X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fzrpn.c;h=ca8e014dc0547624f126e7e002881ac1e1a1be84;hp=01019044d599b1bca3c5d4e9aedf1c879670a228;hb=ef696645cc3b7e0f4027008d1dc589c0f0f90c1f;hpb=ec874a59d0538f9eae969cf885e63f5ebda0f601 diff --git a/index/zrpn.c b/index/zrpn.c index 0101904..ca8e014 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -1,10 +1,88 @@ /* - * Copyright (C) 1994-1997, Index Data I/S + * Copyright (C) 1995-1999, Index Data * All rights reserved. * Sebastian Hammer, Adam Dickmeiss * * $Log: zrpn.c,v $ - * Revision 1.70 1997-10-31 12:34:43 adam + * Revision 1.92 1999-05-26 07:49:13 adam + * C++ compilation. + * + * Revision 1.91 1999/02/02 14:51:13 adam + * Updated WIN32 code specific sections. Changed header. + * + * Revision 1.90 1998/11/16 16:03:43 adam + * Moved loggin utilities to Yaz. Was implemented in file zlogs.c. + * + * Revision 1.89 1998/11/16 10:11:55 adam + * Added addtional info for error 114 - unsupported use attribute. + * + * Revision 1.88 1998/10/18 07:54:52 adam + * Additional info added for diagnostics 114 (Unsupported use attribute) and + * 121 (Unsupported attribute set). + * + * Revision 1.87 1998/09/28 11:19:12 adam + * Fix for Compiled ASN.1. + * + * Revision 1.86 1998/09/22 10:48:20 adam + * Minor changes in search API. + * + * Revision 1.85 1998/09/22 10:03:43 adam + * Changed result sets to be persistent in the sense that they can + * be re-searched if needed. + * Fixed memory leak in rsm_or. + * + * Revision 1.84 1998/09/18 12:41:00 adam + * Fixed bug with numerical relations. + * + * Revision 1.83 1998/09/02 13:53:19 adam + * Extra parameter decode added to search routines to implement + * persistent queries. + * + * Revision 1.82 1998/06/26 11:16:40 quinn + * Added support (un-optimised) for left and left/right truncation + * + * Revision 1.81 1998/06/24 12:16:14 adam + * Support for relations on text operands. Open range support in + * DFA module (i.e. [-j], [g-]). + * + * Revision 1.80 1998/06/23 15:33:34 adam + * Added feature to specify sort criteria in query (type 7 specifies + * sort flags). + * + * Revision 1.79 1998/06/22 11:35:09 adam + * Minor changes. + * + * Revision 1.78 1998/06/08 14:43:17 adam + * Added suport for EXPLAIN Proxy servers - added settings databasePath + * and explainDatabase to facilitate this. Increased maximum number + * of databases and attributes in one register. + * + * Revision 1.77 1998/05/20 10:12:22 adam + * Implemented automatic EXPLAIN database maintenance. + * Modified Zebra to work with ASN.1 compiled version of YAZ. + * + * Revision 1.76 1998/04/02 14:35:29 adam + * First version of Zebra that works with compiled ASN.1. + * + * Revision 1.75 1998/03/05 08:45:13 adam + * New result set model and modular ranking system. Moved towards + * descent server API. System information stored as "SGML" records. + * + * Revision 1.74 1998/02/10 12:03:06 adam + * Implemented Sort. + * + * Revision 1.73 1998/01/29 13:40:11 adam + * Better logging for scan service. + * + * Revision 1.72 1998/01/07 13:53:41 adam + * Queries using simple ranked operands returns right number of hits. + * + * Revision 1.71 1997/12/18 10:54:24 adam + * New method result set method rs_hits that returns the number of + * hits in result-set (if known). The ranked result set returns real + * number of hits but only when not combined with other operands. + * + * Revision 1.70 1997/10/31 12:34:43 adam * Changed a few log statements. * * Revision 1.69 1997/10/29 12:05:02 adam @@ -244,7 +322,7 @@ */ #include #include -#ifdef WINDOWS +#ifdef WIN32 #include #else #include @@ -257,7 +335,6 @@ #include #include #include -#include struct rpn_char_map_info { ZebraMaps zm; @@ -266,16 +343,16 @@ struct rpn_char_map_info { static const char **rpn_char_map_handler (void *vp, const char **from, int len) { - struct rpn_char_map_info *p = vp; + struct rpn_char_map_info *p = (struct rpn_char_map_info *) vp; return zebra_maps_input (p->zm, p->reg_type, from, len); } -static void rpn_char_map_prepare (ZServerInfo *zi, int reg_type, +static void rpn_char_map_prepare (ZebraHandle zh, int reg_type, struct rpn_char_map_info *map_info) { - map_info->zm = zi->zebra_maps; + map_info->zm = zh->zebra_maps; map_info->reg_type = reg_type; - dict_grep_cmap (zi->dict, map_info, rpn_char_map_handler); + dict_grep_cmap (zh->dict, map_info, rpn_char_map_handler); } typedef struct { @@ -287,11 +364,22 @@ typedef struct { static int attr_find (AttrType *src, oid_value *attributeSetP) { - while (src->major < src->zapt->num_attributes) + int num_attributes; + +#ifdef ASN_COMPILED + num_attributes = src->zapt->attributes->num_attributes; +#else + num_attributes = src->zapt->num_attributes; +#endif + while (src->major < num_attributes) { Z_AttributeElement *element; +#ifdef ASN_COMPILED + element = src->zapt->attributes->attributes[src->major]; +#else element = src->zapt->attributeList[src->major]; +#endif if (src->type == *element->attributeType) { switch (element->which) @@ -345,23 +433,37 @@ struct grep_info { #ifdef TERM_COUNT int *term_no; #endif - ISAM_P *isam_p_buf; + ISAM_P *isam_p_buf; int isam_p_size; - int isam_p_indx; + int isam_p_indx; + ZebraHandle zh; + int reg_type; }; -static void add_isam_p (const char *info, struct grep_info *p) +static void term_untrans (ZebraHandle zh, int reg_type, + char *dst, const char *src) +{ + while (*src) + { + const char *cp = zebra_maps_output (zh->zebra_maps, reg_type, &src); + while (*cp) + *dst++ = *cp++; + } + *dst = '\0'; +} + +static void add_isam_p (const char *name, const char *info, + struct grep_info *p) { if (p->isam_p_indx == p->isam_p_size) { ISAM_P *new_isam_p_buf; #ifdef TERM_COUNT int *new_term_no; -#endif - +#endif p->isam_p_size = 2*p->isam_p_size + 100; - new_isam_p_buf = xmalloc (sizeof(*new_isam_p_buf) * - p->isam_p_size); + new_isam_p_buf = (ISAM_P *) xmalloc (sizeof(*new_isam_p_buf) * + p->isam_p_size); if (p->isam_p_buf) { memcpy (new_isam_p_buf, p->isam_p_buf, @@ -371,8 +473,8 @@ static void add_isam_p (const char *info, struct grep_info *p) p->isam_p_buf = new_isam_p_buf; #ifdef TERM_COUNT - new_term_no = xmalloc (sizeof(*new_term_no) * - p->isam_p_size); + new_term_no = (int *) xmalloc (sizeof(*new_term_no) * + p->isam_p_size); if (p->term_no) { memcpy (new_term_no, p->isam_p_buf, @@ -384,12 +486,17 @@ static void add_isam_p (const char *info, struct grep_info *p) } assert (*info == sizeof(*p->isam_p_buf)); memcpy (p->isam_p_buf + p->isam_p_indx, info+1, sizeof(*p->isam_p_buf)); + +#if 0 + term_untrans (p->zh, p->reg_type, term_tmp, name+2); + logf (LOG_DEBUG, "grep: %s", term_tmp); +#endif (p->isam_p_indx)++; } static int grep_handle (char *name, const char *info, void *p) { - add_isam_p (info, p); + add_isam_p (name, info, (struct grep_info *) p); return 0; } @@ -417,11 +524,13 @@ static int term_pre (ZebraMaps zebra_maps, int reg_type, const char **src, } static int term_100 (ZebraMaps zebra_maps, int reg_type, - const char **src, char *dst, int space_split) + const char **src, char *dst, int space_split, + char *dst_term) { const char *s0, *s1; const char **map; int i = 0; + int j = 0; if (!term_pre (zebra_maps, reg_type, src, NULL, NULL)) return 0; @@ -436,20 +545,24 @@ static int term_100 (ZebraMaps zebra_maps, int reg_type, { if (!isalnum (*s1) && *s1 != '-') dst[i++] = '\\'; + dst_term[j++] = *s1; dst[i++] = *s1++; } } dst[i] = '\0'; + dst_term[j] = '\0'; *src = s0; return i; } static int term_101 (ZebraMaps zebra_maps, int reg_type, - const char **src, char *dst, int space_split) + const char **src, char *dst, int space_split, + char *dst_term) { const char *s0, *s1; const char **map; int i = 0; + int j = 0; if (!term_pre (zebra_maps, reg_type, src, "#", "#")) return 0; @@ -460,7 +573,7 @@ static int term_101 (ZebraMaps zebra_maps, int reg_type, { dst[i++] = '.'; dst[i++] = '*'; - s0++; + dst_term[j++] = *s0++; } else { @@ -472,20 +585,24 @@ static int term_101 (ZebraMaps zebra_maps, int reg_type, { if (!isalnum (*s1)) dst[i++] = '\\'; + dst_term[j++] = *s1; dst[i++] = *s1++; } } } dst[i] = '\0'; + dst_term[j++] = '\0'; *src = s0; return i; } static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src, - char *dst, int *errors, int space_split) + char *dst, int *errors, int space_split, + char *dst_term) { int i = 0; + int j = 0; const char *s0, *s1; const char **map; @@ -503,7 +620,10 @@ static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src, while (*s0) { if (strchr ("^\\()[].*+?|-", *s0)) + { + dst_term[j++] = *s0; dst[i++] = *s0++; + } else { s1 = s0; @@ -514,19 +634,22 @@ static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src, { if (!isalnum (*s1)) dst[i++] = '\\'; + dst_term[j++] = *s1; dst[i++] = *s1++; } } } dst[i] = '\0'; + dst_term[j] = '\0'; *src = s0; return i; } static int term_102 (ZebraMaps zebra_maps, int reg_type, const char **src, - char *dst, int space_split) + char *dst, int space_split, char *dst_term) { - return term_103 (zebra_maps, reg_type, src, dst, NULL, space_split); + return term_103 (zebra_maps, reg_type, src, dst, NULL, space_split, + dst_term); } /* gen_regular_rel - generate regular expression from relation @@ -558,7 +681,7 @@ static void gen_regular_rel (char *dst, int val, int islt) } else { - strcpy (dst, "((-"); + strcpy (dst, "(-("); islt = 0; } val = -val; @@ -638,73 +761,188 @@ static void gen_regular_rel (char *dst, int val, int islt) strcat (dst, "))"); } -static int relational_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - const char **term_sub, - char *term_dict, - oid_value attributeSet, - struct grep_info *grep_info, - int *max_pos, - int reg_type) +void string_rel_add_char (char **term_p, const char *src, int *indx) +{ + if (src[*indx] == '\\') + *(*term_p)++ = src[(*indx)++]; + *(*term_p)++ = src[(*indx)++]; +} + +/* + * > abc ([b-].*|a[c-].*|ab[d-].*|abc.+) + * ([^-a].*|a[^-b].*ab[^-c].*|abc.+) + * >= abc ([b-].*|a[c-].*|ab[c-].*) + * ([^-a].*|a[^-b].*|ab[c-].*) + * < abc ([-0].*|a[-a].*|ab[-b].*) + * ([^a-].*|a[^b-].*|ab[^c-].*) + * <= abc ([-0].*|a[-a].*|ab[-b].*|abc) + * ([^a-].*|a[^b-].*|ab[^c-].*|abc) + */ +static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, char *term_dict, + oid_value attributeSet, + int reg_type, int space_split, char *term_dst) { AttrType relation; int relation_value; - int term_value; - int r; + int i; char *term_tmp = term_dict + strlen(term_dict); + char term_component[256]; attr_init (&relation, zapt, 2); relation_value = attr_find (&relation, NULL); - logf (LOG_DEBUG, "relation value=%d", relation_value); + logf (LOG_DEBUG, "string relation value=%d", relation_value); switch (relation_value) { case 1: - if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) + if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component, + space_split, term_dst)) return 0; - term_value = atoi (term_tmp); logf (LOG_DEBUG, "Relation <"); - gen_regular_rel (term_tmp, term_value-1, 1); + + *term_tmp++ = '('; + for (i = 0; term_component[i]; ) + { + int j = 0; + + if (i) + *term_tmp++ = '|'; + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); + + *term_tmp++ = '['; + + *term_tmp++ = '^'; + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '-'; + + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; + } + *term_tmp++ = ')'; + *term_tmp = '\0'; break; case 2: - if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) + if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component, + space_split, term_dst)) return 0; - term_value = atoi (term_tmp); logf (LOG_DEBUG, "Relation <="); - gen_regular_rel (term_tmp, term_value, 1); - break; - case 4: - if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) - return 0; - term_value = atoi (term_tmp); - logf (LOG_DEBUG, "Relation >="); - gen_regular_rel (term_tmp, term_value, 0); + + *term_tmp++ = '('; + for (i = 0; term_component[i]; ) + { + int j = 0; + + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); + *term_tmp++ = '['; + + *term_tmp++ = '^'; + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '-'; + + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; + + *term_tmp++ = '|'; + } + for (i = 0; term_component[i]; ) + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = ')'; + *term_tmp = '\0'; break; case 5: - if (!term_100 (zi->zebra_maps, reg_type, term_sub, term_tmp, 1)) + if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component, + space_split, term_dst)) return 