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 1.16 1996/01/08 08:41:13 adam
49 * Removed unused function.
51 * Revision 1.15 1995/07/20 08:14:34 adam
52 * Qualifiers were observed too often. Instead tokens are treated as
53 * qualifiers only when separated by comma.
55 * Revision 1.14 1995/05/16 09:39:26 adam
58 * Revision 1.13 1995/04/17 09:31:42 adam
59 * Improved handling of qualifiers. Aliases or reserved words.
61 * Revision 1.12 1995/03/20 15:27:43 adam
64 * Revision 1.11 1995/02/23 08:31:59 adam
67 * Revision 1.9 1995/02/16 13:20:06 adam
70 * Revision 1.8 1995/02/14 19:59:42 adam
71 * Removed a syntax error.
73 * Revision 1.7 1995/02/14 19:55:10 adam
74 * Header files ccl.h/cclp.h are gone! They have been merged an
75 * moved to ../include/ccl.h.
76 * Node kind(s) in ccl_rpn_node have changed names.
78 * Revision 1.6 1995/02/14 16:20:55 adam
79 * Qualifiers are read from a file now.
81 * Revision 1.5 1995/02/14 14:12:41 adam
82 * Ranges for ordered qualfiers implemented (e.g. pd=1980-1990).
84 * Revision 1.4 1995/02/14 13:16:29 adam
85 * Left and/or right truncation implemented.
87 * Revision 1.3 1995/02/14 10:25:56 adam
88 * The constructions 'qualifier rel term ...' implemented.
90 * Revision 1.2 1995/02/13 15:15:07 adam
91 * Added handling of qualifiers. Not finished yet.
93 * Revision 1.1 1995/02/13 12:35:20 adam
94 * First version of CCL. Qualifiers aren't handled yet.
105 /* current lookahead token */
106 static struct ccl_token *look_token;
108 /* holds error no if error occur */
109 static int ccl_error;
112 static CCL_bibset bibset;
114 /* returns type of current lookahead */
115 #define KIND (look_token->kind)
117 /* move one token forward */
118 #define ADVANCE look_token = look_token->next
121 * qual_val_type: test for existance of attribute type/value pair.
122 * qa: Attribute array
123 * type: Type of attribute to search for
124 * value: Value of attribute to seach for
125 * return: 1 if found; 0 otherwise.
127 static int qual_val_type (struct ccl_rpn_attr **qa, int type, int value)
130 struct ccl_rpn_attr *q;
134 for (i = 0; (q=qa[i]); i++)
137 if (q->type == type && q->value == value)
145 * strxcat: concatenate strings.
146 * n: Null-terminated Destination string
147 * src: Source string to be appended (not null-terminated)
148 * len: Length of source string.
150 static void strxcat (char *n, const char *src, int len)
160 * copy_token_name: Return copy of CCL token name
161 * tp: Pointer to token info.
162 * return: malloc(3) allocated copy of token name.
164 static char *copy_token_name (struct ccl_token *tp)
166 char *str = malloc (tp->len + 1);
168 memcpy (str, tp->name, tp->len);
174 * mk_node: Create RPN node.
175 * kind: Type of node.
176 * return: pointer to allocated node.
178 static struct ccl_rpn_node *mk_node (enum rpn_node_kind kind)
180 struct ccl_rpn_node *p;
181 p = malloc (sizeof(*p));
188 * ccl_rpn_delete: Delete RPN tree.
189 * rpn: Pointer to tree.
191 void ccl_rpn_delete (struct ccl_rpn_node *rpn)
193 struct ccl_rpn_attr *attr, *attr1;
201 ccl_rpn_delete (rpn->u.p[0]);
202 ccl_rpn_delete (rpn->u.p[1]);
205 free (rpn->u.t.term);
206 for (attr = rpn->u.t.attr_list; attr; attr = attr1)
213 free (rpn->u.setname);
216 ccl_rpn_delete (rpn->u.p[0]);
217 ccl_rpn_delete (rpn->u.p[1]);
223 static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa);
224 static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa);
227 * add_attr: Add attribute (type/value) to RPN term node.
