| 36 | 
  | 
 * [1]  Meineke, et al., J. Comp. Chem. 26, 252-271 (2005).              | 
| 37 | 
  | 
 * [2]  Fennell & Gezelter, J. Chem. Phys. 124, 234104 (2006).           | 
| 38 | 
  | 
 * [3]  Sun, Lin & Gezelter, J. Chem. Phys. 128, 24107 (2008).           | 
| 39 | 
< | 
 * [4]  Vardeman & Gezelter, in progress (2009).                         | 
| 39 | 
> | 
 * [4]  Kuang & Gezelter,  J. Chem. Phys. 133, 164101 (2010). | 
| 40 | 
> | 
 * [5]  Vardeman, Stocker & Gezelter, J. Chem. Theory Comput. 7, 834 (2011). | 
| 41 | 
  | 
 */ | 
| 42 | 
  | 
  | 
| 43 | 
  | 
#include "math/CubicSpline.hpp" | 
| 44 | 
  | 
#include "utils/simError.h" | 
| 45 | 
  | 
#include <cmath> | 
| 46 | 
+ | 
#include <cstdio> | 
| 47 | 
  | 
#include <algorithm> | 
| 46 | 
– | 
#include <iostream> | 
| 48 | 
  | 
 | 
| 49 | 
  | 
using namespace OpenMD; | 
| 50 | 
  | 
using namespace std; | 
| 51 | 
  | 
 | 
| 52 | 
< | 
CubicSpline::CubicSpline() : generated(false), isUniform(true) {} | 
| 52 | 
> | 
CubicSpline::CubicSpline() : generated(false), isUniform(true) { | 
| 53 | 
> | 
  data_.clear(); | 
| 54 | 
> | 
} | 
| 55 | 
  | 
 | 
| 56 | 
< | 
void CubicSpline::addPoint(RealType xp, RealType yp) { | 
| 57 | 
< | 
  data.push_back(make_pair(xp, yp)); | 
| 56 | 
> | 
void CubicSpline::addPoint(const RealType xp, const RealType yp) { | 
| 57 | 
> | 
  data_.push_back(make_pair(xp, yp)); | 
| 58 | 
  | 
} | 
| 59 | 
  | 
 | 
| 60 | 
  | 
void CubicSpline::addPoints(const vector<RealType>& xps,  | 
| 68 | 
  | 
    simError();     | 
| 69 | 
  | 
  } | 
| 70 | 
  | 
 | 
| 71 | 
< | 
  for (int i = 0; i < xps.size(); i++)  | 
| 72 | 
< | 
    data.push_back(make_pair(xps[i], yps[i])); | 
| 71 | 
> | 
  for (unsigned int i = 0; i < xps.size(); i++)  | 
| 72 | 
> | 
    data_.push_back(make_pair(xps[i], yps[i])); | 
| 73 | 
  | 
} | 
| 74 | 
  | 
 | 
| 75 | 
  | 
void CubicSpline::generate() {  | 
| 76 | 
  | 
  // Calculate coefficients defining a smooth cubic interpolatory spline. | 
| 77 | 
  | 
  // | 
| 78 | 
  | 
  // class values constructed: | 
| 79 | 
< | 
  //   n   = number of data points. | 
| 79 | 
> | 
  //   n   = number of data_ points. | 
| 80 | 
  | 
  //   x   = vector of independent variable values  | 
| 81 | 
  | 
  //   y   = vector of dependent variable values | 
| 82 | 
  | 
  //   b   = vector of S'(x[i]) values. | 
| 83 | 
  | 
  //   c   = vector of S"(x[i])/2 values. | 
| 84 | 
  | 
  //   d   = vector of S'''(x[i]+)/6 values (i < n). | 
| 85 | 
  | 
  // Local variables:    | 
| 86 | 
< | 
   | 
| 86 | 
> | 
  | 
| 87 | 
  | 
  RealType fp1, fpn, h, p; | 
| 88 | 
  | 
   | 
| 89 | 
  | 
  // make sure the sizes match | 
| 90 | 
  | 
   | 
| 91 | 
< | 
  n = data.size();   | 
| 89 | 
< | 
  x.resize(n); | 
| 90 | 
< | 
  y.resize(n); | 
| 91 | 
> | 
  n = data_.size();   | 
| 92 | 
  | 
  b.resize(n); | 
