src/gromacs/statistics/statistics.h

   1 /*
   2  * This file is part of the GROMACS molecular simulation package.
   3  *
   4  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
   5  * Copyright (c) 2001-2008, The GROMACS development team.
   6  * Copyright (c) 2010,2014,2015, by the GROMACS development team, led by
   7  * Mark Abraham, David van der Spoel, Berk Hess, and Erik Lindahl,
   8  * and including many others, as listed in the AUTHORS file in the
   9  * top-level source directory and at http://www.gromacs.org.
  10  *
  11  * GROMACS is free software; you can redistribute it and/or
  12  * modify it under the terms of the GNU Lesser General Public License
  13  * as published by the Free Software Foundation; either version 2.1
  14  * of the License, or (at your option) any later version.
  15  *
  16  * GROMACS is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  19  * Lesser General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU Lesser General Public
  22  * License along with GROMACS; if not, see
  23  * http://www.gnu.org/licenses, or write to the Free Software Foundation,
  24  * Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA.
  25  *
  26  * If you want to redistribute modifications to GROMACS, please
  27  * consider that scientific software is very special. Version
  28  * control is crucial - bugs must be traceable. We will be happy to
  29  * consider code for inclusion in the official distribution, but
  30  * derived work must not be called official GROMACS. Details are found
  31  * in the README & COPYING files - if they are missing, get the
  32  * official version at http://www.gromacs.org.
  33  *
  34  * To help us fund GROMACS development, we humbly ask that you cite
  35  * the research papers on the package. Check out http://www.gromacs.org.
  36  */
  37 /*! \libinternal \file
  38  * \brief
  39  * Declares simple statistics toolbox
  40  *
  41  * \authors David van der Spoel <david.vanderspoel@icm.uu.se>
  42  * \inlibraryapi
  43  */
  44 #ifndef GMX_STATISTICS_H
  45 #define GMX_STATISTICS_H
  46
  47 #include <cstdio>
  48
  49 #include "gromacs/utility/real.h"
  50
  51 //! Abstract container type
  52 typedef struct gmx_stats *gmx_stats_t;
  53
  54 //! Error codes returned by the routines
  55 enum {
  56     estatsOK, estatsNO_POINTS, estatsNO_MEMORY, estatsERROR,
  57     estatsINVALID_INPUT, estatsNOT_IMPLEMENTED, estatsNR
  58 };
  59
  60 //! Enum for statistical weights
  61 enum {
  62     elsqWEIGHT_NONE, elsqWEIGHT_X, elsqWEIGHT_Y,
  63     elsqWEIGHT_XY, elsqWEIGHT_NR
  64 };
  65
  66 //! Enum determining which coordinate to histogram
  67 enum {
  68     ehistoX, ehistoY, ehistoNR
  69 };
  70
  71 /*! \brief
  72  * Initiate a data structure
  73  * \return the data structure
  74  */
  75 gmx_stats_t gmx_stats_init();
  76
  77 /*! \brief
  78  * Destroy a data structure
  79  * \param stats The data structure
  80  */
  81 void gmx_stats_free(gmx_stats_t stats);
  82
  83 /*! \brief
  84  * Remove outliers from a straight line, where level in units of
  85  * sigma. Level needs to be larger than one obviously.
  86  * \param[in] stats The data structure
  87  * \param[in] level The sigma level
  88  * \return error code
  89  */
  90 int gmx_stats_remove_outliers(gmx_stats_t stats, double level);
  91
  92 /*! \brief
  93  * Add a point to the data set
  94  * \param[in] stats The data structure
  95  * \param[in] x   The x value
  96  * \param[in] y   The y value
  97  * \param[in] dx  The error in the x value
  98  * \param[in] dy  The error in the y value
  99  * \return error code
 100  */
 101 int gmx_stats_add_point(gmx_stats_t stats, double x, double y,
 102                         double dx, double dy);
 103
 104 /*! \brief
 105  * Add a series of datapoints at once. The arrays dx and dy may
 106  * be NULL in that case zero uncertainties will be assumed.
