src/gmxlib/selection/symrec.c

   1 /*
   2  *
   3  *                This source code is part of
   4  *
   5  *                 G   R   O   M   A   C   S
   6  *
   7  *          GROningen MAchine for Chemical Simulations
   8  *
   9  * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
  10  * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
  11  * Copyright (c) 2001-2009, The GROMACS development team,
  12  * check out http://www.gromacs.org for more information.
  13
  14  * This program is free software; you can redistribute it and/or
  15  * modify it under the terms of the GNU General Public License
  16  * as published by the Free Software Foundation; either version 2
  17  * of the License, or (at your option) any later version.
  18  *
  19  * If you want to redistribute modifications, please consider that
  20  * scientific software is very special. Version control is crucial -
  21  * bugs must be traceable. We will be happy to consider code for
  22  * inclusion in the official distribution, but derived work must not
  23  * be called official GROMACS. Details are found in the README & COPYING
  24  * files - if they are missing, get the official version at www.gromacs.org.
  25  *
  26  * To help us fund GROMACS development, we humbly ask that you cite
  27  * the papers on the package - you can find them in the top README file.
  28  *
  29  * For more info, check our website at http://www.gromacs.org
  30  */
  31 /*! \internal \file
  32  * \brief Implementation of functions in symrec.h.
  33  */
  34 #ifdef HAVE_CONFIG_H
  35 #include <config.h>
  36 #endif
  37
  38 #include <macros.h>
  39 #include <smalloc.h>
  40 #include <string2.h>
  41 #include <typedefs.h>
  42 #include <gmx_fatal.h>
  43
  44 #include <poscalc.h>
  45
  46 #include "selelem.h"
  47 #include "symrec.h"
  48
  49 /*! \internal \brief
  50  * Symbol table for the selection parser.
  51  */
  52 struct gmx_sel_symtab_t
  53 {
  54     /** Pointer to the first symbol in the linked list of symbols. */
  55     gmx_sel_symrec_t *first;
  56 };
  57
  58 /*! \internal \brief
  59  * Single symbol for the selection parser.
  60  */
  61 struct gmx_sel_symrec_t
  62 {
  63     /** Name of the symbol. */
  64     char                           *name;
  65     /** Type of the symbol. */
  66     e_symbol_t                      type;
  67     /** Value of the symbol. */
  68     union {
  69         /** Pointer to the method structure (\ref SYMBOL_METHOD). */
  70         struct gmx_ana_selmethod_t *meth;
  71         /** Pointer to the variable value (\ref SYMBOL_VARIABLE). */
  72         struct t_selelem           *var;
  73     }                               u;
  74     /** Pointer to the next symbol. */
  75     struct gmx_sel_symrec_t        *next;
  76 };
  77
  78 /** List of reserved symbols to register in add_reserved_symbols(). */
  79 static const char *const sym_reserved[] = {
  80     "group",
  81     "to",
  82     "not",
  83     "and",
  84     "or",
  85     "xor",
  86     "yes",
  87     "no",
  88     "on",
  89     "off",
  90     "help",
  91 };
  92
  93 /*!
  94  * \param[in] sym Symbol to query.
  95  * \returns   The name of \p sym.
  96  *
  97  * The returned pointer should not be free'd.
  98  */
  99 char *
 100 _gmx_sel_sym_name(gmx_sel_symrec_t *sym)
 101 {
 102     return sym->name;
 103 }
 104
 105 /*!
 106  * \param[in] sym Symbol to query.
 107  * \returns   The type of \p sym.
 108  */
 109 e_symbol_t
 110 _gmx_sel_sym_type(gmx_sel_symrec_t *sym)
 111 {
 112     return sym->type;
 113 }
 114
 115 /*!
 116  * \param[in] sym Symbol to query.
 117  * \returns   The method associated with \p sym, or NULL if \p sym is not a
 118  *   \ref SYMBOL_METHOD symbol.
 119  */
 120 struct gmx_ana_selmethod_t *
 121 _gmx_sel_sym_value_method(gmx_sel_symrec_t *sym)
 122 {
 123     if (sym->type != SYMBOL_METHOD)
 124     {
 125         gmx_call("symbol is not a method symbol");
 126         return NULL;
 127     }
 128     return sym->u.meth;
 129 }
 130
 131 /*!
