1 /* $Id: zebramap.c,v 1.64 2007-11-05 13:58:01 adam Exp $
2 Copyright (C) 1995-2007
5 This file is part of the Zebra server.
7 Zebra is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 2, or (at your option) any later
12 Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with Zebra; see the file LICENSE.zebra. If not, write to the
19 Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
29 #include <yaz/yaz-util.h>
32 #include <yaz/icu_I18N.h>
36 #define ZEBRA_MAP_TYPE_SORT 1
37 #define ZEBRA_MAP_TYPE_INDEX 2
38 #define ZEBRA_MAP_TYPE_STATICRANK 3
40 #define ZEBRA_REPLACE_ANY 300
55 const char *maptab_name;
57 zebra_maps_t zebra_maps;
62 struct icu_chain *icu_chain;
64 struct zebra_map *next;
72 const char *temp_map_ptr[2];
79 void zebra_maps_close(zebra_maps_t zms)
81 struct zebra_map *zm = zms->map_list;
85 chrmaptab_destroy(zm->maptab);
88 icu_chain_destroy(zm->icu_chain);
95 wrbuf_destroy(zms->wrbuf_1);
96 nmem_destroy(zms->nmem);
100 zebra_map_t zebra_add_map(zebra_maps_t zms, const char *index_type,
103 zebra_map_t zm = (zebra_map_t) nmem_malloc(zms->nmem, sizeof(*zm));
105 zm->zebra_maps = zms;
106 zm->id = nmem_strdup(zms->nmem, index_type);
111 zm->completeness = 0;
113 zm->alwaysmatches = 0;
114 zm->first_in_field = 0;
117 zms->last_map->next = zm;
131 static int parse_command(zebra_maps_t zms, int argc, char **argv,
132 const char *fname, int lineno)
134 zebra_map_t zm = zms->last_map;
137 yaz_log(YLOG_WARN, "%s:%d: Missing arguments for '%s'",
138 fname, lineno, argv[0]);
143 yaz_log(YLOG_WARN, "%s:%d: Too many arguments for '%s'",
144 fname, lineno, argv[0]);
147 if (!yaz_matchstr(argv[0], "index"))
149 zm = zebra_add_map(zms, argv[1], ZEBRA_MAP_TYPE_INDEX);
152 else if (!yaz_matchstr(argv[0], "sort"))
154 zm = zebra_add_map(zms, argv[1], ZEBRA_MAP_TYPE_SORT);
155 zm->u.sort.entry_size = 80;
157 else if (!yaz_matchstr(argv[0], "staticrank"))
159 zm = zebra_add_map(zms, argv[1], ZEBRA_MAP_TYPE_STATICRANK);
160 zm->completeness = 1;
164 yaz_log(YLOG_WARN, "%s:%d: Missing sort/index before '%s'",
165 fname, lineno, argv[0]);
168 else if (!yaz_matchstr(argv[0], "charmap") && argc == 2)
170 if (zm->type != ZEBRA_MAP_TYPE_STATICRANK)
171 zm->maptab_name = nmem_strdup(zms->nmem, argv[1]);
174 yaz_log(YLOG_WARN|YLOG_FATAL, "%s:%d: charmap for "
175 "staticrank is invalid", fname, lineno);
176 yaz_log(YLOG_LOG, "Type is %d", zm->type);
180 else if (!yaz_matchstr(argv[0], "completeness") && argc == 2)
182 zm->completeness = atoi(argv[1]);
184 else if (!yaz_matchstr(argv[0], "position") && argc == 2)
186 zm->positioned = atoi(argv[1]);
188 else if (!yaz_matchstr(argv[0], "alwaysmatches") && argc == 2)
190 if (zm->type != ZEBRA_MAP_TYPE_STATICRANK)
191 zm->alwaysmatches = atoi(argv[1]);
194 yaz_log(YLOG_WARN|YLOG_FATAL, "%s:%d: alwaysmatches for "
195 "staticrank is invalid", fname, lineno);
199 else if (!yaz_matchstr(argv[0], "firstinfield") && argc == 2)
201 zm->first_in_field = atoi(argv[1]);
203 else if (!yaz_matchstr(argv[0], "entrysize") && argc == 2)
205 if (zm->type == ZEBRA_MAP_TYPE_SORT)
206 zm->u.sort.entry_size = atoi(argv[1]);