 107  *
 108  * \param[in] stats The data structure
 109  * \param[in] n   Number of points
 110  * \param[in] x   The array of x values
 111  * \param[in] y   The array of y values
 112  * \param[in] dx  The error in the x value
 113  * \param[in] dy  The error in the y value
 114  * \return error code
 115  */
 116 int gmx_stats_add_points(gmx_stats_t stats, int n, real *x, real *y,
 117                          real *dx, real *dy);
 118
 119 /*! \brief
 120  * Delivers data points from the statistics.
 121  *
 122  * Should be used in a while loop. Variables for either
 123  * pointer may be NULL, in which case the routine can be used as an
 124  * expensive point counter.
 125  * Return the data points one by one. Return estatsOK while there are
 126  *  more points, and returns estatsNOPOINTS when the last point has
 127  *  been returned.
 128  *  If level > 0 then the outliers outside level*sigma are reported
 129  * only.
 130  * \param[in] stats The data structure
 131  * \param[out] x   The array of x values
 132  * \param[out] y   The array of y values
 133  * \param[out] dx  The error in the x value
 134  * \param[out] dy  The error in the y value
 135  * \param[in]  level sigma level (see above)
 136  * \return error code
 137  */
 138 int gmx_stats_get_point(gmx_stats_t stats, real *x, real *y,
 139                         real *dx, real *dy, real level);
 140
 141 /*! \brief
 142  * Fit the data to y = ax + b, possibly weighted, if uncertainties
 143  * have been input. da and db may be NULL.
 144  * \param[in] stats The data structure
 145  * \param[in] weight type of weighting
 146  * \param[out] a slope
 147  * \param[out] b intercept
 148  * \param[out] da sigma in a
 149  * \param[out] db sigma in b
 150  * \param[out] chi2 normalized quality of fit
 151  * \param[out] Rfit correlation coefficient
 152  * \return error code
 153  */
 154 int gmx_stats_get_ab(gmx_stats_t stats, int weight,
 155                      real *a, real *b,
 156                      real *da, real *db, real *chi2, real *Rfit);
 157
 158 /*! \brief
 159  * Fit the data to y = ax, possibly weighted, if uncertainties have
 160  * have been input. da and db may be NULL.
 161  * \param[in] stats The data structure
 162  * \param[in] weight type of weighting
 163  * \param[out] a slope
 164  * \param[out] da sigma in a
 165  * \param[out] chi2 normalized quality of fit
 166  * \param[out] Rfit correlation coefficient
 167  * \return error code
 168  */
 169 int gmx_stats_get_a(gmx_stats_t stats, int weight,
 170                     real *a, real *da, real *chi2, real *Rfit);
 171
 172 /*! \brief
 173  * Get the correlation coefficient.
 174  * \param[in]  stats The data structure
 175  * \param[out] R the correlation coefficient between the data (x and y) as input to the structure.
 176  * \return error code
 177  */
 178 int gmx_stats_get_corr_coeff(gmx_stats_t stats, real *R);
 179
 180 /*! \brief
 181  * Get the root mean square deviation.
 182  * \param[in]  stats The data structure
 183  * \param[out] rmsd  the root mean square deviation between x and y values.
 184  * \return error code
 185  */
 186 int gmx_stats_get_rmsd(gmx_stats_t stats, real *rmsd);
 187
 188 /*! \brief
 189  * Get the number of points.
 190  * \param[in]  stats The data structure
 191  * \param[out] N     number of data points
 192  * \return error code
 193  */
 194 int gmx_stats_get_npoints(gmx_stats_t stats, int *N);
 195
 196 /*! \brief
 197  * Computes and returns the average value.
 198  * \param[in]  stats The data structure
 199  * \param[out] aver  Average value
 200  * \return error code
 201  */
 202 int gmx_stats_get_average(gmx_stats_t stats, real *aver);
 203
 204 /*! \brief
 205  * Computes and returns the standard deviation.
 206  * \param[in]  stats The data structure
 207  * \param[out] sigma  Standard deviation
 208  * \return error code
 209  */
 210 int gmx_stats_get_sigma(gmx_stats_t stats, real *sigma);
 211
 212 /*! \brief
 213  * Computes and returns the standard error.
