X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fzrpn.c;h=c647c83a7a70f68037016a1d439c5df5a4bf0b76;hb=b8844c65622c12fb105bf7fc6bab2fc237af7c14;hp=501a294a3cfe8936a17136d46643b5504298b68f;hpb=b7918920dd34b8c4c93e8d3095203e236523f453;p=idzebra-moved-to-github.git diff --git a/index/zrpn.c b/index/zrpn.c index 501a294..c647c83 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -1,10 +1,70 @@ /* - * Copyright (C) 1994-1995, Index Data I/S + * Copyright (C) 1994-1996, Index Data I/S * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: zrpn.c,v $ - * Revision 1.33 1995-11-01 13:58:28 quinn + * Revision 1.49 1996-06-04 10:18:11 adam + * Search/scan uses character mapping module. + * + * Revision 1.48 1996/05/28 15:15:01 adam + * Bug fix: Didn't handle unknown database correctly. + * + * Revision 1.47 1996/05/15 18:36:28 adam + * Function trans_term transforms unsearchable characters to blanks. + * + * Revision 1.46 1996/05/15 11:57:56 adam + * Fixed bug introduced by set/field mapping in search operations. + * + * Revision 1.45 1996/05/14 11:34:00 adam + * Scan support in multiple registers/databases. + * + * Revision 1.44 1996/05/14 06:16:44 adam + * Compact use/set bytes used in search service. + * + * Revision 1.43 1996/05/09 09:54:43 adam + * Server supports maps from one logical attributes to a list of physical + * attributes. + * The extraction process doesn't make space consuming 'any' keys. + * + * Revision 1.42 1996/05/09 07:28:56 quinn + * Work towards phrases and multiple registers + * + * Revision 1.41 1996/03/20 09:36:43 adam + * Function dict_lookup_grep got extra parameter, init_pos, which marks + * from which position in pattern approximate pattern matching should occur. + * Approximate pattern matching is used in relevance=re-2. + * + * Revision 1.40 1996/02/02 13:44:44 adam + * The public dictionary functions simply use char instead of Dict_char + * to represent search strings. Dict_char is used internally only. + * + * Revision 1.39 1996/01/03 16:22:13 quinn + * operator->roperator + * + * Revision 1.38 1995/12/11 09:12:55 adam + * The rec_get function returns NULL if record doesn't exist - will + * happen in the server if the result set records have been deleted since + * the creation of the set (i.e. the search). + * The server saves a result temporarily if it is 'volatile', i.e. the + * set is register dependent. + * + * Revision 1.37 1995/12/06 15:05:28 adam + * More verbose in count_set. + * + * Revision 1.36 1995/12/06 12:41:27 adam + * New command 'stat' for the index program. + * Filenames can be read from stdin by specifying '-'. + * Bug fix/enhancement of the transformation from terms to regular + * expressons in the search engine. + * + * Revision 1.35 1995/11/27 09:29:00 adam + * Bug fixes regarding conversion to regular expressions. + * + * Revision 1.34 1995/11/16 17:00:56 adam + * Better logging of rpn query. + * + * Revision 1.33 1995/11/01 13:58:28 quinn * Moving data1 to yaz/retrieval * * Revision 1.32 1995/10/27 14:00:11 adam @@ -113,69 +173,18 @@ #include #include #include +#include #include "zserver.h" #include "attribute.h" +#include #include #include #include #include #include -int index_word_prefix_map (char *string, oid_value attrSet, int attrUse, - char *basename) -{ - attent *attp; - - logf (LOG_DEBUG, "oid_value attrSet = %d, attrUse = %d", attrSet, attrUse); - attp = att_getentbyatt (attrSet, attrUse); - if (!attp) - return -1; - logf (LOG_DEBUG, "ord=%d", attp->attset_ordinal); - return index_word_prefix (string, attp->attset_ordinal, - attp->local_attribute, basename); -} - -/* - * attr_print: log attributes - */ -static void attr_print (Z_AttributesPlusTerm *t) -{ - int