228 * p: RPN node of type term.
229 * type: Type of attribute
230 * value: Value of attribute
232 static void add_attr (struct ccl_rpn_node *p, int type, int value)
234 struct ccl_rpn_attr *n;
236 n = malloc (sizeof(*n));
240 n->next = p->u.t.attr_list;
241 p->u.t.attr_list = n;
245 * search_term: Parse CCL search term.
246 * qa: Qualifier attributes already applied.
247 * return: pointer to node(s); NULL on error.
249 static struct ccl_rpn_node *search_term (struct ccl_rpn_attr **qa)
251 struct ccl_rpn_node *p;
252 struct ccl_token *lookahead = look_token;
258 int relation_value = -1;
259 int position_value = -1;
260 int structure_value = -1;
261 int truncation_value = -1;
262 int completeness_value = -1;
264 if (KIND != CCL_TOK_TERM)
266 ccl_error = CCL_ERR_TERM_EXPECTED;
269 /* create the term node, but wait a moment before adding the term */
270 p = mk_node (CCL_RPN_TERM);
271 p->u.t.attr_list = NULL;
276 /* no qualifier(s) applied. Use 'term' if it is defined */
278 qa = malloc (2*sizeof(*qa));
280 qa[0] = ccl_qual_search (bibset, "term", 4);
284 /* go through all attributes and add them to the attribute list */
285 for (i=0; qa && qa[i]; i++)
287 struct ccl_rpn_attr *attr;
289 for (attr = qa[i]; attr; attr = attr->next)
291 { /* deal only with REAL attributes (positive) */
295 if (relation_value != -1)
297 relation_value = attr->value;
300 if (position_value != -1)
302 position_value = attr->value;
305 if (structure_value != -1)
307 structure_value = attr->value;
310 if (truncation_value != -1)
312 truncation_value = attr->value;
315 if (completeness_value != -1)
317 completeness_value = attr->value;
320 add_attr (p, attr->type, attr->value);
323 /* go through each TERM token. If no truncation attribute is yet
324 met, then look for left/right truncation markers (?) and
325 set left_trunc/right_trunc/mid_trunc accordingly */
326 for (no = 0; lookahead->kind == CCL_TOK_TERM; no++)
328 for (i = 0; i<lookahead->len; i++)
329 if (truncation_value == -1 && lookahead->name[i] == '?')
331 if (no == 0 && i == 0 && lookahead->len >= 1)
333 else if (lookahead->next->kind != CCL_TOK_TERM &&
334 i == lookahead->len-1 && i >= 1)
339 len += 1+lookahead->len;
340 lookahead = lookahead->next;
342 /* len now holds the number of characters in the RPN term */
343 /* no holds the number of CCL tokens (1 or more) */
345 if (structure_value == -1 &&
346 qual_val_type (qa, CCL_BIB1_STR, CCL_BIB1_STR_WP))
347 { /* no structure attribute met. Apply either structure attribute
348 WORD or PHRASE depending on number of CCL tokens */
350 add_attr (p, CCL_BIB1_STR, 2);
352 add_attr (p, CCL_BIB1_STR, 1);
355 /* make the RPN token */
356 p->u.t.term = malloc (len);
357 assert (p->u.t.term);
358 p->u.t.term[0] = '\0';
359 for (i = 0; i<no; i++)
361 const char *src_str = look_token->name;
362 int src_len = look_token->len;
364 if (i == 0 && left_trunc)
369 else if (i == no-1 && right_trunc)
372 strcat (p->u.t.term, " ");
373 strxcat (p->u.t.term, src_str, src_len);
376 if (left_trunc && right_trunc)
378 if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_BOTH))
380 ccl_error = CCL_ERR_TRUNC_NOT_BOTH;
385 add_attr (p, CCL_BIB1_TRU, 3);
387 else if (right_trunc)
389 if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_RIGHT))
391 ccl_error = CCL_ERR_TRUNC_NOT_RIGHT;
396 add_attr (p, CCL_BIB1_TRU, 1);
400 if (!qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_LEFT))
402 ccl_error = CCL_ERR_TRUNC_NOT_LEFT;
407 add_attr (p, CCL_BIB1_TRU, 2);
411 if (qual_val_type (qa, CCL_BIB1_TRU, CCL_BIB1_TRU_CAN_NONE))
412 add_attr (p, CCL_BIB1_TRU, 100);
418 * qualifiers: Parse CCL qualifiers and search terms.