0; - term_value = atoi (term_tmp); logf (LOG_DEBUG, "Relation >"); - gen_regular_rel (term_tmp, term_value+1, 0); + + *term_tmp++ = '('; + for (i = 0; term_component[i];) + { + int j = 0; + + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); + *term_tmp++ = '['; + + *term_tmp++ = '^'; + *term_tmp++ = '-'; + string_rel_add_char (&term_tmp, term_component, &i); + + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; + + *term_tmp++ = '|'; + } + for (i = 0; term_component[i];) + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '.'; + *term_tmp++ = '+'; + *term_tmp++ = ')'; + *term_tmp = '\0'; break; + case 4: + if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component, + space_split, term_dst)) + return 0; + logf (LOG_DEBUG, "Relation >="); + + *term_tmp++ = '('; + for (i = 0; term_component[i];) + { + int j = 0; + + if (i) + *term_tmp++ = '|'; + while (j < i) + string_rel_add_char (&term_tmp, term_component, &j); + *term_tmp++ = '['; + + if (term_component[i+1]) + { + *term_tmp++ = '^'; + *term_tmp++ = '-'; + string_rel_add_char (&term_tmp, term_component, &i); + } + else + { + string_rel_add_char (&term_tmp, term_component, &i); + *term_tmp++ = '-'; + } + *term_tmp++ = ']'; + *term_tmp++ = '.'; + *term_tmp++ = '*'; + } + *term_tmp++ = ')'; + *term_tmp = '\0'; + break; + case 3: default: - return 0; + logf (LOG_DEBUG, "Relation ="); + if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component, + space_split, term_dst)) + return 0; + strcat (term_tmp, "("); + strcat (term_tmp, term_component); + strcat (term_tmp, ")"); } - logf (LOG_DEBUG, "dict_lookup_grep: %s", term_tmp); - r = dict_lookup_grep (zi->dict, term_dict, 0, grep_info, max_pos, - 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 field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - const char **term_sub, - oid_value attributeSet, struct grep_info *grep_info, - int reg_type, int complete_flag, - int num_bases, char **basenames) +static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, + oid_value attributeSet, NMEM stream, + struct grep_info *grep_info, + int reg_type, int complete_flag, + int num_bases, char **basenames, + char *term_dst) { - char term_dict[2*IT_MAX_WORD+2]; + char term_dict[2*IT_MAX_WORD+4000]; int j, r, base_no; AttrType truncation; int truncation_value; @@ -715,10 +953,10 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, struct rpn_char_map_info rcmi; int space_split = complete_flag ? 0 : 1; - rpn_char_map_prepare (zi, reg_type, &rcmi); + rpn_char_map_prepare (zh, reg_type, &rcmi); attr_init (&use, zapt, 1); use_value = attr_find (&use, &curAttributeSet); - logf (LOG_DEBUG, "field_term, use value %d", use_value); + logf (LOG_DEBUG, "string_term, use value %d", use_value); attr_init (&truncation, zapt, 5); truncation_value = attr_find (&truncation, NULL); logf (LOG_DEBUG, "truncation value %d", truncation_value); @@ -733,28 +971,46 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, int max_pos, prefix_len = 0; termp = *term_sub; - if ((r=att_getentbyatt (zi, &attp, curAttributeSet, use_value))) + if ((r=att_getentbyatt (zh, &attp, curAttributeSet, use_value))) { logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d", curAttributeSet, use_value, r); if (r == -1) - zi->errCode = 114; + { + char val_str[32]; + sprintf (val_str, "%d", use_value); + zh->errCode = 114; + zh->errString = nmem_strdup (stream, val_str); + } else - zi->errCode = 121; + { + int oid[OID_SIZE]; + struct oident oident; + + oident.proto = PROTO_Z3950; + oident.oclass = CLASS_ATTSET; + oident.value = curAttributeSet; + oid_ent_to_oid (&oident, oid); + + zh->errCode = 121; + zh->errString = nmem_strdup (stream, oident.desc); + } return -1; } - if (zebTargetInfo_curDatabase (zi->zti, basenames[base_no])) + if (zebraExplain_curDatabase (zh->zei, basenames[base_no])) { - zi->errCode = 109; /* Database unavailable */ - zi->errString = basenames[base_no]; + zh->errCode = 109; /* Database unavailable */ + zh->errString = basenames[base_no]; return -1; } for (local_attr = attp.local_attributes; local_attr; local_attr = local_attr->next) { int ord; + char ord_buf[32]; + int i, ord_len; - ord = zebTargetInfo_lookupSU (zi->zti, attp.attset_ordinal, + ord = zebraExplain_lookupSU (zh->zei, attp.attset_ordinal, local_attr->local); if (ord < 0) continue; @@ -762,12 +1018,20 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, term_dict[prefix_len++] = '|'; else term_dict[prefix_len++] = '('; - term_dict[prefix_len++] = 1; - term_dict[prefix_len++] = ord; + + ord_len = key_SU_code (ord, ord_buf); + for (i = 0; ierrCode = 114; + char val_str[32]; + sprintf (val_str, "%d", use_value); + zh->errCode = 114; + zh->errString = nmem_strdup (stream, val_str); return -1; } term_dict[prefix_len++] = ')'; @@ -775,76 +1039,88 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, term_dict[prefix_len++] = reg_type; logf (LOG_DEBUG, "reg_type = %d", term_dict[prefix_len-1]); term_dict[prefix_len] = '\0'; - if (!relational_term (zi, zapt, &termp, term_dict, - attributeSet, grep_info, &max_pos, reg_type)) - { - j = prefix_len; - switch (truncation_value) - { - case -1: /* not specified */ - case 100: /* do not truncate */ - term_dict[j++] = '('; - if (!term_100 (zi->zebra_maps, reg_type, - &termp, term_dict + j, space_split)) - return 0; - strcat (term_dict, ")"); - r = dict_lookup_grep (zi->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=none:%d", r); - break; - case 1: /* right truncation */ - term_dict[j++] = '('; - if (!term_100 (zi->zebra_maps, reg_type, - &termp, term_dict + j, space_split)) - return 0; - strcat (term_dict, ".