2 * Copyright (c) 1995, the EUROPAGATE consortium (see below).
4 * The EUROPAGATE consortium members are:
6 * University College Dublin
7 * Danmarks Teknologiske Videnscenter
8 * An Chomhairle Leabharlanna
9 * Consejo Superior de Investigaciones Cientificas
11 * Permission to use, copy, modify, distribute, and sell this software and
12 * its documentation, in whole or in part, for any purpose, is hereby granted,
15 * 1. This copyright and permission notice appear in all copies of the
16 * software and its documentation. Notices of copyright or attribution
17 * which appear at the beginning of any file must remain unchanged.
19 * 2. The names of EUROPAGATE or the project partners may not be used to
20 * endorse or promote products derived from this software without specific
21 * prior written permission.
23 * 3. Users of this software (implementors and gateway operators) agree to
24 * inform the EUROPAGATE consortium of their use of the software. This
25 * information will be used to evaluate the EUROPAGATE project and the
26 * software, and to plan further developments. The consortium may use
27 * the information in later publications.
29 * 4. Users of this software agree to make their best efforts, when
30 * documenting their use of the software, to acknowledge the EUROPAGATE
31 * consortium, and the role played by the software in their work.
33 * THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT WARRANTY OF ANY KIND,
34 * EXPRESS, IMPLIED, OR OTHERWISE, INCLUDING WITHOUT LIMITATION, ANY
35 * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
36 * IN NO EVENT SHALL THE EUROPAGATE CONSORTIUM OR ITS MEMBERS BE LIABLE
37 * FOR ANY SPECIAL, INCIDENTAL, INDIRECT OR CONSEQUENTIAL DAMAGES OF
38 * ANY KIND, OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA
39 * OR PROFITS, WHETHER OR NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND
40 * ON ANY THEORY OF LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE
41 * USE OR PERFORMANCE OF THIS SOFTWARE.
44 /* CCL find (to rpn conversion)
48 * Revision 1.23 2001-03-20 11:22:58 adam
49 * CCL Truncation character may be defined.
51 * Revision 1.22 2001/03/07 13:24:40 adam
52 * Member and_not in Z_Operator is kept for backwards compatibility.
53 * Added support for definition of CCL operators in field spec file.
55 * Revision 1.21 2001/02/21 13:46:53 adam
58 * Revision 1.20 2000/11/16 13:03:12 adam
59 * Function ccl_rpn_query sets attributeSet to Bib-1.
61 * Revision 1.19 2000/11/16 09:58:02 adam
62 * Implemented local AttributeSet setting for CCL field maps.
64 * Revision 1.18 2000/10/17 19:50:28 adam
65 * Implemented and-list and or-list for CCL module.
67 * Revision 1.17 2000/05/01 09:36:50 adam
68 * Range operator only treated in ordered ranges so that minus (-) can be
69 * used for, say, the and-not operator.
71 * Revision 1.16 2000/03/14 09:06:11 adam
72 * Added POSIX threads support for frontend server.
74 * Revision 1.15 2000/02/24 23:49:13 adam
75 * Fixed memory allocation problem.
77 * Revision 1.14 2000/01/31 13:15:21 adam
78 * Removed uses of assert(3). Cleanup of ODR. CCL parser update so
79 * that some characters are not surrounded by spaces in resulting term.
82 * Revision 1.13 1999/12/22 13:13:32 adam
83 * Search terms may include "operators" without causing error.
85 * Revision 1.12 1999/11/30 13:47:11 adam
86 * Improved installation. Moved header files to include/yaz.
88 * Revision 1.11 1999/03/31 11:15:37 adam
89 * Fixed memory leaks in ccl_find_str and ccl_qual_rm.
91 * Revision 1.10 1998/02/11 11:53:33 adam
92 * Changed code so that it compiles as C++.