419 * la: Token pointer to RELATION token.
420 * qa: Qualifier attributes already applied.
421 * return: pointer to node(s); NULL on error.
423 static struct ccl_rpn_node *qualifiers (struct ccl_token *la,
424 struct ccl_rpn_attr **qa)
426 struct ccl_token *lookahead = look_token;
427 struct ccl_rpn_attr **ap;
433 ccl_error = CCL_ERR_DOUBLE_QUAL;
437 for (lookahead = look_token; lookahead != la; lookahead=lookahead->next)
440 for (i=0; qa[i]; i++)
442 ap = malloc ((no+1) * sizeof(*ap));
444 for (i = 0; look_token != la; i++)
446 ap[i] = ccl_qual_search (bibset, look_token->name, look_token->len);
449 ccl_error = CCL_ERR_UNKNOWN_QUAL;
454 if (KIND == CCL_TOK_COMMA)
461 if (!qual_val_type (ap, CCL_BIB1_REL, CCL_BIB1_REL_ORDER))
463 /* unordered relation */
464 struct ccl_rpn_node *p;
465 if (KIND != CCL_TOK_EQ)
467 ccl_error = CCL_ERR_EQ_EXPECTED;
472 if (KIND == CCL_TOK_LP)
475 if (!(p = find_spec (ap)))
480 if (KIND != CCL_TOK_RP)
482 ccl_error = CCL_ERR_RP_EXPECTED;
490 p = search_terms (ap);
495 if (look_token->len == 1)
497 if (look_token->name[0] == '<')
499 else if (look_token->name[0] == '=')
501 else if (look_token->name[0] == '>')
504 else if (look_token->len == 2)
506 if (!memcmp (look_token->name, "<=", 2))
508 else if (!memcmp (look_token->name, ">=", 2))
510 else if (!memcmp (look_token->name, "<>", 2))
514 ccl_error = CCL_ERR_BAD_RELATION;
517 struct ccl_rpn_node *p;
519 ADVANCE; /* skip relation */
520 if (KIND == CCL_TOK_TERM && look_token->next->kind == CCL_TOK_MINUS)
522 struct ccl_rpn_node *p1;
523 if (!(p1 = search_term (ap)))
528 ADVANCE; /* skip '-' */
529 if (KIND == CCL_TOK_TERM) /* = term - term ? */
531 struct ccl_rpn_node *p2;
533 if (!(p2 = search_term (ap)))
539 p = mk_node (CCL_RPN_AND);
541 add_attr (p1, CCL_BIB1_REL, 4);
543 add_attr (p2, CCL_BIB1_REL, 2);
549 add_attr (p1, CCL_BIB1_REL, 4);
554 else if (KIND == CCL_TOK_MINUS) /* = - term ? */
557 if (!(p = search_term (ap)))
562 add_attr (p, CCL_BIB1_REL, 2);
566 else if (KIND == CCL_TOK_LP)
569 if (!(p = find_spec (ap)))
574 if (KIND != CCL_TOK_RP)
576 ccl_error = CCL_ERR_RP_EXPECTED;
587 if (!(p = search_terms (ap)))
592 add_attr (p, CCL_BIB1_REL, rel);
596 ccl_error = CCL_ERR_TERM_EXPECTED;
603 * search_terms: Parse CCL search terms - including proximity.
604 * qa: Qualifier attributes already applied.