*)"); - dict_lookup_grep (zi->dict, term_dict, 0, grep_info, - &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 */ - term_dict[j++] = '('; - if (!term_101 (zi->zebra_maps, reg_type, - &termp, term_dict + j, space_split)) - return 0; - strcat (term_dict, ")"); - r = dict_lookup_grep (zi->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d", r); - break; - case 102: /* Regexp-1 */ - term_dict[j++] = '('; - if (!term_102 (zi->zebra_maps, reg_type, - &termp, term_dict + j, space_split)) - return 0; - strcat (term_dict, ")"); - logf (LOG_DEBUG, "Regexp-1 tolerance=%d", r); - r = dict_lookup_grep (zi->dict, term_dict, 0, grep_info, - &max_pos, 0, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", - r); - break; - case 103: /* Regexp-2 */ - r = 1; - term_dict[j++] = '('; - if (!term_103 (zi->zebra_maps, reg_type, - &termp, term_dict + j, &r, space_split)) - return 0; - strcat (term_dict, ")"); - logf (LOG_DEBUG, "Regexp-2 tolerance=%d", r); - r = dict_lookup_grep (zi->dict, term_dict, r, grep_info, - &max_pos, 2, grep_handle); - if (r) - logf (LOG_WARN, "dict_lookup_grep err, trunc=eregular: %d", - r); - break; - } + j = prefix_len; + switch (truncation_value) + { + case -1: /* not specified */ + case 100: /* do not truncate */ + if (!string_relation (zh, zapt, &termp, term_dict, + attributeSet, + reg_type, space_split, term_dst)) + return 0; + logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict+prefix_len); + r = dict_lookup_grep (zh->dict, term_dict, 0, grep_info, &max_pos, + 0, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r); + break; + case 1: /* right truncation */ + term_dict[j++] = '('; + if (!term_100 (zh->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + return 0; + strcat (term_dict, ".*)"); + dict_lookup_grep (zh->dict, term_dict, 0, grep_info, + &max_pos, 0, grep_handle); + break; + case 2: /* keft truncation */ + term_dict[j++] = '('; term_dict[j++] = '.'; term_dict[j++] = '*'; + if (!term_100 (zh->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + return 0; + strcat (term_dict, ")"); + dict_lookup_grep (zh->dict, term_dict, 0, grep_info, + &max_pos, 0, grep_handle); + break; + case 3: /* left&right truncation */ + term_dict[j++] = '('; term_dict[j++] = '.'; term_dict[j++] = '*'; + if (!term_100 (zh->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + return 0; + strcat (term_dict, ".*)"); + dict_lookup_grep (zh->dict, term_dict, 0, grep_info, + &max_pos, 0, grep_handle); + break; + zh->errCode = 120; + return -1; + case 101: /* process # in term */ + term_dict[j++] = '('; + if (!term_101 (zh->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + return 0; + strcat (term_dict, ")"); + r = dict_lookup_grep (zh->dict, term_dict, 0, grep_info, + &max_pos, 0, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep err, trunc=#: %d", r); + break; + case 102: /* Regexp-1 */ + term_dict[j++] = '('; + if (!term_102 (zh->zebra_maps, reg_type, + &termp, term_dict + j, space_split, term_dst)) + return 0; + strcat (term_dict, ")"); + logf (LOG_DEBUG, "Regexp-1 tolerance=%d", r); + r = dict_lookup_grep (zh->dict, term_dict, 0, grep_info, + &max_pos, 0, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", + r); + break; + case 103: /* Regexp-2 */ + r = 1; + term_dict[j++] = '('; + if (!term_103 (zh->zebra_maps, reg_type, + &termp, term_dict + j, &r, space_split, term_dst)) + return 0; + strcat (term_dict, ")"); + logf (LOG_DEBUG, "Regexp-2 tolerance=%d", r); + r = dict_lookup_grep (zh->dict, term_dict, r, grep_info, + &max_pos, 2, grep_handle); + if (r) + logf (LOG_WARN, "dict_lookup_grep err, trunc=eregular: %d", + r); + break; } } *term_sub = termp; @@ -852,7 +1128,7 @@ static int field_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, return 1; } -static void trans_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, +static void trans_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, char *termz) { size_t sizez; @@ -865,7 +1141,7 @@ static void trans_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, termz[sizez] = '\0'; } -static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, +static void trans_scan_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, char *termz, int reg_type) { Z_Term *term = zapt->term; @@ -879,7 +1155,7 @@ static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, while ((len = (cp_end - cp)) > 0) { - map = zebra_maps_input (zi->zebra_maps, reg_type, &cp, len); + map = zebra_maps_input (zh->zebra_maps, reg_type, &cp, len); if (**map == *CHR_SPACE) space_map = *map; else @@ -895,68 +1171,7 @@ static void trans_scan_term (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, termz[i] = '\0'; } -static RSET rpn_search_APT_relevance (ZServerInfo *zi, - Z_AttributesPlusTerm *zapt, - oid_value attributeSet, - int reg_type, int complete_flag, - int num_bases, char **basenames) -{ - rset_relevance_parms parms; - char termz[IT_MAX_WORD+1]; - const char *termp = termz; - struct grep_info grep_info; - RSET result; - int term_index = 0; - int r; - - parms.key_size = sizeof(struct it_key); - parms.max_rec = 1000; - parms.cmp = key_compare_it; - parms.get_pos = key_get_pos; - parms.is = zi->isam; - parms.isc = zi->isamc; - parms.no_terms = 0; - - if (zapt->term->which != Z_Term_general) - { - zi->errCode = 124; - return NULL; - } - trans_term (zi, zapt, termz); - -#ifdef TERM_COUNT - grep_info.term_no = 0; -#endif - grep_info.isam_p_indx = 0; - grep_info.isam_p_size = 0; - grep_info.isam_p_buf = NULL; - while (1) - { - r = field_term (zi, zapt, &termp, attributeSet, &grep_info, - reg_type, complete_flag, num_bases, basenames); - if (r <= 0) - break; -#ifdef TERM_COUNT - for (; term_index < grep_info.isam_p_indx; term_index++) - grep_info.term_no[term_index] = parms.no_terms; - parms.no_terms++; -#endif - } - parms.term_no = grep_info.term_no; - parms.isam_positions = grep_info.isam_p_buf; - parms.no_isam_positions = grep_info.isam_p_indx; - if (grep_info.isam_p_indx > 0) - result = rset_create (rset_kind_relevance, &parms); - else - result = rset_create (rset_kind_null, NULL); -#ifdef TERM_COUNT - xfree(grep_info.term_no); -#endif - xfree (grep_info.isam_p_buf); - return result; -} - -static RSET rpn_proximity (ZServerInfo *zi, RSET rset1, RSET rset2, +static RSET rpn_proximity (ZebraHandle zh, RSET rset1, RSET rset2, int ordered, int exclusion, int relation, int distance) { @@ -967,17 +1182,18 @@ static RSET rpn_proximity (ZServerInfo *zi, RSET rset1, RSET rset2, RSFD rsfd_result; RSET result; rset_temp_parms parms; + int term_index; - rsfd1 = rset_open (rset1, RSETF_READ|RSETF_SORT_SYSNO); - more1 = rset_read (rset1, rsfd1, &buf1); + rsfd1 = rset_open (rset1, RSETF_READ); + more1 = rset_read (rset1, rsfd1, &buf1, &term_index); - rsfd2 = rset_open (rset2, RSETF_READ|RSETF_SORT_SYSNO); - more2 = rset_read (rset2, rsfd2, &buf2); + rsfd2 = rset_open (rset2, RSETF_READ); + more2 = rset_read (rset2, rsfd2, &buf2, &term_index); parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zi->res, "setTmpDir"); + parms.temp_path = res_get (zh->res, "setTmpDir"); result = rset_create (rset_kind_temp, &parms); - rsfd_result = rset_open (result, RSETF_WRITE|RSETF_SORT_SYSNO); + rsfd_result = rset_open (result, RSETF_WRITE); logf (LOG_DEBUG, "rpn_proximity excl=%d ord=%d rel=%d dis=%d", exclusion, ordered, relation, distance); @@ -985,9 +1201,9 @@ static RSET rpn_proximity (ZServerInfo *zi, RSET rset1, RSET rset2, { int cmp = key_compare_it (&buf1, &buf2); if (cmp < -1) - more1 = rset_read (rset1, rsfd1, &buf1); + more1 = rset_read (rset1, rsfd1, &buf1, &term_index); else if (cmp > 1) - more2 = rset_read (rset2, rsfd2, &buf2); + more2 = rset_read (rset2, rsfd2, &buf2, &term_index); else { int sysno = buf1.sysno; @@ -995,7 +1211,7 @@ static RSET rpn_proximity (ZServerInfo *zi, RSET rset1, RSET rset2, int n = 0; seqno[n++] = buf1.seqno; - while ((more1 = rset_read (rset1, rsfd1, &buf1)) && + while ((more1 = rset_read (rset1, rsfd1, &buf1, &term_index)) && sysno == buf1.sysno) if (n < 500) seqno[n++] = buf1.seqno; @@ -1037,7 +1253,7 @@ static RSET rpn_proximity (ZServerInfo *zi, RSET rset1, RSET rset2, if (excl) rset_write (result, rsfd_result, &buf2); } - } while ((more2 = rset_read (rset2, rsfd2, &buf2)) && + } while ((more2 = rset_read (rset2, rsfd2, &buf2, &term_index)) && sysno == buf2.sysno); } } @@ -1047,130 +1263,162 @@ static RSET rpn_proximity (ZServerInfo *zi, RSET rset1, RSET rset2, return result; } -static RSET rpn_prox (ZServerInfo *zi, RSET *rset, int rset_no) +static RSET rpn_prox (ZebraHandle zh, RSET *rset, int rset_no) { int i; RSFD *rsfd; int *more; struct it_key **buf; - RSFD rsfd_result; RSET result; - rset_temp_parms parms; + char prox_term[1024]; + int length_prox_term = 0; + int min_nn = 10000000; + int term_index; + const char *flags = NULL; - rsfd = xmalloc (sizeof(*rsfd)*rset_no); - more = xmalloc (sizeof(*more)*rset_no); - buf = xmalloc (sizeof(*buf)*rset_no); + rsfd = (RSFD *) xmalloc (sizeof(*rsfd)*rset_no); + more = (int *) xmalloc (sizeof(*more)*rset_no); + buf = (struct it_key **) xmalloc (sizeof(*buf)*rset_no); for (i = 0; i= 0) - { - rset_close (rset[i], rsfd[i]); - xfree (buf[i]); - --i; - } - xfree (rsfd); - xfree (more); - xfree (buf); - return rset_create (rset_kind_null, NULL); - } + int j; + buf[i] = (struct it_key *) xmalloc (sizeof(**buf)); + rsfd[i] = rset_open (rset[i], RSETF_READ); + if (!(more[i] = rset_read (rset[i], rsfd[i], buf[i], &term_index))) + break; + for (j = 0; jno_rset_terms; j++) + { + const char *nflags = rset[i]->rset_terms[j]->flags; + char *term = rset[i]->rset_terms[j]->name; + int lterm = strlen(term); + if (length_prox_term) + prox_term[length_prox_term++] = ' '; + strcpy (prox_term + length_prox_term, term); + length_prox_term += lterm; + if (min_nn > rset[i]->rset_terms[j]->nn) + min_nn = rset[i]->rset_terms[j]->nn; + flags = nflags; + } } - parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zi->res, "setTmpDir"); - result = rset_create (rset_kind_temp, &parms); - rsfd_result = rset_open (result, RSETF_WRITE|RSETF_SORT_SYSNO); - - while (*more) + if (i != rset_no) { - for (i = 1; i 1) - { - more[i-1] = rset_read (rset[i-1], rsfd[i-1], buf[i-1]); - break; - } - else if (cmp == 1) - { - if (buf[i-1]->seqno+1 != buf[i]->seqno) - { - more[i-1] = rset_read (rset[i-1], rsfd[i-1], buf[i-1]); - break; - } - } - else - { - more[i] = rset_read (rset[i], rsfd[i], buf[i]); - break; - } - } - if (i == rset_no) - { - rset_write (result, rsfd_result, buf[0]); - more[0] = rset_read (*rset, *rsfd, *buf); - } + rset_null_parms parms; + + while (i >= 0) + { + rset_close (rset[i], rsfd[i]); + xfree (buf[i]); + --i; + } + parms.rset_term = rset_term_create (prox_term, -1, flags); + parms.rset_term->nn = 0; + result = rset_create (rset_kind_null, &parms); } - - for (i = 0; inn = min_nn; + parms.key_size = sizeof (struct it_key); + parms.temp_path = res_get (zh->res, "setTmpDir"); + result = rset_create (rset_kind_temp, &parms); + rsfd_result = rset_open (result, RSETF_WRITE); + + while (*more) + { + for (i = 1; i 1) + { + more[i-1] = rset_read (rset[i-1], rsfd[i-1], + buf[i-1], &term_index); + break; + } + else if (cmp == 1) + { + if (buf[i-1]->seqno+1 != buf[i]->seqno) + { + more[i-1] = rset_read (rset[i-1], rsfd[i-1], + buf[i-1], &term_index); + break; + } + } + else + { + more[i] = rset_read (rset[i], rsfd[i], buf[i], + &term_index); + break; + } + } + if (i == rset_no) + { + rset_write (result, rsfd_result, buf[0]); + more[0] = rset_read (*rset, *rsfd, *buf, &term_index); + } + } + + for (i = 0; iterm->which != Z_Term_general) - { - zi->errCode = 124; - return NULL; - } - trans_term (zi, zapt, termz); - #ifdef TERM_COUNT grep_info.term_no = 0; #endif grep_info.isam_p_size = 0; grep_info.isam_p_buf = NULL; + grep_info.zh = zh; + grep_info.reg_type = reg_type; while (1) { logf (LOG_DEBUG, "APT_phrase termp=%s", termp); grep_info.isam_p_indx = 0; - r = field_term (zi, zapt, &termp, attributeSet, &grep_info, - reg_type, complete_flag, num_bases, basenames); + r = string_term (zh, zapt, &termp, attributeSet, stream, &grep_info, + reg_type, complete_flag, num_bases, basenames, + term_dst); if (r < 1) break; - rset[rset_no] = rset_trunc (zi, grep_info.isam_p_buf, - grep_info.isam_p_indx); + logf (LOG_DEBUG, "term: %s", term_dst); + rset[rset_no] = rset_trunc (zh, grep_info.isam_p_buf, + grep_info.isam_p_indx, term_dst, + strlen(term_dst), rank_type); assert (rset[rset_no]); - if (++rset_no >= sizeof(rset)/sizeof(*rset)) + if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset))) break; } #ifdef TERM_COUNT @@ -1178,37 +1426,379 @@ static RSET rpn_search_APT_phrase (ZServerInfo *zi, #endif xfree (grep_info.isam_p_buf); if (rset_no == 0) - return rset_create (rset_kind_null, NULL); + { + rset_null_parms parms; + + parms.rset_term = rset_term_create (term_dst, -1, rank_type); + return rset_create (rset_kind_null, &parms); + } else