 132  * \param[in] sym Symbol to query.
 133  * \returns   The variable expression associated with \p sym, or NULL if
 134  *   \p sym is not a \ref SYMBOL_VARIABLE symbol.
 135  */
 136 struct t_selelem *
 137 _gmx_sel_sym_value_var(gmx_sel_symrec_t *sym)
 138 {
 139     if (sym->type != SYMBOL_VARIABLE)
 140     {
 141         gmx_call("symbol is not a variable symbol");
 142         return NULL;
 143     }
 144     return sym->u.var;
 145 }
 146
 147 /*! \brief
 148  * Adds the reserved symbols to a symbol table.
 149  *
 150  * \param[in,out] tab  Symbol table to which the symbols are added.
 151  *
 152  * Assumes that the symbol table is empty.
 153  */
 154 static void
 155 add_reserved_symbols(gmx_sel_symtab_t *tab)
 156 {
 157     gmx_sel_symrec_t *sym;
 158     gmx_sel_symrec_t *last;
 159     size_t            i;
 160
 161     last = NULL;
 162     for (i = 0; i < asize(sym_reserved); ++i)
 163     {
 164         snew(sym, 1);
 165         sym->name = strdup(sym_reserved[i]);
 166         sym->type = SYMBOL_RESERVED;
 167         sym->next = NULL;
 168         if (last)
 169         {
 170             last->next = sym;
 171         }
 172         else
 173         {
 174             tab->first = sym;
 175         }
 176         last = sym;
 177     }
 178 }
 179
 180 /*! \brief
 181  * Adds the position symbols to the symbol list.
 182  *
 183  * \param[in,out] tab  Symbol table to which the symbols are added.
 184  */
 185 static void
 186 add_position_symbols(gmx_sel_symtab_t *tab)
 187 {
 188     const char       **postypes;
 189     gmx_sel_symrec_t  *sym;
 190     gmx_sel_symrec_t  *last;
 191     int                i;
 192
 193     postypes = gmx_ana_poscalc_create_type_enum(TRUE);
 194     last = tab->first;
 195     while (last && last->next)
 196     {
 197         last = last->next;
 198     }
 199     for (i = 1; postypes[i] != NULL; ++i)
 200     {
 201         snew(sym, 1);
 202         sym->name = strdup(postypes[i]);
 203         sym->type = SYMBOL_POS;
 204         sym->next = NULL;
 205         if (last)
 206         {
 207             last->next = sym;
 208         }
 209         else
 210         {
 211             tab->first = sym;
 212         }
 213         last = sym;
 214     }
 215     sfree(postypes);
 216 }
 217
 218 /*!
 219  * \param[out] tabp Symbol table pointer to initialize.
 220  *
 221  * Reserved and position symbols are added to the created table.
 222  */
 223 int
 224 _gmx_sel_symtab_create(gmx_sel_symtab_t **tabp)
 225 {
 226     gmx_sel_symtab_t *tab;
 227
 228     snew(tab, 1);
 229     add_reserved_symbols(tab);
 230     add_position_symbols(tab);
 231     *tabp = tab;
 232     return 0;
 233 }
 234
 235 /*!
 236  * \param[in] tab Symbol table to free.
 237  *
 238  * The pointer \p tab is invalid after the call.
 239  */
 240 void
 241 _gmx_sel_symtab_free(gmx_sel_symtab_t *tab)
 242 {
 243     gmx_sel_symrec_t *sym;
 244
 245     while (tab->first)
 246     {
 247         sym = tab->first;
 248         tab->first = sym->next;
 249         if (sym->type == SYMBOL_VARIABLE)
 250         {
 251             _gmx_selelem_free(sym->u.var);
 252         }
 253         sfree(sym->name);
 254         sfree(sym);
 255     }
 256     sfree(tab);
 257 }
 258
 259 /*!
 260  * \param[in] tab    Symbol table to search.
 261  * \param[in] name   Symbol name to find.
 262  * \param[in] bExact If FALSE, symbols that begin with \p name are also
 263  *   considered.
