Better selection stdin handling.

[gromacs/qmmm-gamess-us.git] / src / gmxlib / selection / selelem.c

/*
 *
 *                This source code is part of
 *
 *                 G   R   O   M   A   C   S
 *
 *          GROningen MAchine for Chemical Simulations
 *
 * Written by David van der Spoel, Erik Lindahl, Berk Hess, and others.
 * Copyright (c) 1991-2000, University of Groningen, The Netherlands.
 * Copyright (c) 2001-2009, The GROMACS development team,
 * check out http://www.gromacs.org for more information.

 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License
 * as published by the Free Software Foundation; either version 2
 * of the License, or (at your option) any later version.
 *
 * If you want to redistribute modifications, please consider that
 * scientific software is very special. Version control is crucial -
 * bugs must be traceable. We will be happy to consider code for
 * inclusion in the official distribution, but derived work must not
 * be called official GROMACS. Details are found in the README & COPYING
 * files - if they are missing, get the official version at www.gromacs.org.
 *
 * To help us fund GROMACS development, we humbly ask that you cite
 * the papers on the package - you can find them in the top README file.
 *
 * For more info, check our website at http://www.gromacs.org
 */
/*! \internal \file
 * \brief Implementation of functions in selelem.h.
 */
#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include <smalloc.h>

#include <indexutil.h>
#include <poscalc.h>
#include <position.h>
#include <selmethod.h>

#include "evaluate.h"
#include "keywords.h"
#include "selelem.h"

/*!
 * \param[in] sel Selection for which the string is requested
 * \returns   Pointer to a string that corresponds to \p sel->type.
 *
 * The return value points to a string constant and should not be \p free'd.
 * 
 * The function returns NULL if \p sel->type is not one of the valid values.
 */
const char *
_gmx_selelem_type_str(t_selelem *sel)
{
    switch (sel->type)
    {
        case SEL_CONST:      return "CONST";
        case SEL_EXPRESSION: return "EXPR";
        case SEL_BOOLEAN:    return "BOOL";
        case SEL_ROOT:       return "ROOT";
        case SEL_SUBEXPR:    return "SUBEXPR";
        case SEL_SUBEXPRREF: return "REF";
        case SEL_MODIFIER:   return "MODIFIER";
    }
    return NULL;
}

/*!
 * \param[in] val Value structore for which the string is requested.
 * \returns   Pointer to a string that corresponds to \p val->type,
 *   NULL if the type value is invalid.
 *
 * The return value points to a string constant and should not be \p free'd.
 */
const char *
_gmx_sel_value_type_str(gmx_ana_selvalue_t *val)
{
    switch (val->type)
    {
        case NO_VALUE:       return "NONE";
        case INT_VALUE:      return "INT";
        case REAL_VALUE:     return "REAL";
        case STR_VALUE:      return "STR";
        case POS_VALUE:      return "VEC";
        case GROUP_VALUE:    return "GROUP";
    }
    return NULL;
}

/*! \copydoc _gmx_selelem_type_str() */
const char *
_gmx_selelem_boolean_type_str(t_selelem *sel)
{
    switch (sel->u.boolt)
    {
        case BOOL_NOT:  return "NOT"; break;
        case BOOL_AND:  return "AND"; break;
        case BOOL_OR:   return "OR";  break;
        case BOOL_XOR:  return "XOR"; break;
    }
    return NULL;
}

/*!
 * \param[in] type Type of selection element to allocate.
 * \returns   Pointer to the newly allocated and initialized element.
 *
 * \c t_selelem::type is set to \p type,
 * \c t_selelem::v::type is set to \ref GROUP_VALUE for boolean and comparison
 * expressions and \ref NO_VALUE for others,
 * \ref SEL_ALLOCVAL is set for non-root elements (\ref SEL_ALLOCDATA is also
 * set for \ref SEL_BOOLEAN elements),
 * and \c t_selelem::refcount is set to one.
 * All the pointers are set to NULL.
 */
t_selelem *
_gmx_selelem_create(e_selelem_t type)
{
    t_selelem *sel;

    snew(sel, 1);
    sel->name       = NULL;
    sel->type       = type;
    sel->flags      = (type != SEL_ROOT) ? SEL_ALLOCVAL : 0;
    if (type == SEL_BOOLEAN)
    {
        sel->v.type = GROUP_VALUE;
        sel->flags |= SEL_ALLOCDATA;
    }
    else
    {
        sel->v.type = NO_VALUE;
    }
    _gmx_selvalue_clear(&sel->v);
    sel->evaluate   = NULL;
    sel->child      = NULL;
    sel->next       = NULL;
    sel->refcount   = 1;