if (rset_no == 1) return (rset[0]); - result = rpn_prox (zi, rset, rset_no); + result = rpn_prox (zh, rset, rset_no); for (i = 0; i= (int) (sizeof(rset)/sizeof(*rset))) + break; + } +#ifdef TERM_COUNT + xfree(grep_info.term_no); +#endif + xfree (grep_info.isam_p_buf); + if (rset_no == 0) + { + rset_null_parms parms; + + parms.rset_term = rset_term_create (term_dst, -1, rank_type); + return rset_create (rset_kind_null, &parms); + } + result = rset[0]; + for (i = 1; i= (int) (sizeof(rset)/sizeof(*rset))) + break; + } +#ifdef TERM_COUNT + xfree(grep_info.term_no); +#endif + xfree (grep_info.isam_p_buf); + if (rset_no == 0) + { + rset_null_parms parms; + + parms.rset_term = rset_term_create (term_dst, -1, rank_type); + return rset_create (rset_kind_null, &parms); + } + result = rset[0]; + for (i = 1; izebra_maps, reg_type, term_sub, term_tmp, 1, + term_dst)) + return 0; + term_value = atoi (term_tmp); + switch (relation_value) + { + case 1: + logf (LOG_DEBUG, "Relation <"); + gen_regular_rel (term_tmp, term_value-1, 1); + break; + case 2: + logf (LOG_DEBUG, "Relation <="); + gen_regular_rel (term_tmp, term_value, 1); + break; + case 4: + logf (LOG_DEBUG, "Relation >="); + gen_regular_rel (term_tmp, term_value, 0); + break; + case 5: + logf (LOG_DEBUG, "Relation >"); + gen_regular_rel (term_tmp, term_value+1, 0); + break; + case 3: + default: + logf (LOG_DEBUG, "Relation ="); + sprintf (term_tmp, "(0*%d)", term_value); + } + logf (LOG_DEBUG, "dict_lookup_grep: %s", term_tmp); + r = dict_lookup_grep (zh->dict, term_dict, 0, grep_info, max_pos, + 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 numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + const char **term_sub, + oid_value attributeSet, struct grep_info *grep_info, + int reg_type, int complete_flag, + int num_bases, char **basenames, + char *term_dst) +{ + char term_dict[2*IT_MAX_WORD+2]; + int r, base_no; + AttrType use; + int use_value; + oid_value curAttributeSet = attributeSet; + const char *termp; + struct rpn_char_map_info rcmi; + + rpn_char_map_prepare (zh, reg_type, &rcmi); + attr_init (&use, zapt, 1); + use_value = attr_find (&use, &curAttributeSet); + logf (LOG_DEBUG, "numeric_term, use value %d", use_value); + + if (use_value == -1) + use_value = 1016; + + for (base_no = 0; base_no < num_bases; base_no++) + { + attent attp; + data1_local_attribute *local_attr; + int max_pos, prefix_len = 0; + + termp = *term_sub; + if ((r=att_getentbyatt (zh, &attp, curAttributeSet, use_value))) + { + logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d", + curAttributeSet, use_value, r); + if (r == -1) + zh->errCode = 114; + else + zh->errCode = 121; + return -1; + } + if (zebraExplain_curDatabase (zh->zei, basenames[base_no])) + { + zh->errCode = 109; /* Database unavailable */ + zh->errString = basenames[base_no]; + return -1; + } + for (local_attr = attp.local_attributes; local_attr; + local_attr = local_attr->next) + { + int ord; + char ord_buf[32]; + int i, ord_len; + + ord = zebraExplain_lookupSU (zh->zei, attp.attset_ordinal, + local_attr->local); + if (ord < 0) + continue; + if (prefix_len) + term_dict[prefix_len++] = '|'; + else + term_dict[prefix_len++] = '('; + + ord_len = key_SU_code (ord, ord_buf); + for (i = 0; ierrCode = 114; + return -1; + } + term_dict[prefix_len++] = ')'; + term_dict[prefix_len++] = 1; + term_dict[prefix_len++] = reg_type; + logf (LOG_DEBUG, "reg_type = %d", term_dict[prefix_len-1]); + term_dict[prefix_len] = '\0'; + if (!numeric_relation (zh, zapt, &termp, term_dict, + attributeSet, grep_info, &max_pos, reg_type, + term_dst)) + return 0; + } + *term_sub = termp; + logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); + return 1; +} + +static RSET rpn_search_APT_numeric (ZebraHandle zh, + Z_AttributesPlusTerm *zapt, + const char *termz, + oid_value attributeSet, + NMEM stream, + int reg_type, int complete_flag, + const char *rank_type, + int num_bases, char **basenames) +{ + char term_dst[IT_MAX_WORD+1]; + const char *termp = termz; + RSET rset[60], result; + int i, r, rset_no = 0; + struct grep_info grep_info; + +#ifdef TERM_COUNT + grep_info.term_no = 0; +#endif + grep_info.isam_p_size = 0; + grep_info.isam_p_buf = NULL; + grep_info.zh = zh; + grep_info.reg_type = reg_type; + + while (1) + { + logf (LOG_DEBUG, "APT_numeric termp=%s", termp); + grep_info.isam_p_indx = 0; + r = numeric_term (zh, zapt, &termp, attributeSet, &grep_info, + reg_type, complete_flag, num_bases, basenames, + term_dst); + if (r < 1) + break; + logf (LOG_DEBUG, "term: %s", term_dst); + rset[rset_no] = rset_trunc (zh, grep_info.isam_p_buf, + grep_info.isam_p_indx, term_dst, + strlen(term_dst), rank_type); + assert (rset[rset_no]); + if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset))) + break; + } +#ifdef TERM_COUNT + xfree(grep_info.term_no); +#endif + xfree (grep_info.isam_p_buf); + if (rset_no == 0) + { + rset_null_parms parms; + + parms.rset_term = rset_term_create (term_dst, -1, rank_type); + return rset_create (rset_kind_null, &parms); + } + result = rset[0]; + for (i = 1; iterm->which != Z_Term_general) - { - zi->errCode = 124; - return NULL; - } + parms.rset_term = rset_term_create (termz, -1, rank_type); parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zi->res, "setTmpDir"); + parms.temp_path = res_get (zh->res, "setTmpDir"); result = rset_create (rset_kind_temp, &parms); - rsfd = rset_open (result, RSETF_WRITE|RSETF_SORT_SYSNO); - - trans_term (zi, zapt, termz); + rsfd = rset_open (result, RSETF_WRITE); key.sysno = atoi (termz); + key.seqno = 1; if (key.sysno <= 0) key.sysno = 1; rset_write (result, rsfd, &key); @@ -1216,51 +1806,173 @@ static RSET rpn_search_APT_local (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, return result; } -static RSET rpn_search_APT (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, - oid_value attributeSet, +static RSET rpn_sort_spec (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + oid_value attributeSet, NMEM stream, + Z_SortKeySpecList *sort_sequence, + const char *rank_type) +{ + rset_null_parms parms; + int i; + int sort_relation_value; + AttrType sort_relation_type; + int use_value; + AttrType