 214  * \param[in]  stats The data structure
 215  * \param[out] error Standard error
 216  * \return error code
 217  */
 218 int gmx_stats_get_error(gmx_stats_t stats, real *error);
 219
 220 /*! \brief
 221  * Pointers may be null, in which case no assignment will be done.
 222  * \param[in]  stats The data structure
 223  * \param[out] aver  Average value
 224  * \param[out] sigma  Standard deviation
 225  * \param[out] error Standard error
 226  * \return error code
 227  */
 228 int gmx_stats_get_ase(gmx_stats_t stats, real *aver, real *sigma, real *error);
 229
 230 /*! \brief
 231  * Dump the x, y, dx, dy data to a text file
 232  * \param[in]  stats The data structure
 233  * \param[in] fp  File pointer
 234  * \return error code
 235  */
 236 int gmx_stats_dump_xy(gmx_stats_t stats, FILE *fp);
 237
 238 /*! \brief
 239  * Make a histogram of the data present.
 240  *
 241  * Uses either binwidth to
 242  * determine the number of bins, or nbins to determine the binwidth,
 243  * therefore one of these should be zero, but not the other. If *nbins = 0
 244  * the number of bins will be returned in this variable. ehisto should be one of
 245  * ehistoX or ehistoY. If
 246  * normalized not equal to zero, the integral of the histogram will be
 247  * normalized to one. The output is in two arrays, *x and *y, to which
 248  * you should pass a pointer. Memory for the arrays will be allocated
 249  * as needed. Function returns one of the estats codes.
 250  * \param[in]  stats The data structure
 251  * \param[in] binwidth For the histogram
 252  * \param[in] nbins    Number of bins
 253  * \param[in] ehisto   Type (see enum above)
 254  * \param[in] normalized see above
 255  * \param[out] x see above
 256  * \param[out] y see above
 257  * \return error code
 258  */
 259 int gmx_stats_make_histogram(gmx_stats_t stats, real binwidth, int *nbins,
 260                              int ehisto,
 261                              int normalized, real **x, real **y);
 262
 263 /*! \brief
 264  * Return message belonging to error code
 265  * \param[in] estats error code
 266  */
 267 const char *gmx_stats_message(int estats);
 268
 269 /****************************************************
 270  * Some statistics utilities for convenience: useful when a complete data
 271  * set is available already from another source, e.g. an xvg file.
 272  ****************************************************/
 273 /*! \brief
 274  * Fit a straight line y=ax thru the n data points x, y, return the
 275  * slope in *a.
 276  * \param[in] n number of points
 277  * \param[in] x data points x
 278  * \param[in] y data point y
 279  * \param[out] a slope
 280  * \return error code
 281  */
 282 int lsq_y_ax(int n, real x[], real y[], real *a);
 283
 284 /*! \brief
 285  * Fit a straight line y=ax+b thru the n data points x, y.
 286  * \param[in] n number of points
 287  * \param[in] x data points x
 288  * \param[in] y data point y
 289  * \param[out] a slope
 290  * \param[out] b intercept
 291  * \param[out] r correlation coefficient
 292  * \param[out] chi2 quality of fit
 293  * \return error code
 294  */
 295 int lsq_y_ax_b(int n, real x[], real y[], real *a, real *b, real *r,
 296                real *chi2);
 297
 298 /*! \copydoc lsq_y_ax_b
 299  */
 300 int lsq_y_ax_b_xdouble(int n, double x[], real y[],
 301                        real *a, real *b, real *r, real *chi2);
 302
 303 /*! \brief
 304  * Fit a straight line y=ax+b thru the n data points x, y.
 305  * \param[in] n number of points
 306  * \param[in] x data points x
 307  * \param[in] y data point y
 308  * \param[in] dy uncertainty in data point y
 309  * \param[out] a slope
 310  * \param[out] b intercept
 311  * \param[out] da error in slope
 312  * \param[out] db error in intercept
 313  * \param[out] r correlation coefficient
 314  * \param[out] chi2 quality of fit
 315  * \return error code
 316  */
 317 int lsq_y_ax_b_error(int n, real x[], real y[], real dy[],
 318                      real *a, real *b, real *da, real *db,
 319                      real *r, real *chi2);
 320
 321 #endif