X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=util%2Fzebramap.c;h=3d75a4160ea8b2a51c9e2a0fb26ccc8459f7ed9f;hb=ef696645cc3b7e0f4027008d1dc589c0f0f90c1f;hp=8fd2227563287055e271474accae0030f5f3c021;hpb=5c693d36af8be6f6642257160b3c6441d2e2d762;p=idzebra-moved-to-github.git diff --git a/util/zebramap.c b/util/zebramap.c index 8fd2227..3d75a41 100644 --- a/util/zebramap.c +++ b/util/zebramap.c @@ -1,10 +1,35 @@ /* - * Copyright (C) 1994-1998, Index Data I/S + * Copyright (C) 1994-1999, Index Data * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: zebramap.c,v $ - * Revision 1.7 1998-02-10 12:03:07 adam + * Revision 1.15 1999-05-26 07:49:14 adam + * C++ compilation. + * + * Revision 1.14 1999/02/19 10:37:40 adam + * Minor fix. + * + * Revision 1.13 1999/02/18 15:01:04 adam + * Structure=key uses register type 0. + * + * Revision 1.12 1999/02/12 13:29:25 adam + * Implemented position-flag for registers. + * + * Revision 1.11 1998/10/13 20:09:19 adam + * Changed call to readconf_line. + * + * Revision 1.10 1998/06/23 15:33:37 adam + * Added feature to specify sort criteria in query (type 7 specifies + * sort flags). + * + * Revision 1.9 1998/04/02 14:35:30 adam + * First version of Zebra that works with compiled ASN.1. + * + * Revision 1.8 1998/03/05 08:42:44 adam + * Minor changes to zebramap data structures. Query mapping rules changed. + * + * Revision 1.7 1998/02/10 12:03:07 adam * Implemented Sort. * * Revision 1.6 1998/01/29 13:36:01 adam @@ -39,10 +64,22 @@ #include #include +#define ZEBRA_MAP_TYPE_SORT 1 +#define ZEBRA_MAP_TYPE_INDEX 2 + struct zebra_map { - int reg_type; + unsigned reg_id; int completeness; - int sort_flag; + int positioned; + int type; + union { + struct { + int dummy; + } index; + struct { + int entry_size; + } sort; + } u; chrmaptab maptab; const char *maptab_name; struct zebra_map *next; @@ -76,6 +113,7 @@ static void zebra_map_read (ZebraMaps zms, const char *name) char line[512]; char *argv[10]; int argc; + int lineno = 0; struct zebra_map **zm = 0, *zp; if (!(f = yaz_path_fopen(zms->tabpath, name, "r"))) @@ -83,65 +121,77 @@ static void zebra_map_read (ZebraMaps zms, const char *name) logf(LOG_WARN|LOG_ERRNO, "%s", name); return ; } - while ((argc = readconf_line(f, line, 512, argv, 10))) + while ((argc = readconf_line(f, &lineno, line, 512, argv, 10))) { - if (!strcmp (argv[0], "index") && argc == 2) + if (!yaz_matchstr (argv[0], "index") && argc == 2) { if (!zm) zm = &zms->map_list; else zm = &(*zm)->next; - *zm = nmem_malloc (zms->nmem, sizeof(**zm)); - (*zm)->reg_type = argv[1][0]; + *zm = (struct zebra_map *) nmem_malloc (zms->nmem, sizeof(**zm)); + (*zm)->reg_id = argv[1][0]; (*zm)->maptab_name = NULL; (*zm)->maptab = NULL; - (*zm)->sort_flag = 0; + (*zm)->type = ZEBRA_MAP_TYPE_INDEX; (*zm)->completeness = 0; + (*zm)->positioned = 1; } - else if (!strcmp (argv[0], "sort") && argc == 2) + else if (!yaz_matchstr (argv[0], "sort") && argc == 2) { if (!zm) zm = &zms->map_list; else zm = &(*zm)->next; - *zm = nmem_malloc (zms->nmem, sizeof(**zm)); - (*zm)->reg_type = argv[1][0]; + *zm = (struct zebra_map *) nmem_malloc (zms->nmem, sizeof(**zm)); + (*zm)->reg_id = argv[1][0]; (*zm)->maptab_name = NULL; - (*zm)->sort_flag = 1; + (*zm)->type = ZEBRA_MAP_TYPE_SORT; + (*zm)->u.sort.entry_size = 80; (*zm)->maptab = NULL; (*zm)->completeness = 0; + (*zm)->positioned = 0; } - else if (zm && !strcmp (argv[0], "charmap") && argc == 2) + else if (zm && !yaz_matchstr (argv[0], "charmap") && argc == 2) { (*zm)->maptab_name = nmem_strdup (zms->nmem, argv[1]); } - else if (zm && !strcmp (argv[0], "completeness") && argc == 2) + else if (zm && !yaz_matchstr (argv[0], "completeness") && argc == 2) { (*zm)->completeness = atoi (argv[1]); } + else if (zm && !yaz_matchstr (argv[0], "position") && argc == 2) + { + (*zm)->positioned = atoi (argv[1]); + } + else if (zm && !yaz_matchstr (argv[0], "entrysize") && argc == 2) + { + if ((*zm)->type == ZEBRA_MAP_TYPE_SORT) + (*zm)->u.sort.entry_size = atoi (argv[1]); + } } if (zm) (*zm)->next = NULL; fclose (f); for (zp = zms->map_list; zp; zp = zp->next) - zms->lookup_array[zp->reg_type & 255] = zp; + zms->lookup_array[zp->reg_id] = zp; } static void zms_map_handle (void *p, const char *name, const char *value) { - ZebraMaps zms = p; + ZebraMaps zms = (ZebraMaps) p; zebra_map_read (zms, value); } -ZebraMaps zebra_maps_open (const char *tabpath, Res res) +ZebraMaps zebra_maps_open (Res res) { - ZebraMaps zms = xmalloc (sizeof(*zms)); + ZebraMaps zms = (ZebraMaps) xmalloc (sizeof(*zms)); int i; zms->nmem = nmem_create (); - zms->tabpath = nmem_strdup (zms->nmem, tabpath); + zms->tabpath = nmem_strdup (zms->nmem, res_get_def (res, "profilePath", ".")); zms->map_list = NULL; zms->temp_map_str[0] = '\0'; @@ -150,31 +200,41 @@ ZebraMaps zebra_maps_open (const char *tabpath, Res res) zms->temp_map_ptr[0] = zms->temp_map_str; zms->temp_map_ptr[1] = NULL; - zms->lookup_array = + zms->lookup_array = (struct zebra_map**) nmem_malloc (zms->nmem, sizeof(*zms->lookup_array)*256); for (i = 0; i<256; i++) zms->lookup_array[i] = 0; - if (!res_trav (res, "index", zms, zms_map_handle)) + if (!res || !res_trav (res, "index", zms, zms_map_handle)) zebra_map_read (zms, "default.idx"); return zms; } -struct zebra_map *zebra_map_get (ZebraMaps zms, int reg_type) +struct zebra_map *zebra_map_get (ZebraMaps zms, unsigned reg_id) { - return zms->lookup_array[reg_type]; + return zms->lookup_array[reg_id]; } -chrmaptab zebra_charmap_get (ZebraMaps zms, int reg_type) +chrmaptab zebra_charmap_get (ZebraMaps zms, unsigned reg_id) { - struct zebra_map *zm = zebra_map_get (zms, reg_type); + struct zebra_map *zm = zebra_map_get (zms, reg_id); if (!zm) { - logf (LOG_WARN, "Unknown register type: %c", reg_type); - return NULL; + zm = (struct zebra_map *) nmem_malloc (zms->nmem, sizeof(*zm)); + logf (LOG_WARN, "Unknown register type: %c", reg_id); + + zm->reg_id = reg_id; + zm->maptab_name = NULL; + zm->maptab = NULL; + zm->type = ZEBRA_MAP_TYPE_INDEX; + zm->completeness = 0; + zm->next = zms->map_list; + zms->map_list = zm->next; + + zms->lookup_array[zm->reg_id & 255] = zm; } if (!zm->maptab) { - if (!strcmp (zm->maptab_name, "@")) + if (!zm->maptab_name || !yaz_matchstr (zm->maptab_name, "@")) return NULL; if (!(zm->maptab = chrmaptab_create (zms->tabpath, zm->maptab_name, 0))) @@ -186,12 +246,12 @@ chrmaptab zebra_charmap_get (ZebraMaps zms, int reg_type) return zm->maptab; } -const char **zebra_maps_input (ZebraMaps zms, int reg_type, +const char **zebra_maps_input (ZebraMaps zms, unsigned reg_id, const char **from, int len) { chrmaptab maptab; - maptab = zebra_charmap_get (zms, reg_type); + maptab = zebra_charmap_get (zms, reg_id); if (maptab) return chr_map_input(maptab, from, len); @@ -201,13 +261,14 @@ const char **zebra_maps_input (ZebraMaps zms, int reg_type, return zms->temp_map_ptr; } -const char *zebra_maps_output(ZebraMaps zms, int reg_type, const char **from) +const char *zebra_maps_output(ZebraMaps zms, unsigned reg_id, + const char **from) { chrmaptab maptab; unsigned char i = (unsigned char) **from; static char buf[2] = {0,0}; - maptab = zebra_charmap_get (zms, reg_type); + maptab = zebra_charmap_get (zms, reg_id); if (maptab) return chr_map_output (maptab, from, 1); (*from)++; @@ -273,9 +334,13 @@ static int attr_find (AttrType *src, oid_value *attributeSetP) static void attr_init_APT (AttrType *src, Z_AttributesPlusTerm *zapt, int type) { - +#ifdef ASN_COMPILED + src->attributeList = zapt->attributes->attributes; + src->num_attributes = zapt->attributes->num_attributes; +#else src->attributeList = zapt->attributeList; src->num_attributes = zapt->num_attributes; +#endif src->type = type; src->major = 0; src->minor = 0; @@ -292,19 +357,27 @@ static void attr_init_AttrList (AttrType *src, Z_AttributeList *list, int type) /* ------------------------------------ */ -int zebra_maps_is_complete (ZebraMaps zms, int reg_type) +int zebra_maps_is_complete (ZebraMaps zms, unsigned reg_id) { - struct zebra_map *zm = zebra_map_get (zms, reg_type); + struct zebra_map *zm = zebra_map_get (zms, reg_id); if (zm) return zm->completeness; return 0; } -int zebra_maps_is_sort (ZebraMaps zms, int reg_type) +int zebra_maps_is_positioned (ZebraMaps zms, unsigned reg_id) { - struct zebra_map *zm = zebra_map_get (zms, reg_type); + struct zebra_map *zm = zebra_map_get (zms, reg_id); if (zm) - return zm->sort_flag; + return zm->positioned; + return 0; +} + +int zebra_maps_is_sort (ZebraMaps zms, unsigned reg_id) +{ + struct zebra_map *zm = zebra_map_get (zms, reg_id); + if (zm) + return zm->type == ZEBRA_MAP_TYPE_SORT; return 0; } @@ -317,60 +390,77 @@ int zebra_maps_sort (ZebraMaps zms, Z_SortAttributes *sortAttributes) } int zebra_maps_attr (ZebraMaps zms, Z_AttributesPlusTerm *zapt, - int *reg_type, char **search_type, int *complete_flag) + unsigned *reg_id, char **search_type, char **rank_type, + int *complete_flag, int *sort_flag) { AttrType completeness; AttrType structure; AttrType relation; + AttrType sort_relation; int completeness_value; int structure_value; int relation_value; + int sort_relation_value; attr_init_APT (&structure, zapt, 4); attr_init_APT (&completeness, zapt, 6); attr_init_APT (&relation, zapt, 2); + attr_init_APT (&sort_relation, zapt, 7); completeness_value = attr_find (&completeness, NULL); structure_value = attr_find (&structure, NULL); relation_value = attr_find (&relation, NULL); + sort_relation_value = attr_find (&sort_relation, NULL); if (completeness_value == 2 || completeness_value == 3) *complete_flag = 1; else *complete_flag = 0; - *reg_type = 0; + *reg_id = 0; + *sort_flag = (sort_relation_value > 0) ? 1 : 0; *search_type = "phrase"; + *rank_type = "void"; if (relation_value == 102) - *search_type = "ranked"; + *rank_type = "rank"; + if (*complete_flag) + *reg_id = 'p'; + else + *reg_id = 'w'; switch (structure_value) { case 6: /* word list */ + *search_type = "and-list"; + break; case 105: /* free-form-text */ + *search_type = "or-list"; + break; case 106: /* document-text */ - *search_type = "ranked"; - *reg_type = 'w'; - break; + *search_type = "or-list"; + break; case -1: case 1: /* phrase */ case 2: /* word */ - case 3: /* key */ case 108: /* string */ - if (*complete_flag) - *reg_type = 'p'; - else - *reg_type = 'w'; + *search_type = "phrase"; break; case 107: /* local-number */ *search_type = "local"; - *reg_type = 0; + *reg_id = 0; + break; case 109: /* numeric string */ - *reg_type = 'n'; + *reg_id = 'n'; + *search_type = "numeric"; break; case 104: /* urx */ - *reg_type = 'u'; + *reg_id = 'u'; + *search_type = "phrase"; break; + case 3: /* key */ + *reg_id = '0'; + *search_type = "phrase"; + break; default: return -1; }