    return sel;
}

/*!
 * \param[in,out] sel   Selection element to set the type for.
 * \param[in]     vtype Value type for the selection element.
 * \returns       0 on success, EINVAL if the value type is invalid.
 *
 * If the new type is \ref GROUP_VALUE or \ref POS_VALUE, the
 * \ref SEL_ALLOCDATA flag is also set.
 *
 * This function should only be called at most once for each element,
 * preferably right after calling _gmx_selelem_create().
 */
int
_gmx_selelem_set_vtype(t_selelem *sel, e_selvalue_t vtype)
{
    if (sel->type == SEL_BOOLEAN && vtype != GROUP_VALUE)
    {
        gmx_bug("internal error");
        return EINVAL;
    }
    if (sel->v.type != NO_VALUE && vtype != sel->v.type)
    {
        gmx_call("_gmx_selelem_set_vtype() called more than once");
        return EINVAL;
    }
    sel->v.type = vtype;
    if (vtype == GROUP_VALUE || vtype == POS_VALUE)
    {
        sel->flags |= SEL_ALLOCDATA;
    }
    return 0;
}

/*!
 * \param[in] sel Selection to free.
 */
void
_gmx_selelem_free_values(t_selelem *sel)
{
    int   i, n;

    if ((sel->flags & SEL_ALLOCDATA) && sel->v.u.ptr)
    {
        /* The number of position/group structures is constant, so the
         * backup of using sel->v.nr should work for them.
         * For strings, we report an error if we don't know the allocation
         * size here. */
        n = (sel->v.nalloc > 0) ? sel->v.nalloc : sel->v.nr;
        switch (sel->v.type)
        {
            case STR_VALUE:
                if (sel->v.nalloc == 0)
                {
                    gmx_bug("SEL_ALLOCDATA should only be set for allocated STR_VALUE values");
                    break;
                }
                for (i = 0; i < n; ++i)
                {
                    sfree(sel->v.u.s[i]);
                }
                break;
            case POS_VALUE:
                for (i = 0; i < n; ++i)
                {
                    gmx_ana_pos_deinit(&sel->v.u.p[i]);
                }
                break;
            case GROUP_VALUE:
                for (i = 0; i < n; ++i)
                {
                    gmx_ana_index_deinit(&sel->v.u.g[i]);
                }
                break;
            default: /* No special handling for other types */
                break;
        }
    }
    if (sel->flags & SEL_ALLOCVAL)
    {
        sfree(sel->v.u.ptr);
    }
    _gmx_selvalue_setstore(&sel->v, NULL);
}

/*!
 * \param[in] sel Selection to free.
 */
void
_gmx_selelem_free_exprdata(t_selelem *sel)
{
    int i;

    if (sel->type == SEL_EXPRESSION || sel->type == SEL_MODIFIER)
    {
        /* Free method data */
        if (sel->u.expr.mdata)
        {
            if (sel->u.expr.method && sel->u.expr.method->free)
            {
                sel->u.expr.method->free(sel->u.expr.mdata);
            }
            sfree(sel->u.expr.mdata);
            sel->u.expr.mdata = NULL;
        }
        /* Free the method itself */
        if (sel->u.expr.method)
        {
            /* If the method has not yet been initialized, we must free the
             * memory allocated for parameter values here. */
            if (!(sel->flags & SEL_METHODINIT))
            {
                for (i = 0; i < sel->u.expr.method->nparams; ++i)
                {
                    gmx_ana_selparam_t *param = &sel->u.expr.method->param[i];

                    if ((param->flags & (SPAR_VARNUM | SPAR_ATOMVAL))
                        && param->val.type != GROUP_VALUE
                        && param->val.type != POS_VALUE)
                    {
                        /* We don't need to check for enum values here, because
                         * SPAR_ENUMVAL cannot be combined with the flags
                         * required above. If it ever will be, this results
                         * in a double free within this function, which should
                         * be relatively easy to debug.
                         */
                        if (param->val.type == STR_VALUE)
                        {
                            int j;

                            for (j = 0; j < param->val.nr; ++j)
                            {
                                sfree(param->val.u.s[j]);
                            }
                        }
                        sfree(param->val.u.ptr);
                    }
                }
            }
            /* And even if it is, the arrays allocated for enum values need
             * to be freed. */
            for (i = 0; i < sel->u.expr.method->nparams; ++i)
            {
                gmx_ana_selparam_t *param = &sel->u.expr.method->param[i];

                if (param->flags & SPAR_ENUMVAL)
                {
                    sfree(param->val.u.ptr);
                }