| 93 | 
  | 
  c.resize(n); | 
| 94 | 
  | 
  d.resize(n); | 
| 98 | 
  | 
  bool sorted = true; | 
| 99 | 
  | 
   | 
| 100 | 
  | 
  for (int i = 1; i < n; i++) { | 
| 101 | 
< | 
    if ( (data[i].first - data[i-1].first ) <= 0.0 ) sorted = false; | 
| 101 | 
> | 
    if ( (data_[i].first - data_[i-1].first ) <= 0.0 ) sorted = false; | 
| 102 | 
  | 
  } | 
| 103 | 
  | 
   | 
| 104 | 
  | 
  // sort if necessary | 
| 105 | 
  | 
   | 
| 106 | 
< | 
  if (!sorted) sort(data.begin(), data.end());   | 
| 106 | 
> | 
  if (!sorted) sort(data_.begin(), data_.end());   | 
| 107 | 
  | 
   | 
| 107 | 
– | 
  // Copy spline data out to separate arrays: | 
| 108 | 
– | 
   | 
| 109 | 
– | 
  for (int i = 0; i < n; i++) { | 
| 110 | 
– | 
    x[i] = data[i].first; | 
| 111 | 
– | 
    y[i] = data[i].second; | 
| 112 | 
– | 
  } | 
| 113 | 
– | 
   | 
| 108 | 
  | 
  // Calculate coefficients for the tridiagonal system: store | 
| 109 | 
  | 
  // sub-diagonal in B, diagonal in D, difference quotient in C.   | 
| 110 | 
  | 
   | 
| 111 | 
< | 
  b[0] = data[1].first - data[0].first; | 
| 112 | 
< | 
  c[0] = (data[1].second - data[0].second) / b[0]; | 
| 111 | 
> | 
  b[0] = data_[1].first - data_[0].first; | 
| 112 | 
> | 
  c[0] = (data_[1].second - data_[0].second) / b[0]; | 
| 113 | 
  | 
   | 
| 114 | 
  | 
  if (n == 2) { | 
| 115 | 
  | 
 | 
| 116 | 
  | 
    // Assume the derivatives at both endpoints are zero. Another | 
| 117 | 
  | 
    // assumption could be made to have a linear interpolant between | 
| 118 | 
  | 
    // the two points.  In that case, the b coefficients below would be | 
| 119 | 
< | 
    // (data[1].second - data[0].second) / (data[1].first - data[0].first) | 
| 119 | 
> | 
    // (data_[1].second - data_[0].second) / (data_[1].first - data_[0].first) | 
| 120 | 
  | 
    // and the c and d coefficients would both be zero. | 
| 121 | 
  | 
    b[0] = 0.0; | 
| 122 | 
< | 
    c[0] = -3.0 * pow((data[1].second - data[0].second) / | 
| 123 | 
< | 
                      (data[1].first-data[0].first), 2); | 
| 124 | 
< | 
    d[0] = -2.0 * pow((data[1].second - data[0].second) /  | 
| 125 | 
< | 
                      (data[1].first-data[0].first), 3); | 
| 122 | 
> | 
    c[0] = -3.0 * pow((data_[1].second - data_[0].second) / | 
| 123 | 
> | 
                      (data_[1].first-data_[0].first), 2); | 
| 124 | 
> | 
    d[0] = -2.0 * pow((data_[1].second - data_[0].second) /  | 
| 125 | 
> | 
                      (data_[1].first-data_[0].first), 3); | 
| 126 | 
  | 
    b[1] = b[0]; | 
| 127 | 
  | 
    c[1] = 0.0; | 
| 128 | 
  | 
    d[1] = 0.0; | 
| 129 | 
< | 
    dx = 1.0 / (data[1].first - data[0].first); | 
| 129 | 
> | 
    dx = 1.0 / (data_[1].first - data_[0].first); | 
| 130 | 
  | 
    isUniform = true; | 
| 131 | 
  | 
    generated = true; | 
| 132 | 
  | 
    return; | 
