X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Fzrpn.c;h=866af77f406ce5891dd5bca79f7dfbcdacf7024f;hb=44cec91a73f331f5825fea0f2ef640cc28fdda11;hp=24a7ae71f60dd83f239a6e601f651f1999173d46;hpb=9b4779254ca63687e44be9f3d2f0efecfff0c1ba;p=idzebra-moved-to-github.git diff --git a/index/zrpn.c b/index/zrpn.c index 24a7ae7..866af77 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -4,7 +4,35 @@ * Sebastian Hammer, Adam Dickmeiss * * $Log: zrpn.c,v $ - * Revision 1.35 1995-11-27 09:29:00 adam + * 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 @@ -119,6 +147,7 @@ #include #include #include +#include #include "zserver.h" #include "attribute.h" @@ -474,7 +503,7 @@ 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); @@ -610,13 +639,20 @@ 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 void verbatim_char (int ch, int *indx, char *dst) +{ + if (!isalnum (ch)) + dst[(*indx)++] = '\\'; + dst[(*indx)++] = ch; +} + static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, const char *term_sub, oid_value attributeSet, struct grep_info *grep_info, @@ -654,58 +690,74 @@ static int trunc_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, 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 */ - sprintf (term_dict + strlen(term_dict), - "([]%d %s)", strlen(term_sub), term_sub); - 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 */ - sprintf (term_dict + strlen(term_dict), - "([]%d %s.*)", strlen(term_sub), term_sub); + 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 */ - strcat (term_dict, "("); - j = strlen(term_dict); + term_dict[j++] = '('; for (i=0; term_sub[i]; i++) - if (i > 2 && term_sub[i] == '#') + if (term_sub[i] == '#' && i > 2) { term_dict[j++] = '.'; term_dict[j++] = '*'; } else - { - term_dict[j++] = '\\'; - term_dict[j++] = term_sub[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=#: %d", r); break; case 102: /* regular expression */ - sprintf (term_dict + strlen(term_dict), "(%s)", term_sub); - 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; } } if (max_pos <= strlen(basenames[base_no])) @@ -1098,12 +1150,15 @@ 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); @@ -1112,19 +1167,21 @@ 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); @@ -1155,13 +1212,46 @@ 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)) @@ -1171,9 +1261,10 @@ 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, @@ -1195,7 +1286,10 @@ int rpn_search (ZServerInfo *zi, 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); @@ -1210,8 +1304,7 @@ struct scan_info { 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; @@ -1246,7 +1339,7 @@ static int scan_handle (Dict_char *name, const char *info, int pos, } -static int dummy_handle (Dict_char *name, const char *info, void *p) +static int dummy_handle (char *name, const char *info, void *p) { return 0; } @@ -1290,7 +1383,7 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, use_value = 1016; i = index_word_prefix (termz, 1, use_value, *basenames); - dict_lookup_grep (zi->wordDict, termz, 0, NULL, &max_pos, + dict_lookup_grep (zi->wordDict, termz, 0, NULL, &max_pos, 0, dummy_handle); if (max_pos <= strlen(*basenames)) {