of, i; - for (of = 0; of < t->num_attributes; of++) - { - Z_AttributeElement *element; - element = t->attributeList[of]; - - switch (element->which) - { - case Z_AttributeValue_numeric: - logf (LOG_DEBUG, "attributeType=%d value=%d", - *element->attributeType, - *element->value.numeric); - break; - case Z_AttributeValue_complex: - logf (LOG_DEBUG, "attributeType=%d complex", - *element->attributeType); - for (i = 0; ivalue.complex->num_list; i++) - { - if (element->value.complex->list[i]->which == - Z_StringOrNumeric_string) - logf (LOG_DEBUG, " string: '%s'", - element->value.complex->list[i]->u.string); - else if (element->value.complex->list[i]->which == - Z_StringOrNumeric_numeric) - logf (LOG_DEBUG, " numeric: '%d'", - *element->value.complex->list[i]->u.numeric); - } - break; - default: - assert (0); - } - } -} - typedef struct { int type; int major; @@ -507,7 +516,7 @@ static void add_isam_p (const char *info, struct grep_info *p) (p->isam_p_indx)++; } -static int grep_handle (Dict_char *name, const char *info, void *p) +static int grep_handle (char *name, const char *info, void *p) { logf (LOG_DEBUG, "dict name: %s", name); add_isam_p (info, p); @@ -643,15 +652,22 @@ static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, } logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict); r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, max_pos, - grep_handle); + 0, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r); logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); return 1; } -static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - const char *term_sub, +static void verbatim_char (int ch, int *indx, char *dst) +{ + if (!isalnum (ch)) + dst[(*indx)++] = '\\'; + dst[(*indx)++] = ch; +} + +static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, + const char *term_sub, int regType, oid_value attributeSet, struct grep_info *grep_info, int num_bases, char **basenames) { @@ -675,79 +691,122 @@ static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, for (base_no = 0; base_no < num_bases; base_no++) { - int max_pos; - int prefix_len = index_word_prefix_map (term_dict, curAttributeSet, - use_value, - basenames[base_no]); - if (prefix_len < 0) + attent *attp; + data1_local_attribute *local_attr; + int max_pos, prefix_len = 0; + + attp = att_getentbyatt (curAttributeSet, use_value); + if (!attp) { + logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d", + curAttributeSet, use_value); zi->errCode = 114; return -1; } + if (zebTargetInfo_curDatabase (zi->zti, basenames[base_no])) + { + zi->errCode = 109; /* Database unavailable */ + zi->errString = basenames[base_no]; + return -1; + } + for (local_attr = attp->local_attributes; local_attr; + local_attr = local_attr->next) + { + int ord; + + ord = zebTargetInfo_lookupSU (zi->zti, attp->attset_ordinal, + local_attr->local); + if (ord < 0) + continue; + if (prefix_len) + term_dict[prefix_len++] = '|'; + else + term_dict[prefix_len++] = '('; + term_dict[prefix_len++] = 1; + term_dict[prefix_len++] = ord; + } + if (!prefix_len) + { + zi->errCode = 114; + return -1; + } + term_dict[prefix_len++] = ')'; + term_dict[prefix_len++] = 1; + term_dict[prefix_len++] = regType; + term_dict[prefix_len] = '\0'; if (!relational_term (zi, zapt, term_sub, term_dict, attributeSet, grep_info, &max_pos)) { + const char *cp; + + j = prefix_len; switch (truncation_value) { case -1: /* not specified */ case 100: /* do not truncate */ - strcat (term_dict, "("); - strcat (term_dict, term_sub); - strcat (term_dict, ")"); - logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict); + term_dict[j++] = '('; + for (i = 0; term_sub[i]; i++) + verbatim_char (term_sub[i], &j, term_dict); + strcpy (term_dict+j, ")"); r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, grep_handle); + &max_pos, 0, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep err, trunc=none:%d", r); break; case 1: /* right truncation */ - strcat (term_dict, term_sub); - strcat (term_dict, ".