94 * Revision 1.9 1997/09/29 08:56:37 adam
95 * Changed CCL parser to be thread safe. New type, CCL_parser, declared
96 * and a create/destructers ccl_parser_create/ccl_parser/destory has
99 * Revision 1.8 1997/09/01 08:48:11 adam
100 * New windows NT/95 port using MSV5.0. Only a few changes made
103 * Revision 1.7 1997/05/14 06:53:26 adam
106 * Revision 1.6 1997/04/30 08:52:06 quinn
109 * Revision 1.5 1996/10/11 15:00:24 adam
110 * CCL parser from Europagate Email gateway 1.0.
112 * Revision 1.16 1996/01/08 08:41:13 adam
113 * Removed unused function.
115 * Revision 1.15 1995/07/20 08:14:34 adam
116 * Qualifiers were observed too often. Instead tokens are treated as
117 * qualifiers only when separated by comma.
119 * Revision 1.14 1995/05/16 09:39:26 adam
122 * Revision 1.13 1995/04/17 09:31:42 adam
123 * Improved handling of qualifiers. Aliases or reserved words.
125 * Revision 1.12 1995/03/20 15:27:43 adam
128 * Revision 1.11 1995/02/23 08:31:59 adam
131 * Revision 1.9 1995/02/16 13:20:06 adam
134 * Revision 1.8 1995/02/14 19:59:42 adam
135 * Removed a syntax error.
137 * Revision 1.7 1995/02/14 19:55:10 adam
138 * Header files ccl.h/cclp.h are gone! They have been merged an
139 * moved to ../include/ccl.h.
140 * Node kind(s) in ccl_rpn_node have changed names.
142 * Revision 1.6 1995/02/14 16:20:55 adam
143 * Qualifiers are read from a file now.
145 * Revision 1.5 1995/02/14 14:12:41 adam
146 * Ranges for ordered qualfiers implemented (e.g. pd=1980-1990).
148 * Revision 1.4 1995/02/14 13:16:29 adam
149 * Left and/or right truncation implemented.
151 * Revision 1.3 1995/02/14 10:25:56 adam
152 * The constructions 'qualifier rel term ...' implemented.
154 * Revision 1.2 1995/02/13 15:15:07 adam
155 * Added handling of qualifiers. Not finished yet.
157 * Revision 1.1 1995/02/13 12:35:20 adam
158 * First version of CCL. Qualifiers aren't handled yet.
167 /* returns type of current lookahead */
168 #define KIND (cclp->look_token->kind)
170 /* move one token forward */
171 #define ADVANCE cclp->look_token = cclp->look_token->next
174 * qual_val_type: test for existance of attribute type/value pair.
175 * qa: Attribute array
176 * type: Type of attribute to search for
177 * value: Value of attribute to seach for
178 * return: 1 if found; 0 otherwise.
180 static int qual_val_type (struct ccl_rpn_attr **qa, int type, int value,
184 struct ccl_rpn_attr *q;
188 for (i = 0; (q=qa[i]); i++)
191 if (q->type == type && q->value == value)
203 * strxcat: concatenate strings.
204 * n: Null-terminated Destination string
205 * src: Source string to be appended (not null-terminated)
206 * len: Length of source string.
208 static void strxcat (char *n, const char *src, int len)
218 * copy_token_name: Return copy of CCL token name
219 * tp: Pointer to token info.
220 * return: malloc(3) allocated copy of token name.
222 static char *copy_token_name (struct ccl_token *tp)
224 char *str = (char *)malloc (tp->len + 1);
226 memcpy (str, tp->name, tp->len);
232 * mk_node: Create RPN node.
233 * kind: Type of node.
234 * return: pointer to allocated node.
236 static struct ccl_rpn_node *mk_node (int kind)
238 struct ccl_rpn_node *p;
239 p = (struct ccl_rpn_node *)malloc (sizeof(*p));
246 * ccl_rpn_delete: Delete RPN tree.
247 * rpn: Pointer to tree.