210 "%s:%d: entrysize only valid in sort section",
215 else if (!yaz_matchstr(argv[0], "locale"))
217 zm->locale = nmem_strdup(zms->nmem, argv[1]);
219 else if (!yaz_matchstr(argv[0], "icuchain"))
222 zm->doc = xmlParseFile(argv[1]);
225 yaz_log(YLOG_WARN, "%s:%d: Could not load icuchain config '%s'",
226 fname, lineno, argv[1]);
233 xmlNode *xml_node = xmlDocGetRootElement(zm->doc);
235 icu_chain_xml_config(xml_node, zm->locale,
236 zm->type == ZEBRA_MAP_TYPE_SORT,
240 yaz_log(YLOG_WARN, "%s:%d: Failed to load ICU chain %s",
241 fname, lineno, argv[1]);
244 yaz_log(YLOG_WARN, "%s:%d: ICU support unavailable",
250 yaz_log(YLOG_WARN, "%s:%d: XML support unavailable",
257 yaz_log(YLOG_WARN, "%s:%d: Unrecognized directive '%s'",
258 fname, lineno, argv[0]);
264 ZEBRA_RES zebra_maps_read_file(zebra_maps_t zms, const char *fname)
273 if (!(f = yaz_fopen(zms->tabpath, fname, "r", zms->tabroot)))
275 yaz_log(YLOG_ERRNO|YLOG_FATAL, "%s", fname);
278 while ((argc = readconf_line(f, &lineno, line, 512, argv, 10)))
280 int r = parse_command(zms, argc, argv, fname, lineno);
289 (zms->no_files_read)++;
293 zebra_maps_t zebra_maps_open(Res res, const char *base_path,
294 const char *profile_path)
296 zebra_maps_t zms = (zebra_maps_t) xmalloc(sizeof(*zms));
298 zms->nmem = nmem_create();
299 zms->tabpath = profile_path ? nmem_strdup(zms->nmem, profile_path) : 0;
302 zms->tabroot = nmem_strdup(zms->nmem, base_path);
306 zms->temp_map_str[0] = '\0';
307 zms->temp_map_str[1] = '\0';
309 zms->temp_map_ptr[0] = zms->temp_map_str;
310 zms->temp_map_ptr[1] = NULL;
312 zms->wrbuf_1 = wrbuf_alloc();
314 zms->no_files_read = 0;
318 zebra_map_t zebra_map_get(zebra_maps_t zms, const char *id)
321 for (zm = zms->map_list; zm; zm = zm->next)
322 if (!strcmp(zm->id, id))
327 zebra_map_t zebra_map_get_or_add(zebra_maps_t zms, const char *id)
329 struct zebra_map *zm = zebra_map_get(zms, id);
332 zm = zebra_add_map(zms, id, ZEBRA_MAP_TYPE_INDEX);
334 /* no reason to warn if no maps are read from file */
335 if (zms->no_files_read)
336 yaz_log(YLOG_WARN, "Unknown register type: %s", id);
338 zm->maptab_name = nmem_strdup(zms->nmem, "@");
339 zm->completeness = 0;
345 chrmaptab zebra_charmap_get(zebra_map_t zm)
349 if (!zm->maptab_name || !yaz_matchstr(zm->maptab_name, "@"))
351 if (!(zm->maptab = chrmaptab_create(zm->zebra_maps->tabpath,
353 zm->zebra_maps->tabroot)))
354 yaz_log(YLOG_WARN, "Failed to read character table %s",
357 yaz_log(YLOG_DEBUG, "Read character table %s", zm->maptab_name);
362 const char **zebra_maps_input(zebra_map_t zm,
363 const char **from, int len, int first)
365 chrmaptab maptab = zebra_charmap_get(zm);
367 return chr_map_input(maptab, from, len, first);
369 zm->zebra_maps->temp_map_str[0] = **from;
372 return zm->zebra_maps->temp_map_ptr;
375 const char **zebra_maps_search(zebra_map_t zm,
376 const char **from, int len, int *q_map_match)
381 maptab = zebra_charmap_get(zm);
385 map = chr_map_q_input(maptab, from, len, 0);
391 map = chr_map_input(maptab, from, len, 0);
395 zm->zebra_maps->temp_map_str[0] = **from;
398 return zm->zebra_maps->temp_map_ptr;
401 const char *zebra_maps_output(zebra_map_t zm,