 264  * \returns   Pointer to the symbol with name \p name, or NULL if not found.
 265  *
 266  * If no exact match is found and \p bExact is FALSE, returns a symbol that
 267  * begins with \p name if a unique matching symbol is found.
 268  */
 269 gmx_sel_symrec_t *
 270 _gmx_sel_find_symbol(gmx_sel_symtab_t *tab, const char *name, bool bExact)
 271 {
 272     return _gmx_sel_find_symbol_len(tab, name, strlen(name), bExact);
 273 }
 274
 275 /*!
 276  * \param[in] tab    Symbol table to search.
 277  * \param[in] name   Symbol name to find.
 278  * \param[in] len    Only consider the first \p len characters of \p name.
 279  * \param[in] bExact If FALSE, symbols that begin with \p name are also
 280  *   considered.
 281  * \returns   Pointer to the symbol with name \p name, or NULL if not found.
 282  *
 283  * If no exact match is found and \p bExact is FALSE, returns a symbol that
 284  * begins with \p name if a unique matching symbol is found.
 285  *
 286  * The parameter \p len is there to allow using this function from scanner.l
 287  * without modifying the text to be scanned or copying it.
 288  */
 289 gmx_sel_symrec_t *
 290 _gmx_sel_find_symbol_len(gmx_sel_symtab_t *tab, const char *name, size_t len,
 291                          bool bExact)
 292 {
 293     gmx_sel_symrec_t *sym;
 294     gmx_sel_symrec_t *match;
 295     bool              bUnique;
 296     bool              bMatch;
 297
 298     match = NULL;
 299     bUnique = TRUE;
 300     bMatch  = FALSE;
 301     sym = tab->first;
 302     while (sym)
 303     {
 304         if (!strncmp(sym->name, name, len))
 305         {
 306             if (strlen(sym->name) == len)
 307             {
 308                 return sym;
 309             }
 310             if (bMatch)
 311             {
 312                 bUnique = FALSE;
 313             }
 314             bMatch = TRUE;
 315             if (sym->type == SYMBOL_METHOD)
 316             {
 317                 match = sym;
 318             }
 319         }
 320         sym = sym->next;
 321     }
 322     if (bExact)
 323     {
 324         return NULL;
 325     }
 326
 327     if (!bUnique)
 328     {
 329         fprintf(stderr, "parse error: ambiguous symbol\n");
 330         return NULL;
 331     }
 332     return match;
 333 }
 334
 335 /*!
 336  * \param[in] tab   Symbol table to search.
 337  * \param[in] type  Type of symbol to find.
 338  * \returns   The first symbol in \p tab with type \p type,
 339  *   or NULL if there are no such symbols.
 340  */
 341 gmx_sel_symrec_t *
 342 _gmx_sel_first_symbol(gmx_sel_symtab_t *tab, e_symbol_t type)
 343 {
 344     gmx_sel_symrec_t *sym;
 345
 346     sym = tab->first;
 347     while (sym)
 348     {
 349         if (sym->type == type)
 350         {
 351             return sym;
 352         }
 353         sym = sym->next;
 354     }
 355     return NULL;
 356 }
 357
 358 /*!
 359  * \param[in] after Start the search after this symbol.
 360  * \param[in] type  Type of symbol to find.
 361  * \returns   The next symbol after \p after with type \p type,
 362  *   or NULL if there are no more symbols.
 363  */
 364 gmx_sel_symrec_t *
 365 _gmx_sel_next_symbol(gmx_sel_symrec_t *after, e_symbol_t type)
 366 {
 367     gmx_sel_symrec_t *sym;
 368
 369     sym = after->next;
 370     while (sym)
 371     {
 372         if (sym->type == type)
 373         {
 374             return sym;
 375         }
 376         sym = sym->next;
 377     }
 378     return NULL;
 379 }
 380
 381 /*! \brief
 382  * Internal utility function used in adding symbols to a symbol table.
 383  *
 384  * \param[in,out] tab   Symbol table to add the symbol to.
 385  * \param[in]     name  Name of the symbol to add.
 386  * \param[out]    ctype On error, the type of the conflicting symbol is
 387  *   written to \p *ctype.