*"); + term_dict[j++] = '('; + for (i = 0; term_sub[i]; i++) + verbatim_char (term_sub[i], &j, term_dict); + strcpy (term_dict+j, ".*)"); dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, grep_handle); + &max_pos, 0, grep_handle); break; case 2: /* left truncation */ case 3: /* left&right truncation */ zi->errCode = 120; return -1; case 101: /* process # in term */ - for (j = strlen(term_dict), i = 0; term_sub[i] && i < 2; i++) - term_dict[j++] = term_sub[i]; - for (; term_sub[i]; i++) - if (term_sub[i] == '#') + term_dict[j++] = '('; + for (i=0; term_sub[i]; i++) + if (term_sub[i] == '#' && i > 2) { term_dict[j++] = '.'; term_dict[j++] = '*'; } else - term_dict[j++] = term_sub[i]; - term_dict[j] = '\0'; + verbatim_char (term_sub[i], &j, term_dict); + strcpy (term_dict+j, ")"); r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, grep_handle); + &max_pos, 0, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d", r); break; case 102: /* regular expression */ - strcat (term_dict, "("); - strcat (term_dict, term_sub); - strcat (term_dict, ")"); - logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict); + sprintf (term_dict + j, "(%s)", term_sub); r = dict_lookup_grep (zi->wordDict, term_dict, 0, grep_info, - &max_pos, grep_handle); + &max_pos, 0, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", r); break; + case 103: /* regular expression with error correction */ + cp = term_sub; + r = 0; + if (*cp == '*' && cp[1] && cp[2]) + { + r = atoi (cp+1); + cp += 2; + } + sprintf (term_dict + j, "(%s)", cp); + r = dict_lookup_grep (zi->wordDict, term_dict, r, grep_info, + &max_pos, j, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep err, trunc=eregular: %d", + r); + break; } } - logf (LOG_DEBUG, "max_pos = %d", max_pos); - if (max_pos <= strlen(basenames[base_no])) - { - zi->errCode = 109; /* Database unavailable */ - zi->errString = basenames[base_no]; - return -1; - } } logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); return 0; @@ -756,14 +815,42 @@ static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, static void trans_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, char *termz) { - size_t i, sizez; + size_t sizez; Z_Term *term = zapt->term; sizez = term->u.general->len; - if (sizez > IT_MAX_WORD) - sizez = IT_MAX_WORD; - for (i = 0; i < sizez; i++) - termz[i] = index_char_cvt (term->u.general->buf[i]); + if (sizez > IT_MAX_WORD-1) + sizez = IT_MAX_WORD-1; + memcpy (termz, term->u.general->buf, sizez); + termz[sizez] = '\0'; +} + +static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, + char *termz) +{ + Z_Term *term = zapt->term; + char **map; + const char *cp = (const char *) term->u.general->buf; + const char *cp_end = cp + term->u.general->len; + const char *src; + int i = 0; + int prev_space = 0; + int len; + + while ((len = (cp_end - cp)) > 0) + { + map = map_chrs_input (&cp, len); + if (**map == *CHR_SPACE) + { + if (prev_space) + continue; + prev_space = 1; + } + else + prev_space = 0; + for (src = *map; *src; src++) + termz[i++] = *src; + } termz[i] = '\0'; } @@ -790,26 +877,38 @@ static RSET rpn_search_APT_relevance (ZServerInfo *zi, return NULL; } trans_term (zi, zapt, termz); + grep_info.isam_p_indx = 0; grep_info.isam_p_size = 0; grep_info.isam_p_buf = NULL; while (1) { - if ((p1 = strchr (p0, ' '))) + char **map; + char *p2 = NULL; + p1 = p0; + while (*p1) { - memcpy (term_sub, p0, p1-p0); - term_sub[p1-p0] = '\0'; + map = map_chrs_input (&p1, strlen(p1)); + if (**map == *CHR_SPACE) + break; + p2 = p1; } - else - strcpy (term_sub, p0); - if (trunc_term (zi, zapt, term_sub, attributeSet, &grep_info, + if (p1 == p0) + break; + memcpy (term_sub, p0, p2-p0); + term_sub[p2-p0] = '\0'; + while (*p1) + { + map = map_chrs_input (&p1, strlen(p1)); + if (**map == *CHR_SPACE) + break; + } + if (field_term (zi, zapt, term_sub, 'w', attributeSet, &grep_info, num_bases, basenames)) return NULL; if (!p1) break; p0 = p1; - while (*++p0 == ' ') - ; } parms.isam_positions = grep_info.isam_p_buf; parms.no_isam_positions = grep_info.isam_p_indx; @@ -821,10 +920,10 @@ static RSET rpn_search_APT_relevance (ZServerInfo *zi, return result; } -static RSET rpn_search_APT_word (ZServerInfo *zi, - Z_AttributesPlusTerm *zapt, - oid_value attributeSet, - int num_bases, char **basenames) +static RSET rpn_search_APT_cphrase (ZServerInfo *zi, + Z_AttributesPlusTerm *zapt, + oid_value attributeSet, + int num_bases, char **basenames) { rset_isam_parms parms; char termz[IT_MAX_WORD+1]; @@ -842,7 +941,7 @@ static RSET rpn_search_APT_word (ZServerInfo *zi, grep_info.isam_p_size = 0; grep_info.isam_p_buf = NULL; - if (trunc_term (zi, zapt, termz, attributeSet, &grep_info, + if (field_term (zi, zapt, termz, 'p', attributeSet, &grep_info, num_bases, basenames)) return NULL; if (grep_info.isam_p_indx < 1) @@ -970,16 +1069,30 @@ static RSET rpn_search_APT_phrase (ZServerInfo *zi, while (1) { - if ((p1 = strchr (p0, ' '))) + char **map; + char *p2 = NULL; + p1 = p0; + while (*p1) { - memcpy (term_sub, p0, p1-p0); - term_sub[p1-p0] = '\0'; + map = map_chrs_input (&p1, strlen(p1)); + if (**map == *CHR_SPACE) + break; + p2 = p1; + } + if (p1 == p0) + break; + memcpy (term_sub, p0, p2-p0); + term_sub[p2-p0] = '\0'; + p0 = p1; + while (*p1) + { + map = map_chrs_input (&p1, strlen(p1)); + if (**map != *CHR_SPACE) + break; + p0 = p1; } - else - strcpy (term_sub, p0); - grep_info.isam_p_indx = 0; - if (trunc_term (zi, zapt, term_sub, attributeSet, &grep_info, + if (field_term (zi, zapt, term_sub, 'w', attributeSet, &grep_info, num_bases, basenames)) return NULL; if (grep_info.isam_p_indx == 0) @@ -1001,9 +1114,6 @@ static RSET rpn_search_APT_phrase (ZServerInfo *zi, break; if (!p1) break; - p0 = p1; - while (*++p0 == ' ') - ; } xfree (grep_info.isam_p_buf); if (rset_no == 0) @@ -1036,6 +1146,7 @@ static RSET rpn_search_APT_local (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, rsfd = rset_open (result, RSETF_WRITE|RSETF_SORT_SYSNO); trans_term (zi, zapt, termz); + key.sysno = atoi (termz); if (key.sysno <= 0) key.sysno = 1; @@ -1050,25 +1161,34 @@ static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, { AttrType relation; AttrType structure; - int relation_value, structure_value; + AttrType completeness; + int relation_value, structure_value, completeness_value; attr_init (&relation, zapt, 2); attr_init (&structure, zapt, 4); + attr_init (&completeness, zapt, 6); relation_value = attr_find (&relation, NULL); structure_value = attr_find (&structure, NULL); + completeness_value = attr_find (&completeness, NULL); switch (structure_value) { case -1: if (relation_value == 102) /* relevance relation */ return rpn_search_APT_relevance (zi, zapt, attributeSet, num_bases, basenames); + if (completeness_value == 2 || completeness_value == 3) + return rpn_search_APT_cphrase (zi, zapt, attributeSet, + num_bases, basenames); return rpn_search_APT_phrase (zi, zapt, attributeSet, num_bases, basenames); case 1: /* phrase */ if (relation_value == 102) /* relevance relation */ return rpn_search_APT_relevance (zi, zapt, attributeSet, num_bases, basenames); + if (completeness_value == 2 || completeness_value == 3) + return rpn_search_APT_cphrase (zi, zapt, attributeSet, + num_bases, basenames); return rpn_search_APT_phrase (zi, zapt, attributeSet, num_bases, basenames); break; @@ -1076,8 +1196,11 @@ static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, if (relation_value == 102) /* relevance relation */ return rpn_search_APT_relevance (zi, zapt, attributeSet, num_bases, basenames); - return rpn_search_APT_word (zi, zapt, attributeSet, - num_bases, basenames); + if (completeness_value == 2 || completeness_value == 3) + return rpn_search_APT_cphrase (zi, zapt, attributeSet, + num_bases, basenames); + return rpn_search_APT_phrase (zi, zapt, attributeSet, + num_bases, basenames); case 3: /* key */ break; case 4: /* year */ @@ -1106,8 +1229,8 @@ static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, case 107: /* local-number */ return rpn_search_APT_local (zi, zapt, attributeSet); case 108: /* string */ - return rpn_search_APT_word (zi, zapt, attributeSet, - num_bases, basenames); + return rpn_search_APT_phrase (zi, zapt, attributeSet, + num_bases, basenames); case 109: /* numeric string */ break; } @@ -1132,12 +1255,15 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, if (zs->which == Z_RPNStructure_complex) { rset_bool_parms bool_parms; + int soft = 0; bool_parms.rset_l = rpn_search_structure (zi, zs->u.complex->s1, attributeSet, num_bases, basenames); if (bool_parms.rset_l == NULL) return NULL; + if (rset_is_ranked(bool_parms.rset_l)) + soft = 1; bool_parms.rset_r = rpn_search_structure (zi, zs->u.complex->s2, attributeSet, num_bases, basenames); @@ -1146,19 +1272,21 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, rset_delete (bool_parms.rset_l); return NULL; } + if (rset_is_ranked(bool_parms.rset_r)) + soft = 1; bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare; - switch (zs->u.complex->operator->which) + switch (zs->u.complex->roperator->which) { case Z_Operator_and: - r = rset_create (rset_kind_and, &bool_parms); + r = rset_create (soft ? rset_kind_sand:rset_kind_and, &bool_parms); break; case Z_Operator_or: - r = rset_create (rset_kind_or, &bool_parms); + r = rset_create (soft ? rset_kind_sor:rset_kind_or, &bool_parms); break; case Z_Operator_and_not: - r = rset_create (rset_kind_not, &bool_parms); + r = rset_create (soft ? rset_kind_snot:rset_kind_not, &bool_parms); break; default: assert (0); @@ -1189,13 +1317,46 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, return r; } +void count_set_save (RSET *r, int *count) +{ + int psysno = 0; + int kno = 0; + struct it_key key; + RSFD rfd, wfd; + RSET w; + rset_temp_parms parms; + + logf (LOG_DEBUG, "count_set_save"); + *count = 0; + parms.key_size = sizeof(struct it_key); + w = rset_create (rset_kind_temp, &parms); + wfd = rset_open (w, RSETF_WRITE|RSETF_SORT_SYSNO); + rfd = rset_open (*r, RSETF_READ|RSETF_SORT_SYSNO); + while (rset_read (*r, rfd, &key)) + { + if (key.sysno != psysno) + { + rset_write (w, wfd, &key); + psysno = key.sysno; + (*count)++; + } + kno++; + } + rset_close (*r, rfd); + rset_delete (*r); + rset_close (w, wfd); + *r = w; + logf (LOG_DEBUG, "%d keys, %d distinct sysnos", kno, *count); +} + static void count_set (RSET r, int *count) { int psysno = 0; + int kno = 0; struct it_key key; RSFD rfd; - logf (LOG_DEBUG, "rpn_save_set"); + logf (LOG_DEBUG, "count_set"); *count = 0; rfd = rset_open (r, RSETF_READ|RSETF_SORT_SYSNO); while (rset_read (r, rfd, &key)) @@ -1205,9 +1366,10 @@ static void count_set (RSET r, int *count) psysno = key.sysno; (*count)++; } + kno++; } rset_close (r, rfd); - logf (LOG_DEBUG, "%d distinct sysnos", *count); + logf (LOG_DEBUG, "%d keys, %d distinct