605 * return: pointer to node(s); NULL on error.
607 static struct ccl_rpn_node *search_terms (struct ccl_rpn_attr **qa)
609 struct ccl_rpn_node *p1, *p2, *pn;
610 p1 = search_term (qa);
615 if (KIND == CCL_TOK_PROX)
618 p2 = search_term (qa);
624 pn = mk_node (CCL_RPN_PROX);
629 else if (KIND == CCL_TOK_TERM)
631 p2 = search_term (qa);
637 pn = mk_node (CCL_RPN_PROX);
649 * search_elements: Parse CCL search elements
650 * qa: Qualifier attributes already applied.
651 * return: pointer to node(s); NULL on error.
653 static struct ccl_rpn_node *search_elements (struct ccl_rpn_attr **qa)
655 struct ccl_rpn_node *p1;
656 struct ccl_token *lookahead;
657 if (KIND == CCL_TOK_LP)
663 if (KIND != CCL_TOK_RP)
665 ccl_error = CCL_ERR_RP_EXPECTED;
672 else if (KIND == CCL_TOK_SET)
675 if (KIND == CCL_TOK_EQ)
677 if (KIND != CCL_TOK_TERM)
679 ccl_error = CCL_ERR_SETNAME_EXPECTED;
682 p1 = mk_node (CCL_RPN_SET);
683 p1->u.setname = copy_token_name (look_token);
687 lookahead = look_token;
689 while (lookahead->kind==CCL_TOK_TERM)
691 lookahead = lookahead->next;
692 if (lookahead->kind == CCL_TOK_REL || lookahead->kind == CCL_TOK_EQ)
693 return qualifiers (lookahead, qa);
694 if (lookahead->kind != CCL_TOK_COMMA)
696 lookahead = lookahead->next;
698 return search_terms (qa);
702 * find_spec: Parse CCL find specification
703 * qa: Qualifier attributes already applied.
704 * return: pointer to node(s); NULL on error.
706 static struct ccl_rpn_node *find_spec (struct ccl_rpn_attr **qa)
708 struct ccl_rpn_node *p1, *p2, *pn;
709 if (!(p1 = search_elements (qa)))
717 p2 = search_elements (qa);
723 pn = mk_node (CCL_RPN_AND);
730 p2 = search_elements (qa);
736 pn = mk_node (CCL_RPN_OR);
743 p2 = search_elements (qa);
749 pn = mk_node (CCL_RPN_NOT);
761 * ccl_find: Parse CCL find - token representation
762 * abibset: Bibset to be used for the parsing
763 * list: List of tokens
764 * error: Pointer to integer. Holds error no. on completion.
765 * pos: Pointer to char position. Holds approximate error position.
766 * return: RPN tree on successful completion; NULL otherwise.
768 struct ccl_rpn_node *ccl_find (CCL_bibset abibset, struct ccl_token *list,
769 int *error, const char **pos)
771 struct ccl_rpn_node *p;
775 p = find_spec (NULL);
776 if (p && KIND != CCL_TOK_EOL)
778 if (KIND == CCL_TOK_RP)
779 ccl_error = CCL_ERR_BAD_RP;
781 ccl_error = CCL_ERR_OP_EXPECTED;
785 *pos = look_token->name;
794 * ccl_find_str: Parse CCL find - string representation
795 * bibset: Bibset to be used for the parsing
796 * str: String to be parsed
797 * error: Pointer to integer. Holds error no. on completion.
798 * pos: Pointer to char position. Holds approximate error position.
799 * return: RPN tree on successful completion; NULL otherwise.
801 struct ccl_rpn_node *ccl_find_str (CCL_bibset bibset, const char *str,
802 int *error, int *pos)
804 struct ccl_token *list;
805 struct ccl_rpn_node *rpn;
806 const char *char_pos;
808 list = ccl_tokenize (str);
809 rpn = ccl_find (bibset, list, error, &char_pos);
811 *pos = char_pos - str;