249 void ccl_rpn_delete (struct ccl_rpn_node *rpn)
251 struct ccl_rpn_attr *attr, *attr1;
259 ccl_rpn_delete (rpn->u.p[0]);
260 ccl_rpn_delete (rpn->u.p[1]);
263 free (rpn->u.t.term);
264 for (attr = rpn->u.t.attr_list; attr; attr = attr1)
273 free (rpn->u.setname);
276 ccl_rpn_delete (rpn->u.p[0]);
277 ccl_rpn_delete (rpn->u.p[1]);
283 static struct ccl_rpn_node *find_spec (CCL_parser cclp,
284 struct ccl_rpn_attr **qa);
286 static int is_term_ok (int look, int *list)
288 for (;*list >= 0; list++)
294 static struct ccl_rpn_node *search_terms (CCL_parser cclp,
295 struct ccl_rpn_attr **qa);
298 * add_attr: Add attribute (type/value) to RPN term node.
299 * p: RPN node of type term.
300 * type: Type of attribute
301 * value: Value of attribute
302 * set: Attribute set name
304 static void add_attr (struct ccl_rpn_node *p, const char *set,
307 struct ccl_rpn_attr *n;
309 n = (struct ccl_rpn_attr *)malloc (sizeof(*n));
313 n->set = (char*) malloc (strlen(set)+1);
314 strcpy (n->set, set);
320 n->next = p->u.t.attr_list;
321 p->u.t.attr_list = n;
325 * search_term: Parse CCL search term.
327 * qa: Qualifier attributes already applied.
328 * return: pointer to node(s); NULL on error.
330 static struct ccl_rpn_node *search_term_x (CCL_parser cclp,
331 struct ccl_rpn_attr **qa,
334 struct ccl_rpn_attr *qa_tmp[2];
335 struct ccl_rpn_node *p_top = 0;
336 struct ccl_token *lookahead = cclp->look_token;
340 const char *truncation_aliases;
343 ccl_qual_search_special(cclp->bibset, "truncation");
344 if (!truncation_aliases)
345 truncation_aliases = "?";
349 /* no qualifier(s) applied. Use 'term' if it is defined */
353 qa[0] = ccl_qual_search (cclp, "term", 4);
356 if (qual_val_type (qa, CCL_BIB1_STR, CCL_BIB1_STR_AND_LIST, 0))
358 if (qual_val_type (qa, CCL_BIB1_STR, CCL_BIB1_STR_OR_LIST, 0))
362 struct ccl_rpn_node *p;
367 int relation_value = -1;
368 int position_value = -1;
369 int structure_value = -1;
370 int truncation_value = -1;
371 int completeness_value = -1;
374 if (and_list || or_list)
377 /* go through each TERM token. If no truncation attribute is yet
378 met, then look for left/right truncation markers (?) and
379 set left_trunc/right_trunc/mid_trunc accordingly */
380 for (no = 0; no < max && is_term_ok(lookahead->kind, term_list); no++)
382 for (i = 0; i<lookahead->len; i++)
383 if (truncation_value == -1 && strchr(truncation_aliases,
386 if (no == 0 && i == 0 && lookahead->len >= 1)
388 else if (!is_term_ok(lookahead->next->kind, term_list) &&
389 i == lookahead->len-1 && i >= 1)
394 len += 1+lookahead->len;
395 lookahead = lookahead->next;
399 break; /* no more terms . stop . */
404 p = mk_node (CCL_RPN_OR);
406 p = mk_node (CCL_RPN_AND);
408 p = mk_node (CCL_RPN_AND);
413 /* create the term node, but wait a moment before adding the term */