 388  * \returns       Pointer to the new symbol record, or NULL if \p name
 389  *   conflicts with an existing symbol.
 390  */
 391 static gmx_sel_symrec_t *
 392 add_symbol(gmx_sel_symtab_t *tab, const char *name, e_symbol_t *ctype)
 393 {
 394     gmx_sel_symrec_t *sym, *psym;
 395     int               len;
 396
 397     /* Check if there is a conflicting symbol */
 398     psym = NULL;
 399     sym  = tab->first;
 400     while (sym)
 401     {
 402         if (!strcasecmp(sym->name, name))
 403         {
 404             *ctype = sym->type;
 405             return NULL;
 406         }
 407         psym = sym;
 408         sym  = sym->next;
 409     }
 410
 411     /* Create a new symbol record */
 412     if (psym == NULL)
 413     {
 414         snew(tab->first, 1);
 415         sym = tab->first;
 416     }
 417     else
 418     {
 419         snew(psym->next, 1);
 420         sym = psym->next;
 421     }
 422     sym->name = strdup(name);
 423     return sym;
 424 }
 425
 426 /*!
 427  * \param[in,out] tab    Symbol table to add the symbol to.
 428  * \param[in]     name   Name of the new symbol.
 429  * \param[in]     sel    Value of the variable.
 430  * \returns       Pointer to the created symbol record, or NULL if there was a
 431  *   symbol with the same name.
 432  */
 433 gmx_sel_symrec_t *
 434 _gmx_sel_add_var_symbol(gmx_sel_symtab_t *tab, const char *name,
 435                         struct t_selelem *sel)
 436 {
 437     gmx_sel_symrec_t *sym;
 438     e_symbol_t        ctype;
 439
 440     sym = add_symbol(tab, name, &ctype);
 441     if (!sym)
 442     {
 443         fprintf(stderr, "parse error: ");
 444         switch (ctype)
 445         {
 446             case SYMBOL_RESERVED:
 447             case SYMBOL_POS:
 448                 fprintf(stderr, "variable name (%s) conflicts with a reserved keyword\n",
 449                         name);
 450                 break;
 451             case SYMBOL_VARIABLE:
 452                 fprintf(stderr, "duplicate variable name (%s)\n", name);
 453                 break;
 454             case SYMBOL_METHOD:
 455                 fprintf(stderr, "variable name (%s) conflicts with a selection keyword\n",
 456                         name);
 457                 break;
 458         }
 459         return NULL;
 460     }
 461
 462     sym->type  = SYMBOL_VARIABLE;
 463     sym->u.var = sel;
 464     sel->refcount++;
 465     return sym;
 466 }
 467
 468 /*!
 469  * \param[in,out] tab    Symbol table to add the symbol to.
 470  * \param[in]     name   Name of the new symbol.
 471  * \param[in]     method Method that this symbol represents.
 472  * \returns       Pointer to the created symbol record, or NULL if there was a
 473  *   symbol with the same name.
 474  */
 475 gmx_sel_symrec_t *
 476 _gmx_sel_add_method_symbol(gmx_sel_symtab_t *tab, const char *name,
 477                            struct gmx_ana_selmethod_t *method)
 478 {
 479     gmx_sel_symrec_t *sym;
 480     e_symbol_t        ctype;
 481
 482     sym = add_symbol(tab, name, &ctype);
 483     if (!sym)
 484     {
 485         fprintf(stderr, "parse error: ");
 486         switch (ctype)
 487         {
 488             case SYMBOL_RESERVED:
 489             case SYMBOL_POS:
 490                 fprintf(stderr, "method name (%s) conflicts with a reserved keyword\n",
 491                         name);
 492                 break;
 493             case SYMBOL_VARIABLE:
 494                 fprintf(stderr, "method name (%s) conflicts with a variable name\n",
 495                         name);
 496                 break;
 497             case SYMBOL_METHOD:
 498                 fprintf(stderr, "duplicate method name (%s)\n", name);
 499                 break;
 500         }
 501         return NULL;
 502     }
 503
 504     sym->type   = SYMBOL_METHOD;
 505     sym->u.meth = method;
 506     return sym;
 507 }