real eval not available"); }
+
+    virtual void Evaluate (const BaseMappedIntegrationPoint & mip, FlatVector<Complex> values) const override
+    {
+      values(0) = sh.Eval(mip.GetPoint());
+    }
+    
+    virtual void Evaluate (const BaseMappedIntegrationRule & ir, BareSliceMatrix<Complex> values) const override
+    {
+      for (int i = 0; i < ir.Size(); i++)
+        {
+          auto & mip = ir[i];
+          values(i,0) = sh.Eval(mip.GetPoint());
+        }
+    }
+
+    auto & SH() { return sh; }
+  };
+
+
+
+  template <typename RADIAL>
+  class MultiPoleCF : public CoefficientFunction
+  {
+    MultiPole<RADIAL> mp;
+    Vec<3> center;
+  public:
+    MultiPoleCF (int order, double kappa, Vec<3> acenter, double scale = 1)
+      : CoefficientFunction(1, true), mp(order, kappa, scale), center(acenter) { }
+
+    Complex & Coef(int n, int m) { return mp.Coef(n,m); } 
+    auto & SH() { return mp.SH(); }
+    auto & MP() { return mp; }
+    Vec<3> Center() const { return center; }
+    
+    virtual double Evaluate (const BaseMappedIntegrationPoint & ip) const override
+    { throw Exception("real eval not available"); }
+
+    virtual void Evaluate (const BaseMappedIntegrationPoint & mip, FlatVector<Complex> values) const override
+    {
+      values(0) = mp.Eval(mip.GetPoint()-center);
+    }
+
+    template <typename TARGET>
+    void ShiftZ (double z, MultiPole<TARGET> & target) { mp.ShiftZ(z, target); }
+
+    using CoefficientFunction::Transform;        
+    template <typename TARGET>
+    void Transform (MultiPoleCF<TARGET> & target)
+    {
+      mp.Transform (target.MP(), target.Center()-center);
+    }
+  };
+
+  
+  class SingularMLMultiPoleCF : public CoefficientFunction
+  {
+    shared_ptr<SingularMLMultiPole> mlmp;
+  public:
+    SingularMLMultiPoleCF (Vec<3> center, double r, int order, double kappa)
+      : CoefficientFunction(1, true), mlmp{make_shared<SingularMLMultiPole>(center, r, order, kappa)} { }
+    
+    virtual double Evaluate (const BaseMappedIntegrationPoint & ip) const override
+    { throw Exception("real eval not available"); }
+    
+    virtual void Evaluate (const BaseMappedIntegrationPoint & mip, FlatVector<Complex> values) const override
+    {
+      values(0) = mlmp->Evaluate(mip.GetPoint());
+    }
+    
+    shared_ptr<SingularMLMultiPole> MLMP() { return mlmp; }
+  };
+  
+
+  class RegularMLMultiPoleCF : public CoefficientFunction
+  {
+    RegularMLMultiPole mlmp;
+  public:
+    RegularMLMultiPoleCF (shared_ptr<SingularMLMultiPoleCF> asingmp, Vec<3> center, double r, int order)
+      : CoefficientFunction(1, true), mlmp(asingmp->MLMP(), center, r, order) { } 
+    
+    virtual double Evaluate (const BaseMappedIntegrationPoint & ip) const override
+    { throw Exception("real eval not available"); }
+
+    virtual void Evaluate (const BaseMappedIntegrationPoint & mip, FlatVector<Complex> values) const override
+    {
+      values(0) = mlmp.Evaluate(mip.GetPoint());
+    }
+
+    RegularMLMultiPole & MLMP() { return mlmp; }
+  };
+
+  
+}
+#endif