414 p = mk_node (CCL_RPN_TERM);
415 p->u.t.attr_list = NULL;
418 /* make the top node point to us.. */
425 /* go through all attributes and add them to the attribute list */
426 for (i=0; qa && qa[i]; i++)
428 struct ccl_rpn_attr *attr;
430 for (attr = qa[i]; attr; attr = attr->next)
432 { /* deal only with REAL attributes (positive) */
436 if (relation_value != -1)
438 relation_value = attr->value;
441 if (position_value != -1)
443 position_value = attr->value;
446 if (structure_value != -1)
448 structure_value = attr->value;
451 if (truncation_value != -1)
453 truncation_value = attr->value;
454 left_trunc = right_trunc = mid_trunc = 0;
457 if (completeness_value != -1)
459 completeness_value = attr->value;
462 add_attr (p, attr->set, attr->type, attr->value);
465 /* len now holds the number of characters in the RPN term */
466 /* no holds the number of CCL tokens (1 or more) */
468 if (structure_value == -1 &&
469 qual_val_type (qa, CCL_BIB1_STR, CCL_BIB1_STR_WP, &attset))
470 { /* no structure attribute met. Apply either structure attribute
471 WORD or PHRASE depending on number of CCL tokens */
473 add_attr (p, attset, CCL_BIB1_STR, 2);
475 add_attr (p, attset, CCL_BIB1_STR, 1);
478 /* make the RPN token */
479 p->u.t.term = (char *)malloc (len);
480 ccl_assert (p->u.t.term);
481 p->u.t.term[0] = '\0';
482 for (i = 0; i<no; i++)
484 const char *src_str = cclp->look_token->name;
485 int src_len = cclp->look_token->len;
487 if (i == 0 && left_trunc)
492 else if (i == no-1 && right_trunc)
496 int len = strlen(p->u.t.term);
498 !strchr("-+", *src_str) &&
499 !strchr("-+", p->u.t.term[len-1]))
501 strcat (p->u.t.term, " ");
504 strxcat (p->u.t.term, src_str, src_len);
507 if (left_trunc && right_trunc)
509 if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_BOTH,
512 cclp->error_code = CCL_ERR_TRUNC_NOT_BOTH;
516 add_attr (p, attset, CCL_BIB1_TRU, 3);
518 else if (right_trunc)
520 if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_RIGHT,
523 cclp->error_code = CCL_ERR_TRUNC_NOT_RIGHT;
527 add_attr (p, attset, CCL_BIB1_TRU, 1);
531 if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_LEFT,
534 cclp->error_code = CCL_ERR_TRUNC_NOT_LEFT;
538 add_attr (p, attset, CCL_BIB1_TRU, 2);
542 if (qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_NONE,
544 add_attr (p, attset, CCL_BIB1_TRU, 100);
548 cclp->error_code = CCL_ERR_TERM_EXPECTED;
552 static struct ccl_rpn_node *search_term (CCL_parser cclp,
553 struct ccl_rpn_attr **qa)
555 static int list[] = {CCL_TOK_TERM, CCL_TOK_COMMA, -1};
556 return search_term_x(cclp, qa, list);
560 * qualifiers: Parse CCL qualifiers and search terms.
562 * la: Token pointer to RELATION token.
563 * qa: Qualifier attributes already applied.
564 * return: pointer to node(s); NULL on error.