404 chrmaptab maptab = zebra_charmap_get(zm);
407 return chr_map_output(maptab, from, 1);
411 /* ------------------------------------ */
413 int zebra_maps_is_complete(zebra_map_t zm)
416 return zm->completeness;
420 int zebra_maps_is_positioned(zebra_map_t zm)
423 return zm->positioned;
427 int zebra_maps_is_index(zebra_map_t zm)
430 return zm->type == ZEBRA_MAP_TYPE_INDEX;
434 int zebra_maps_is_staticrank(zebra_map_t zm)
437 return zm->type == ZEBRA_MAP_TYPE_STATICRANK;
441 int zebra_maps_is_sort(zebra_map_t zm)
444 return zm->type == ZEBRA_MAP_TYPE_SORT;
448 int zebra_maps_is_alwaysmatches(zebra_map_t zm)
451 return zm->alwaysmatches;
455 int zebra_maps_is_first_in_field(zebra_map_t zm)
458 return zm->first_in_field;
462 int zebra_maps_sort(zebra_maps_t zms, Z_SortAttributes *sortAttributes,
468 attr_init_AttrList(&use, sortAttributes->list, 1);
469 attr_init_AttrList(&structure, sortAttributes->list, 4);
472 structure_value = attr_find(&structure, 0);
473 if (structure_value == 109)
475 return attr_find(&use, NULL);
478 int zebra_maps_attr(zebra_maps_t zms, Z_AttributesPlusTerm *zapt,
479 const char **index_type, char **search_type, char *rank_type,
480 int *complete_flag, int *sort_flag)
482 AttrType completeness;
485 AttrType sort_relation;
488 int completeness_value;
490 const char *structure_str = 0;
492 int sort_relation_value;
496 attr_init_APT(&structure, zapt, 4);
497 attr_init_APT(&completeness, zapt, 6);
498 attr_init_APT(&relation, zapt, 2);
499 attr_init_APT(&sort_relation, zapt, 7);
500 attr_init_APT(&weight, zapt, 9);
501 attr_init_APT(&use, zapt, 1);
503 completeness_value = attr_find(&completeness, NULL);
504 structure_value = attr_find_ex(&structure, NULL, &structure_str);
505 relation_value = attr_find(&relation, NULL);
506 sort_relation_value = attr_find(&sort_relation, NULL);
507 weight_value = attr_find(&weight, NULL);
508 use_value = attr_find(&use, NULL);
510 if (completeness_value == 2 || completeness_value == 3)
516 *sort_flag =(sort_relation_value > 0) ? 1 : 0;
517 *search_type = "phrase";
518 strcpy(rank_type, "void");
519 if (relation_value == 102)
521 if (weight_value == -1)
523 sprintf(rank_type, "rank,w=%d,u=%d", weight_value, use_value);
529 switch (structure_value)
531 case 6: /* word list */
532 *search_type = "and-list";
534 case 105: /* free-form-text */
535 *search_type = "or-list";
537 case 106: /* document-text */
538 *search_type = "or-list";
543 case 108: /* string */
544 *search_type = "phrase";
546 case 107: /* local-number */
547 *search_type = "local";
550 case 109: /* numeric string */
552 *search_type = "numeric";
556 *search_type = "phrase";
560 *search_type = "phrase";
564 *search_type = "phrase";
568 *search_type = "phrase";
571 if (structure_str && *structure_str)
572 *index_type = structure_str;
582 WRBUF zebra_replace(zebra_map_t zm, const char *ex_list,
583 const char *input_str, int input_len)
585 wrbuf_rewind(zm->zebra_maps->wrbuf_1);
586 wrbuf_write(zm->zebra_maps->wrbuf_1, input_str, input_len);
587 return zm->zebra_maps->wrbuf_1;
593 * indent-tabs-mode: nil
595 * vim: shiftwidth=4 tabstop=8 expandtab