| 135 | 
  | 
  d[0] = 2.0 * b[0]; | 
| 136 | 
  | 
   | 
| 137 | 
  | 
  for (int i = 1; i < n-1; i++) { | 
| 138 | 
< | 
    b[i] = data[i+1].first - data[i].first; | 
| 138 | 
> | 
    b[i] = data_[i+1].first - data_[i].first; | 
| 139 | 
  | 
    if ( fabs( b[i] - b[0] ) / b[0] > 1.0e-5) isUniform = false; | 
| 140 | 
< | 
    c[i] = (data[i+1].second - data[i].second) / b[i]; | 
| 140 | 
> | 
    c[i] = (data_[i+1].second - data_[i].second) / b[i]; | 
| 141 | 
  | 
    d[i] = 2.0 * (b[i] + b[i-1]); | 
| 142 | 
  | 
  } | 
| 143 | 
  | 
   | 
| 144 | 
  | 
  d[n-1] = 2.0 * b[n-2]; | 
| 145 | 
  | 
   | 
| 146 | 
  | 
  // Calculate estimates for the end slopes using polynomials | 
| 147 | 
< | 
  // that interpolate the data nearest the end. | 
| 147 | 
> | 
  // that interpolate the data_ nearest the end. | 
| 148 | 
  | 
   | 
| 149 | 
  | 
  fp1 = c[0] - b[0]*(c[1] - c[0])/(b[0] + b[1]); | 
| 150 | 
  | 
  if (n > 3) fp1 = fp1 + b[0]*((b[0] + b[1]) * (c[2] - c[1]) /  | 
| 151 | 
  | 
                               (b[1] + b[2]) -  | 
| 152 | 
< | 
                               c[1] + c[0]) / (data[3].first - data[0].first); | 
| 152 | 
> | 
                               c[1] + c[0]) / (data_[3].first - data_[0].first); | 
| 153 | 
  | 
   | 
| 154 | 
  | 
  fpn = c[n-2] + b[n-2]*(c[n-2] - c[n-3])/(b[n-3] + b[n-2]); | 
| 155 | 
  | 
 | 
| 156 | 
  | 
  if (n > 3)  fpn = fpn + b[n-2] *  | 
| 157 | 
  | 
    (c[n-2] - c[n-3] - (b[n-3] + b[n-2]) *  | 
| 158 | 
< | 
     (c[n-3] - c[n-4])/(b[n-3] + b[n-4]))/(data[n-1].first - data[n-4].first); | 
| 158 | 
> | 
     (c[n-3] - c[n-4])/(b[n-3] + b[n-4]))/(data_[n-1].first - data_[n-4].first); | 
| 159 | 
  | 
   | 
| 160 | 
  | 
   | 
| 161 | 
  | 
  // Calculate the right hand side and store it in C. | 
| 181 | 
  | 
  // Calculate the coefficients defining the spline. | 
| 182 | 
  | 
   | 
| 183 | 
  | 
  for (int i = 0; i < n-1; i++) { | 
| 184 | 
< | 
    h = data[i+1].first - data[i].first; | 
| 184 | 
> | 
    h = data_[i+1].first - data_[i].first; | 
| 185 | 
  | 
    d[i] = (c[i+1] - c[i]) / (3.0 * h); | 
| 186 | 
< | 
    b[i] = (data[i+1].second - data[i].second)/h - h * (c[i] + h * d[i]); | 
| 186 | 
> | 
    b[i] = (data_[i+1].second - data_[i].second)/h - h * (c[i] + h * d[i]); | 
| 187 | 
  | 
  } | 
| 188 | 
  | 
   | 
| 189 | 
  | 
  b[n-1] = b[n-2] + h * (2.0 * c[n-2] + h * 3.0 * d[n-2]); | 
| 190 | 
  | 
   | 
| 191 | 
< | 
  if (isUniform) dx = 1.0 / (data[1].first - data[0].first);  | 
| 191 | 
> | 
  if (isUniform) dx = 1.0 / (data_[1].first - data_[0].first);  | 
| 192 | 
  | 
   | 
| 193 | 
  | 
  generated = true; | 
| 194 | 
  | 
  return; | 
| 205 | 
  | 
  if (!generated) generate(); | 
| 206 | 
  | 
  RealType dt; | 
| 207 | 
  | 
   | 
| 208 | 
< | 
  if ( t < data[0].first || t > data[n-1].first ) {     | 
| 208 | 
> | 