566 static struct ccl_rpn_node *qualifiers (CCL_parser cclp, struct ccl_token *la,
567 struct ccl_rpn_attr **qa)
569 struct ccl_token *lookahead = cclp->look_token;
570 struct ccl_rpn_attr **ap;
577 cclp->error_code = CCL_ERR_DOUBLE_QUAL;
581 for (lookahead = cclp->look_token; lookahead != la;
582 lookahead=lookahead->next)
585 for (i=0; qa[i]; i++)
587 ap = (struct ccl_rpn_attr **)malloc ((no+1) * sizeof(*ap));
589 for (i = 0; cclp->look_token != la; i++)
591 ap[i] = ccl_qual_search (cclp, cclp->look_token->name,
592 cclp->look_token->len);
595 cclp->error_code = CCL_ERR_UNKNOWN_QUAL;
600 if (KIND == CCL_TOK_COMMA)
607 if (!qual_val_type(ap, CCL_BIB1_REL, CCL_BIB1_REL_ORDER, &attset))
609 /* unordered relation */
610 struct ccl_rpn_node *p;
611 if (KIND != CCL_TOK_EQ)
613 cclp->error_code = CCL_ERR_EQ_EXPECTED;
618 if (KIND == CCL_TOK_LP)
621 if (!(p = find_spec (cclp, ap)))
626 if (KIND != CCL_TOK_RP)
628 cclp->error_code = CCL_ERR_RP_EXPECTED;
636 p = search_terms (cclp, ap);
640 /* ordered relation ... */
642 if (cclp->look_token->len == 1)
644 if (cclp->look_token->name[0] == '<')
646 else if (cclp->look_token->name[0] == '=')
648 else if (cclp->look_token->name[0] == '>')
651 else if (cclp->look_token->len == 2)
653 if (!memcmp (cclp->look_token->name, "<=", 2))
655 else if (!memcmp (cclp->look_token->name, ">=", 2))
657 else if (!memcmp (cclp->look_token->name, "<>", 2))
661 cclp->error_code = CCL_ERR_BAD_RELATION;
664 struct ccl_rpn_node *p;
666 ADVANCE; /* skip relation */
667 if (KIND == CCL_TOK_TERM &&
668 cclp->look_token->next->len == 1 &&
669 cclp->look_token->next->name[0] == '-')
671 struct ccl_rpn_node *p1;
672 if (!(p1 = search_term (cclp, ap)))
677 ADVANCE; /* skip '-' */
678 if (KIND == CCL_TOK_TERM) /* = term - term ? */
680 struct ccl_rpn_node *p2;
682 if (!(p2 = search_term (cclp, ap)))
688 p = mk_node (CCL_RPN_AND);
690 add_attr (p1, attset, CCL_BIB1_REL, 4);
692 add_attr (p2, attset, CCL_BIB1_REL, 2);
698 add_attr (p1, attset, CCL_BIB1_REL, 4);
703 else if (cclp->look_token->len == 1 &&
704 cclp->look_token->name[0] == '"') /* = - term ? */
707 if (!(p = search_term (cclp, ap)))
712 add_attr (p, attset, CCL_BIB1_REL, 2);
716 else if (KIND == CCL_TOK_LP)
719 if (!(p = find_spec (cclp, ap)))
724 if (KIND != CCL_TOK_RP)
726 cclp->error_code = CCL_ERR_RP_EXPECTED;
737 if (!(p = search_terms (cclp, ap)))
742 add_attr (p, attset, CCL_BIB1_REL, rel);
746 cclp->error_code = CCL_ERR_TERM_EXPECTED;
753 * search_terms: Parse CCL search terms - including proximity.
755 * qa: Qualifier attributes already applied.
756 * return: pointer to node(s); NULL on error.
758 static struct ccl_rpn_node *search_terms (CCL_parser cclp,
759 struct ccl_rpn_attr **qa)
761 static int list[] = {
762 CCL_TOK_TERM, CCL_TOK_COMMA,CCL_TOK_EQ, CCL_TOK_REL, -1};
763 struct ccl_rpn_node *p1, *p2, *pn;
764 p1 = search_term_x (cclp, qa, list);
769 if (KIND == CCL_TOK_PROX)
772 p2 = search_term_x (cclp, qa, list);
778 pn = mk_node (CCL_RPN_PROX);
783 else if (is_term_ok(KIND, list))
785 p2 = search_term_x (cclp, qa, list);
791 pn = mk_node (CCL_RPN_PROX);
803 * search_elements: Parse CCL search elements
805 * qa: Qualifier attributes already applied.
806 * return: pointer to node(s); NULL on error.