            }
            sfree(sel->u.expr.method->param);
            sfree(sel->u.expr.method);
            sel->u.expr.method = NULL;
        }
        /* Free position data */
        if (sel->u.expr.pos)
        {
            gmx_ana_pos_free(sel->u.expr.pos);
            sel->u.expr.pos = NULL;
        }
        /* Free position calculation data */
        if (sel->u.expr.pc)
        {
            gmx_ana_poscalc_free(sel->u.expr.pc);
            sel->u.expr.pc = NULL;
        }
    }
}

/*!
 * \param[in] sel Selection to free.
 *
 * Decrements \ref t_selelem::refcount "sel->refcount" and frees the
 * memory allocated for \p sel and all its children if the reference count
 * reaches zero.
 */
void
_gmx_selelem_free(t_selelem *sel)
{
    t_selelem *child, *prev;
    
    /* Decrement the reference counter and do nothing if references remain */
    sel->refcount--;
    if (sel->refcount > 0)
    {
        return;
    }

    /* Free the children */
    child = sel->child;
    while (child)
    {
        prev = child;
        child = child->next;
        _gmx_selelem_free(prev);
    }

    /* Free value storage */
    _gmx_selelem_free_values(sel);

    /* Free other storage */
    _gmx_selelem_free_exprdata(sel);
    if (sel->type == SEL_SUBEXPR || sel->type == SEL_ROOT
        || (sel->type == SEL_CONST && sel->v.type == GROUP_VALUE))
    {
        gmx_ana_index_deinit(&sel->u.cgrp);
    }

    /* Free temporary compiler data if present */
    _gmx_selelem_free_compiler_data(sel);

    sfree(sel);
}

/*!
 * \param[in] first First selection to free.
 *
 * Frees \p first and all selections accessible through the
 * \ref t_selelem::next "first->next" pointer.
 */
void
_gmx_selelem_free_chain(t_selelem *first)
{
    t_selelem *child, *prev;

    child = first;
    while (child)
    {
        prev = child;
        child = child->next;
        _gmx_selelem_free(prev);
    }
}

/*! \brief
 * Writes out a human-readable name for the evaluation function.
 *
 * \param[in] fp  File handle to receive the output.
 * \param[in] sel Selection element for which the evaluation function is printed.
 */
static void
print_evaluation_func(FILE *fp, t_selelem *sel)
{
    fprintf(fp, " eval=");
    if (!sel->evaluate)
        fprintf(fp, "none");
    else if (sel->evaluate == &_gmx_sel_evaluate_root)
        fprintf(fp, "root");
    else if (sel->evaluate == &_gmx_sel_evaluate_static)
        fprintf(fp, "static");
    else if (sel->evaluate == &_gmx_sel_evaluate_subexpr_pass)
        fprintf(fp, "subexpr_pass");
    else if (sel->evaluate == &_gmx_sel_evaluate_subexpr)
        fprintf(fp, "subexpr");
    else if (sel->evaluate == &_gmx_sel_evaluate_subexprref_pass)
        fprintf(fp, "ref_pass");
    else if (sel->evaluate == &_gmx_sel_evaluate_subexprref)
        fprintf(fp, "ref");
    else if (sel->evaluate == &_gmx_sel_evaluate_method)
        fprintf(fp, "method");
    else if (sel->evaluate == &_gmx_sel_evaluate_modifier)
        fprintf(fp, "mod");
    else if (sel->evaluate == &_gmx_sel_evaluate_not)
        fprintf(fp, "not");
    else if (sel->evaluate == &_gmx_sel_evaluate_and)
        fprintf(fp, "and");
    else if (sel->evaluate == &_gmx_sel_evaluate_or)
        fprintf(fp, "or");
    else
        fprintf(fp, "%p", (void*)(sel->evaluate));
}

/*!
 * \param[in] fp      File handle to receive the output.
 * \param[in] sel     Root of the selection subtree to print.
 * \param[in] bValues If TRUE, the evaluated values of selection elements
 *   are printed as well.
 * \param[in] level   Indentation level, starting from zero.
 */
void
_gmx_selelem_print_tree(FILE *fp, t_selelem *sel, bool bValues, int level)
{
    t_selelem *child;
    int          i;