  if ( t < data_[0].first || t > data_[n-1].first ) {     | 
| 209 | 
  | 
    sprintf( painCave.errMsg, | 
| 210 | 
  | 
             "CubicSpline::getValueAt was passed a value outside the range of the spline!\n"); | 
| 211 | 
  | 
    painCave.severity = OPENMD_ERROR; | 
| 220 | 
  | 
 | 
| 221 | 
  | 
  if (isUniform) {     | 
| 222 | 
  | 
     | 
| 223 | 
< | 
    j = max(0, min(n-1, int((t - data[0].first) * dx)));    | 
| 223 | 
> | 
    j = max(0, min(n-1, int((t - data_[0].first) * dx)));    | 
| 224 | 
  | 
 | 
| 225 | 
  | 
  } else {  | 
| 226 | 
  | 
 | 
| 227 | 
  | 
    j = n-1; | 
| 228 | 
  | 
     | 
| 229 | 
  | 
    for (int i = 0; i < n; i++) { | 
| 230 | 
< | 
      if ( t < data[i].first ) { | 
| 230 | 
> | 
      if ( t < data_[i].first ) { | 
| 231 | 
  | 
        j = i-1; | 
| 232 | 
  | 
        break; | 
| 233 | 
  | 
      }       | 
| 236 | 
  | 
   | 
| 237 | 
  | 
  //  Evaluate the cubic polynomial. | 
| 238 | 
  | 
   | 
| 239 | 
< | 
  dt = t - data[j].first; | 
| 240 | 
< | 
  return data[j].second + dt*(b[j] + dt*(c[j] + dt*d[j])); | 
| 239 | 
> | 
  dt = t - data_[j].first; | 
| 240 | 
> | 
  return data_[j].second + dt*(b[j] + dt*(c[j] + dt*d[j])); | 
| 241 | 
  | 
   | 
| 242 | 
  | 
} | 
| 243 | 
  | 
 | 
| 253 | 
  | 
  if (!generated) generate(); | 
| 254 | 
  | 
  RealType dt; | 
| 255 | 
  | 
   | 
| 256 | 
< | 
  if ( t < data.front().first || t > data.back().first ) {     | 
| 256 | 
> | 
  if ( t < data_.front().first || t > data_.back().first ) {     | 
| 257 | 
  | 
    sprintf( painCave.errMsg, | 
| 258 | 
  | 
             "CubicSpline::getValueAndDerivativeAt was passed a value outside the range of the spline!\n"); | 
| 259 | 
  | 
    painCave.severity = OPENMD_ERROR; | 
| 268 | 
  | 
 | 
| 269 | 
  | 
  if (isUniform) {     | 
| 270 | 
  | 
     | 
| 271 | 
< | 
    j = max(0, min(n-1, int((t - data[0].first) * dx)));    | 
| 271 | 
> | 
    j = max(0, min(n-1, int((t - data_[0].first) * dx)));    | 
| 272 | 
  | 
 | 
| 273 | 
  | 
  } else {  | 
| 274 | 
  | 
 | 
| 275 | 
  | 
    j = n-1; | 
| 276 | 
  | 
     | 
| 277 | 
  | 
    for (int i = 0; i < n; i++) { | 
| 278 | 
< | 
      if ( t < data[i].first ) { | 
| 278 | 
> | 
      if ( t < data_[i].first ) { | 
| 279 | 
  | 
        j = i-1; | 
| 280 | 
  | 
        break; | 
| 281 | 
  | 
      }       | 
| 284 | 
  | 
   | 
| 285 | 
  | 
  //  Evaluate the cubic polynomial. | 
| 286 | 
  | 
   | 
| 287 | 
< | 
  dt = t - data[j].first; | 
| 287 | 
> | 
  dt = t - data_[j].first; | 
| 288 | 
  | 
 | 
| 289 | 
< | 
  RealType yval = data[j].second + dt*(b[j] + dt*(c[j] + dt*d[j])); | 
| 289 | 
> | 
  RealType yval = data_[j].second + dt*(b[j] + dt*(c[j] + dt*d[j])); | 
| 290 | 
  | 
  RealType dydx = b[j] + dt*(2.0 * c[j] + 3.0 * dt * d[j]); | 
| 291 | 
  | 
   | 
| 292 | 
  | 
  return make_pair(yval, dydx); |