808 static struct ccl_rpn_node *search_elements (CCL_parser cclp,
809 struct ccl_rpn_attr **qa)
811 struct ccl_rpn_node *p1;
812 struct ccl_token *lookahead;
813 if (KIND == CCL_TOK_LP)
816 p1 = find_spec (cclp, qa);
819 if (KIND != CCL_TOK_RP)
821 cclp->error_code = CCL_ERR_RP_EXPECTED;
828 else if (KIND == CCL_TOK_SET)
831 if (KIND == CCL_TOK_EQ)
833 if (KIND != CCL_TOK_TERM)
835 cclp->error_code = CCL_ERR_SETNAME_EXPECTED;
838 p1 = mk_node (CCL_RPN_SET);
839 p1->u.setname = copy_token_name (cclp->look_token);
843 lookahead = cclp->look_token;
845 while (lookahead->kind==CCL_TOK_TERM)
847 lookahead = lookahead->next;
848 if (lookahead->kind == CCL_TOK_REL || lookahead->kind == CCL_TOK_EQ)
849 return qualifiers (cclp, lookahead, qa);
850 if (lookahead->kind != CCL_TOK_COMMA)
852 lookahead = lookahead->next;
854 return search_terms (cclp, qa);
858 * find_spec: Parse CCL find specification
860 * qa: Qualifier attributes already applied.
861 * return: pointer to node(s); NULL on error.
863 static struct ccl_rpn_node *find_spec (CCL_parser cclp,
864 struct ccl_rpn_attr **qa)
866 struct ccl_rpn_node *p1, *p2, *pn;
867 if (!(p1 = search_elements (cclp, qa)))
875 p2 = search_elements (cclp, qa);
881 pn = mk_node (CCL_RPN_AND);
888 p2 = search_elements (cclp, qa);
894 pn = mk_node (CCL_RPN_OR);
901 p2 = search_elements (cclp, qa);
907 pn = mk_node (CCL_RPN_NOT);
918 struct ccl_rpn_node *ccl_parser_find (CCL_parser cclp, struct ccl_token *list)
920 struct ccl_rpn_node *p;
924 cclp->look_token = list;
925 p = find_spec (cclp, NULL);
926 if (p && KIND != CCL_TOK_EOL)
928 if (KIND == CCL_TOK_RP)
929 cclp->error_code = CCL_ERR_BAD_RP;
931 cclp->error_code = CCL_ERR_OP_EXPECTED;
935 cclp->error_pos = cclp->look_token->name;
937 cclp->error_code = CCL_ERR_OK;
939 cclp->error_code = cclp->error_code;
944 * ccl_find: Parse CCL find - token representation
945 * bibset: Bibset to be used for the parsing
946 * list: List of tokens
947 * error: Pointer to integer. Holds error no. on completion.
948 * pos: Pointer to char position. Holds approximate error position.
949 * return: RPN tree on successful completion; NULL otherwise.
951 struct ccl_rpn_node *ccl_find (CCL_bibset bibset, struct ccl_token *list,
952 int *error, const char **pos)
954 struct ccl_rpn_node *p;
955 CCL_parser cclp = ccl_parser_create ();
957 cclp->bibset = bibset;
959 p = ccl_parser_find (cclp, list);
961 *error = cclp->error_code;
962 *pos = cclp->error_pos;
964 ccl_parser_destroy (cclp);
970 * ccl_find_str: Parse CCL find - string representation
971 * bibset: Bibset to be used for the parsing
972 * str: String to be parsed
973 * error: Pointer to integer. Holds error no. on completion.
974 * pos: Pointer to char position. Holds approximate error position.
975 * return: RPN tree on successful completion; NULL otherwise.
977 struct ccl_rpn_node *ccl_find_str (CCL_bibset bibset, const char *str,
978 int *error, int *pos)
980 CCL_parser cclp = ccl_parser_create ();
981 struct ccl_token *list;
982 struct ccl_rpn_node *p;
984 cclp->bibset = bibset;
986 list = ccl_parser_tokenize (cclp, str);
987 p = ccl_parser_find (cclp, list);
989 *error = cclp->error_code;
991 *pos = cclp->error_pos - str;
992 ccl_parser_destroy (cclp);
993 ccl_token_del (list);