sysnos", kno, *count); } int rpn_search (ZServerInfo *zi, @@ -1218,41 +1380,45 @@ int rpn_search (ZServerInfo *zi, oident *attrset; oid_value attributeSet; + dict_grep_cmap (zi->wordDict, map_chrs_input); + zlog_rpn (rpn); + zi->errCode = 0; zi->errString = NULL; - + attrset = oid_getentbyoid (rpn->attributeSetId); attributeSet = attrset->value; - rset = rpn_search_structure (zi, rpn->RPNStructure, attributeSet, num_bases, basenames); if (!rset) return zi->errCode; - count_set (rset, hits); + if (rset_is_volatile(rset)) + count_set_save(&rset,hits); + else + count_set (rset, hits); resultSetAdd (zi, setname, 1, rset); if (zi->errCode) logf (LOG_DEBUG, "search error: %d", zi->errCode); return zi->errCode; } +struct scan_info_entry { + char *term; + ISAM_P isam_p; +}; + struct scan_info { - struct scan_entry *list; + struct scan_info_entry *list; ODR odr; int before, after; - ISAM isam; char prefix[20]; }; -static int scan_handle (Dict_char *name, const char *info, int pos, - void *client) +static int scan_handle (char *name, const char *info, int pos, void *client) { int len_prefix, idx; - ISAM_P isam_p; - RSET rset; struct scan_info *scan_info = client; - rset_isam_parms parms; - len_prefix = strlen(scan_info->prefix); if (memcmp (name, scan_info->prefix, len_prefix)) return 1; @@ -1260,113 +1426,261 @@ static int scan_handle (Dict_char *name, const char *info, int pos, idx = scan_info->after - pos + scan_info->before; else idx = - pos - 1; + logf (LOG_DEBUG, "%-3d %s", idx, name+len_prefix); scan_info->list[idx].term = odr_malloc (scan_info->odr, strlen(name + len_prefix)+1); strcpy (scan_info->list[idx].term, name + len_prefix); - assert (*info == sizeof(isam_p)); - memcpy (&isam_p, info+1, sizeof(isam_p)); - parms.is = scan_info->isam; - parms.pos = isam_p; -#if 1 - rset = rset_create (rset_kind_isam, &parms); - count_set (rset, &scan_info->list[idx].occurrences); - rset_delete (rset); -#else - scan_info->list[idx].occurrences = 1; -#endif - logf (LOG_DEBUG, "pos=%3d idx=%3d name=%s", pos, idx, name); + assert (*info == sizeof(ISAM_P)); + memcpy (&scan_info->list[idx].isam_p, info+1, sizeof(ISAM_P)); return 0; } -static int dummy_handle (Dict_char *name, const char *info, void *p) -{ - return 0; +static void scan_term_untrans (ODR odr, char **dstp, const char *src) +{ + char *dst = odr_malloc (odr, strlen(src)*2+1); + *dstp = dst; + + while (*src) + { + const char *cp = map_chrs_output (&src); + while (*cp) + *dst++ = *cp++; + } + *dst = '\0'; } int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, + oid_value attributeset, int num_bases, char **basenames, int *position, int *num_entries, struct scan_entry **list, int *status) { - int i, j, sizez, max_pos; + int i; int pos = *position; int num = *num_entries; int before; int after; + int base_no; char termz[IT_MAX_WORD+20]; AttrType use; int use_value; - Z_Term *term = zapt->term; - struct scan_info scan_info; + AttrType completeness; + int completeness_value; + struct scan_info *scan_info_array; + struct scan_entry *glist; + int ords[32], ord_no = 0; + int ptr[32]; logf (LOG_DEBUG, "scan, position = %d, num = %d", pos, num); - if (num_bases != 1) - return 111; - scan_info.before = before = pos-1; - scan_info.after = after = 1+num-pos; - scan_info.odr = zi->odr; - - logf (LOG_DEBUG, "scan, before = %d, after = %d", before, after); - - scan_info.isam = zi->wordIsam; - scan_info.list = odr_malloc (zi->odr, (before+after)* - sizeof(*scan_info.list)); - for (j = 0; jerrCode = 114; + } + if (zebTargetInfo_curDatabase (zi->zti, basenames[base_no])) + { + zi->errString = basenames[base_no]; + return zi->errCode = 109; /* Database unavailable */ + } + for (local_attr = attp->local_attributes; local_attr && ord_no < 32; + local_attr = local_attr->next) + { + int ord; - dict_lookup_grep (zi->wordDict, termz, 0, NULL, &max_pos, - dummy_handle); - if (max_pos <= strlen(*basenames)) + ord = zebTargetInfo_lookupSU (zi->zti, attp->attset_ordinal, + local_attr->local); + if (ord > 0) + ords[ord_no++] = ord; + } + } + if (ord_no == 0) + return zi->errCode = 113; + before = pos-1; + after = 1+num-pos; + scan_info_array = odr_malloc (zi->odr, ord_no * sizeof(*scan_info_array)); + for (i = 0; i < ord_no; i++) { - zi->errString = *basenames; - return zi->errCode = 109; /* Database unavailable */ + int j, prefix_len = 0; + int before_tmp = before, after_tmp = after; + struct scan_info *scan_info = scan_info_array + i; + + scan_info->before = before; + scan_info->after = after; + scan_info->odr = zi->odr; + + scan_info->list = odr_malloc (zi->odr, (before+after)* + sizeof(*scan_info->list)); + for (j = 0; jlist[j].term = NULL; + termz[prefix_len++] = ords[i]; + termz[prefix_len++] = + (completeness_value==2 || completeness_value==3) ? 'p': 'w'; + termz[prefix_len] = 0; + strcpy (scan_info->prefix, termz); + + trans_scan_term (zi, zapt, termz+prefix_len); + + dict_scan (zi->wordDict, termz, &before_tmp, &after_tmp, scan_info, + scan_handle); } - strcpy (scan_info.prefix, termz); - sizez = term->u.general->len; - if (sizez > IT_MAX_WORD) - sizez = IT_MAX_WORD; - for (j = 0; ju.general->buf[j]); - termz[j+i] = '\0'; + glist = odr_malloc (zi->odr, (before+after)*sizeof(*glist)); + for (i = 0; i < ord_no; i++) + ptr[i] = before; - dict_scan (zi->wordDict, termz, &before, &after, &scan_info, scan_handle); - *status = BEND_SCAN_SUCCESS; - - for (i = 0; iodr, &glist[i+before].term, mterm); + parms.is = zi->wordIsam; + parms.pos = scan_info_array[j0].list[ptr[j0]].isam_p; + rset = rset_create (rset_kind_isam, &parms); + + ptr[j0]++; + for (j = j0+1; jwordIsam; + parms.pos = scan_info_array[j].list[ptr[j]].isam_p; + rset2 = rset_create (rset_kind_isam, &parms); + + bool_parms.key_size = sizeof(struct it_key); + bool_parms.cmp = key_compare; + bool_parms.rset_l = rset; + bool_parms.rset_r = rset2; + + rset = rset_create (rset_kind_or, &bool_parms); + + ptr[j]++; + } + } + count_set (rset, &glist[i+before].occurrences); + rset_delete (rset); + } + if (i < after) + { + *num_entries -= (after-i); *status = BEND_SCAN_PARTIAL; + } + + for (i = 0; i 0)) + { + j0 = j; + mterm = tst; + } + } + if (j0 == -1) break; - if (i) - *status = BEND_SCAN_PARTIAL; - *position -= i; - *num_entries -= i; - *list = scan_info.list+i; /* list is set to first 'real' entry */ + scan_term_untrans (zi->odr, &glist[before-1-i].term, mterm); + + parms.is = zi->wordIsam; + parms.pos = scan_info_array[j0].list[before-1-ptr[j0]].isam_p; + rset = rset_create (rset_kind_isam, &parms); + + ptr[j0]++; + + for (j = j0+1; jwordIsam; + parms.pos = scan_info_array[j].list[before-1-ptr[j]].isam_p; + rset2 = rset_create (rset_kind_isam, &parms); + + bool_parms.key_size = sizeof(struct it_key); + bool_parms.cmp = key_compare; + bool_parms.rset_l = rset; + bool_parms.rset_r = rset2; + + rset = rset_create (rset_kind_or, &bool_parms); - if (*num_entries == 0) /* signal 'unsupported use-attribute' */ - zi->errCode = 114; /* if no entries was found */ + ptr[j]++; + } + } + count_set (rset, &glist[before-1-i].occurrences); + rset_delete (rset); + } + i = before-i; + if (i) + { + *status = BEND_SCAN_PARTIAL; + *position -= i; + *num_entries -= i; + } + *list = glist + i; /* list is set to first 'real' entry */ + logf (LOG_DEBUG, "position = %d, num_entries = %d", *position, *num_entries); if (zi->errCode) logf (LOG_DEBUG, "scan error: %d", zi->errCode); - return 0; + return zi->errCode; } - -