    fprintf(fp, "%*c %s %s", level*2+1, '*',
            _gmx_selelem_type_str(sel), _gmx_sel_value_type_str(&sel->v));
    if (sel->name)
    {
        fprintf(fp, " \"%s\"", sel->name);
    }
    fprintf(fp, " flg=");
    if (sel->flags & SEL_FLAGSSET)
    {
        fprintf(fp, "s");
    }
    if (sel->flags & SEL_SINGLEVAL)
    {
        fprintf(fp, "S");
    }
    if (sel->flags & SEL_ATOMVAL)
    {
        fprintf(fp, "A");
    }
    if (sel->flags & SEL_VARNUMVAL)
    {
        fprintf(fp, "V");
    }
    if (sel->flags & SEL_DYNAMIC)
    {
        fprintf(fp, "D");
    }
    if (!(sel->flags & SEL_VALFLAGMASK))
    {
        fprintf(fp, "0");
    }
    if (sel->type == SEL_CONST)
    {
        if (sel->v.type == INT_VALUE)
        {
            fprintf(fp, " %d", sel->v.u.i[0]);
        }
        else if (sel->v.type == REAL_VALUE)
        {
            fprintf(fp, " %f", sel->v.u.r[0]);
        }
        else if (sel->v.type == GROUP_VALUE)
        {
            gmx_ana_index_t *g = sel->v.u.g;
            if (!g || g->isize == 0)
                g = &sel->u.cgrp;
            fprintf(fp, " (%d atoms)", g->isize);
        }
    }
    else if (sel->type == SEL_BOOLEAN)
    {
        fprintf(fp, " %s", _gmx_selelem_boolean_type_str(sel));
    }
    else if (sel->type == SEL_EXPRESSION
             && sel->u.expr.method->name == sm_compare.name)
    {
        _gmx_selelem_print_compare_info(fp, sel->u.expr.mdata);
    }
    if (sel->evaluate)
    {
        print_evaluation_func(fp, sel);
    }
    if (sel->refcount > 1)
    {
        fprintf(fp, " refc=%d", sel->refcount);
    }
    if (!(sel->flags & SEL_ALLOCVAL))
    {
        fprintf(fp, " (ext. output)");
    }
    fprintf(fp, "\n");

    if ((sel->type == SEL_CONST && sel->v.type == GROUP_VALUE) || sel->type == SEL_ROOT)
    {
        gmx_ana_index_t *g = sel->v.u.g;
        if (!g || g->isize == 0 || sel->evaluate != NULL)
        {
            g = &sel->u.cgrp;
        }
        if (g->isize > 0)
        {
            fprintf(fp, "%*c group:", level*2+1, ' ');
            if (g->isize <= 20)
            {
                for (i = 0; i < g->isize; ++i)
                {
                    fprintf(fp, " %d", g->index[i] + 1);
                }
            }
            else
            {
                fprintf(fp, " %d atoms", g->isize);
            }
            fprintf(fp, "\n");
        }
    }
    else if (sel->type == SEL_EXPRESSION)
    {
        if (sel->u.expr.pc)
        {
            fprintf(fp, "%*c COM", level*2+3, '*');
            fprintf(fp, "\n");
        }
    }

    if (bValues && sel->type != SEL_CONST && sel->type != SEL_ROOT && sel->v.u.ptr)
    {
        fprintf(fp, "%*c value: ", level*2+1, ' ');
        switch (sel->v.type)
        {
            case POS_VALUE:
                fprintf(fp, "(%f, %f, %f)",
                        sel->v.u.p->x[0][XX], sel->v.u.p->x[0][YY], sel->v.u.p->x[0][ZZ]);
                break;
            case GROUP_VALUE:
                fprintf(fp, "%d atoms", sel->v.u.g->isize);
                if (sel->v.u.g->isize < 20)
                {
                    if (sel->v.u.g->isize > 0)
                    {
                        fprintf(fp, ":");
                    }
                    for (i = 0; i < sel->v.u.g->isize; ++i)
                    {
                        fprintf(fp, " %d", sel->v.u.g->index[i] + 1);
                    }
                }
                break;
            default:
                fprintf(fp, "???");
                break;
        }
        fprintf(fp, "\n");
    }

    /* Print the subexpressions with one more level of indentation */
    child = sel->child;
    while (child)
    {
        if (!(sel->type == SEL_SUBEXPRREF && child->type == SEL_SUBEXPR))
        {
            _gmx_selelem_print_tree(fp, child, bValues, level+1);
        }
        child = child->next;
    }
}

/*!
 * \param[in] root Root of the subtree to query.
 * \returns TRUE if \p root or any any of its elements require topology
 *   information, FALSE otherwise.
 */
bool
_gmx_selelem_requires_top(t_selelem *root)
{
    t_selelem *child;

    if (root->type == SEL_EXPRESSION || root->type == SEL_MODIFIER)
    {
        if (root->u.expr.method && (root->u.expr.method->flags & SMETH_REQTOP))
        {
            return TRUE;
        }
        if (root->u.expr.pc && gmx_ana_poscalc_requires_top(root->u.expr.pc))
        {
            return TRUE;
        }
    }
    child = root->child;
    while (child)
    {
        if (_gmx_selelem_requires_top(child))
        {
            return TRUE;
        }
        child = child->next;
    }
    return FALSE;
}