use_type; + Z_SortKeySpec *sks; + Z_SortKey *sk; + Z_AttributeElement *ae; + int oid[OID_SIZE]; + oident oe; + + attr_init (&sort_relation_type, zapt, 7); + sort_relation_value = attr_find (&sort_relation_type, &attributeSet); + + attr_init (&use_type, zapt, 1); + use_value = attr_find (&use_type, &attributeSet); + + if (!sort_sequence->specs) + { + sort_sequence->num_specs = 10; + sort_sequence->specs = (Z_SortKeySpec **) + nmem_malloc (stream, sort_sequence->num_specs * + sizeof(*sort_sequence->specs)); + for (i = 0; inum_specs; i++) + sort_sequence->specs[i] = 0; + } + if (zapt->term->which != Z_Term_general) + i = 0; + else + i = atoi_n ((char *) zapt->term->u.general->buf, + zapt->term->u.general->len); + if (i >= sort_sequence->num_specs) + i = 0; + + oe.proto = PROTO_Z3950; + oe.oclass = CLASS_ATTSET; + oe.value = attributeSet; + if (!oid_ent_to_oid (&oe, oid)) + return 0; + + sks = (Z_SortKeySpec *) nmem_malloc (stream, sizeof(*sks)); + sks->sortElement = (Z_SortElement *) + nmem_malloc (stream, sizeof(*sks->sortElement)); + sks->sortElement->which = Z_SortElement_generic; + sk = sks->sortElement->u.generic = (Z_SortKey *) + nmem_malloc (stream, sizeof(*sk)); + sk->which = Z_SortKey_sortAttributes; + sk->u.sortAttributes = (Z_SortAttributes *) + nmem_malloc (stream, sizeof(*sk->u.sortAttributes)); + + sk->u.sortAttributes->id = oid; + sk->u.sortAttributes->list = (Z_AttributeList *) + nmem_malloc (stream, sizeof(*sk->u.sortAttributes->list)); + sk->u.sortAttributes->list->num_attributes = 1; + sk->u.sortAttributes->list->attributes = (Z_AttributeElement **) + nmem_malloc (stream, sizeof(*sk->u.sortAttributes->list->attributes)); + ae = *sk->u.sortAttributes->list->attributes = (Z_AttributeElement *) + nmem_malloc (stream, sizeof(**sk->u.sortAttributes->list->attributes)); + ae->attributeSet = 0; + ae->attributeType = (int *) + nmem_malloc (stream, sizeof(*ae->attributeType)); + *ae->attributeType = 1; + ae->which = Z_AttributeValue_numeric; + ae->value.numeric = (int *) + nmem_malloc (stream, sizeof(*ae->value.numeric)); + *ae->value.numeric = use_value; + + sks->sortRelation = (int *) + nmem_malloc (stream, sizeof(*sks->sortRelation)); + if (sort_relation_value == 1) + *sks->sortRelation = Z_SortRelation_ascending; + else if (sort_relation_value == 2) + *sks->sortRelation = Z_SortRelation_descending; + else + *sks->sortRelation = Z_SortRelation_ascending; + + sks->caseSensitivity = (int *) + nmem_malloc (stream, sizeof(*sks->caseSensitivity)); + *sks->caseSensitivity = 0; + +#ifdef ASN_COMPILED + sks->which = Z_SortKeySpec_null; + sks->u.null = odr_nullval (); +#else + sks->missingValueAction = 0; +#endif + + sort_sequence->specs[i] = sks; + + parms.rset_term = rset_term_create ("", -1, rank_type); + return rset_create (rset_kind_null, &parms); +} + + +static RSET rpn_search_APT (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + oid_value attributeSet, NMEM stream, + Z_SortKeySpecList *sort_sequence, int num_bases, char **basenames) { - int reg_type; + unsigned reg_id; char *search_type = NULL; + char *rank_type = NULL; int complete_flag; + int sort_flag; + char termz[IT_MAX_WORD+1]; - zebra_maps_attr (zi->zebra_maps, zapt, ®_type, &search_type, - &complete_flag); + zebra_maps_attr (zh->zebra_maps, zapt, ®_id, &search_type, + &rank_type, &complete_flag, &sort_flag); - logf (LOG_DEBUG, "reg_type=%c", reg_type); + logf (LOG_DEBUG, "reg_id=%c", reg_id); logf (LOG_DEBUG, "complete_flag=%d", complete_flag); logf (LOG_DEBUG, "search_type=%s", search_type); + logf (LOG_DEBUG, "rank_type=%s", rank_type); + + if (zapt->term->which != Z_Term_general) + { + zh->errCode = 124; + return NULL; + } + trans_term (zh, zapt, termz); + + if (sort_flag) + return rpn_sort_spec (zh, zapt, attributeSet, stream, sort_sequence, + rank_type); + if (!strcmp (search_type, "phrase")) { - return rpn_search_APT_phrase (zi, zapt, attributeSet, - reg_type, complete_flag, + return rpn_search_APT_phrase (zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, num_bases, basenames); } - else if (!strcmp (search_type, "ranked")) + else if (!strcmp (search_type, "and-list")) { - return rpn_search_APT_relevance (zi, zapt, attributeSet, - reg_type, complete_flag, - num_bases, basenames); + return rpn_search_APT_and_list (zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, + num_bases, basenames); + } + else if (!strcmp (search_type, "or-list")) + { + return rpn_search_APT_or_list (zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, + num_bases, basenames); } else if (!strcmp (search_type, "local")) { - return rpn_search_APT_local (zi, zapt, attributeSet); + return rpn_search_APT_local (zh, zapt, termz, attributeSet, stream, + rank_type); + } + else if (!strcmp (search_type, "numeric")) + { + return rpn_search_APT_numeric (zh, zapt, termz, attributeSet, stream, + reg_id, complete_flag, rank_type, + num_bases, basenames); } - zi->errCode = 118; + zh->errCode = 118; return NULL; } -static RSET rpn_search_ref (ZServerInfo *zi, Z_ResultSetId *resultSetId) -{ - ZServerSet *s; - - if (!(s = resultSetGet (zi, resultSetId))) - return rset_create (rset_kind_null, NULL); - return s->rset; -} - -static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, - oid_value attributeSet, +static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, + oid_value attributeSet, NMEM stream, + Z_SortKeySpecList *sort_sequence, int num_bases, char **basenames) { RSET r = NULL; @@ -1268,55 +1980,71 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, { Z_Operator *zop = zs->u.complex->roperator; rset_bool_parms bool_parms; - int soft = 0; - - bool_parms.rset_l = rpn_search_structure (zi, zs->u.complex->s1, - attributeSet, + bool_parms.rset_l = rpn_search_structure (zh, zs->u.complex->s1, + attributeSet, stream, + sort_sequence, 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, + bool_parms.rset_r = rpn_search_structure (zh, zs->u.complex->s2, + attributeSet, stream, + sort_sequence, num_bases, basenames); if (bool_parms.rset_r == NULL) { 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_it; switch (zop->which) { case Z_Operator_and: - r = rset_create (soft ? rset_kind_sand:rset_kind_and, &bool_parms); + r = rset_create (rset_kind_and, &bool_parms); break; case Z_Operator_or: - r = rset_create (soft ? rset_kind_sor:rset_kind_or, &bool_parms); + r = rset_create (rset_kind_or, &bool_parms); break; case Z_Operator_and_not: - r = rset_create (soft ? rset_kind_snot:rset_kind_not, &bool_parms); + r = rset_create (rset_kind_not, &bool_parms); break; case Z_Operator_prox: +#ifdef ASN_COMPILED + if (zop->u.prox->which != Z_ProximityOperator_known) + { + zh->errCode = 132; + return NULL; + } +#else if (zop->u.prox->which != Z_ProxCode_known) { - zi->errCode = 132; + zh->errCode = 132; return NULL; } +#endif + +#ifdef ASN_COMPILED + if (*zop->u.prox->u.known != Z_ProxUnit_word) + { + char *val = (char *) nmem_malloc (stream, 16); + zh->errCode = 132; + zh->errString = val; + sprintf (val, "%d", *zop->u.prox->u.known); + return NULL; + } +#else if (*zop->u.prox->proximityUnitCode != Z_ProxUnit_word) { - char *val = odr_malloc (zi->odr, 16); - zi->errCode = 132; - zi->errString = val; + char *val = (char *) nmem_malloc (stream, 16); + zh->errCode = 132; + zh->errString = val; sprintf (val, "%d", *zop->u.prox->proximityUnitCode); return NULL; } - r = rpn_proximity (zi, bool_parms.rset_l, bool_parms.rset_r, +#endif + r = rpn_proximity (zh, bool_parms.rset_l, bool_parms.rset_r, *zop->u.prox->ordered, (!zop->u.prox->exclusion ? 0 : *zop->u.prox->exclusion), @@ -1324,7 +2052,7 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, *zop->u.prox->distance); break; default: - zi->errCode = 110; + zh->errCode = 110; return NULL; } } @@ -1333,113 +2061,82 @@ static RSET rpn_search_structure (ZServerInfo *zi, Z_RPNStructure *zs, if (zs->u.simple->which == Z_Operand_APT) { logf (LOG_DEBUG, "rpn_search_APT"); - r = rpn_search_APT (zi, zs->u.simple->u.attributesPlusTerm, - attributeSet, num_bases, basenames); + r = rpn_search_APT (zh, zs->u.simple->u.attributesPlusTerm, + attributeSet, stream, sort_sequence, + num_bases, basenames); } else if (zs->u.simple->which == Z_Operand_resultSetId) { logf (LOG_DEBUG, "rpn_search_ref"); - r = rpn_search_ref (zi, zs->u.simple->u.resultSetId); + r = resultSetRef (zh, zs->u.simple->u.resultSetId); + if (!r) + r = rset_create (rset_kind_null, NULL); } else { - zi->errCode = 3; + zh->errCode = 3; return NULL; } } else { - zi->errCode = 3; + zh->errCode = 3; return NULL; } return r; } -void count_set_save (ZServerInfo *zi, RSET *r, int *count) -{ - int psysno = 0; - int kno = 0; - struct it_key key; - RSFD rfd, wfd; - RSET w; - rset_temp_parms parms; - int maxResultSetSize = atoi (res_get_def (zi->res, - "maxResultSetSize", "400")); - logf (LOG_DEBUG, "count_set_save"); - *count = 0; - parms.key_size = sizeof(struct it_key); - parms.temp_path = res_get (zi->res, "setTmpDir"); - 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) - { - if (*count < maxResultSetSize) - rset_write (w, wfd, &key); - (*count)++; - psysno = key.sysno; - } - 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, "count_set"); - *count = 0; - rfd = rset_open (r, RSETF_READ|RSETF_SORT_SYSNO); - while (rset_read (r, rfd, &key)) - { - if (key.sysno != psysno) - { - psysno = key.sysno; - (*count)++; - } - kno++; - } - rset_close (r, rfd); - logf (LOG_DEBUG, "%d keys, %d distinct sysnos", kno, *count); -} - -int rpn_search (ZServerInfo *zi, - Z_RPNQuery *rpn, int num_bases, char **basenames, - const char *setname, int *hits) +RSET rpn_search (ZebraHandle zh, NMEM nmem, + Z_RPNQuery *rpn, int num_bases, char **basenames, + const char *setname, + ZebraSet sset) { RSET rset; oident *attrset; oid_value attributeSet; - - zlog_rpn (rpn); - - zi->errCode = 0; - zi->errString = NULL; - + Z_SortKeySpecList *sort_sequence; + int sort_status, i; + + zh->errCode = 0; + zh->errString = NULL; + zh->hits = 0; + + sort_sequence = (Z_SortKeySpecList *) + nmem_malloc (nmem, sizeof(*sort_sequence)); + sort_sequence->num_specs = 10; + sort_sequence->specs = (Z_SortKeySpec **) + nmem_malloc (nmem, sort_sequence->num_specs * + sizeof(*sort_sequence->specs)); + for (i = 0; inum_specs; i++) + sort_sequence->specs[i] = 0; + attrset = oid_getentbyoid (rpn->attributeSetId); attributeSet = attrset->value; - rset = rpn_search_structure (zi, rpn->RPNStructure, attributeSet, - num_bases, basenames); + rset = rpn_search_structure (zh, rpn->RPNStructure, attributeSet, + nmem, sort_sequence, num_bases, basenames); if (!rset) - return zi->errCode; - if (rset_is_volatile(rset)) - count_set_save(zi, &rset,hits); + return 0; + + if (zh->errCode) + logf (LOG_DEBUG, "search error: %d", zh->errCode); + + for (i = 0; sort_sequence->specs[i]; i++) + ; + sort_sequence->num_specs = i; + if (!i) + resultSetRank (zh, sset, rset); else - count_set (rset, hits); - resultSetAdd (zi, setname, 1, rset); - if (zi->errCode) - logf (LOG_DEBUG, "search error: %d", zi->errCode); - return zi->errCode; + { + logf (LOG_DEBUG, "resultSetSortSingle in rpn_search"); + resultSetSortSingle (zh, nmem, sset, rset, + sort_sequence, &sort_status); + if (zh->errCode) + { + logf (LOG_DEBUG, "resultSetSortSingle status = %d", zh->errCode); + } + } + return rset; } struct scan_info_entry { @@ -1457,7 +2154,7 @@ struct scan_info { static int scan_handle (char *name, const char *info, int pos, void *client) { int len_prefix, idx; - struct scan_info *scan_info = client; + struct scan_info *scan_info = (struct scan_info *) client; len_prefix = strlen(scan_info->prefix); if (memcmp (name, scan_info->prefix, len_prefix)) @@ -1466,35 +2163,55 @@ static int scan_handle (char *name, const char *info, int pos, void *client) idx = scan_info->after - pos + scan_info->before; else idx = - pos - 1; - scan_info->list[idx].term = odr_malloc (scan_info->odr, - strlen(name + len_prefix)+1); + scan_info->list[idx].term = (char *) + 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 (&scan_info->list[idx].isam_p, info+1, sizeof(ISAM_P)); return 0; } +static void scan_term_untrans (ZebraHandle zh, NMEM stream, int reg_type, + char **dst, const char *src) +{ + char term_dst[1024]; + + term_untrans (zh, reg_type, term_dst, src); + + *dst = (char *) nmem_malloc (stream, strlen(term_dst)+1); + strcpy (*dst, term_dst); +} -static void scan_term_untrans (ZServerInfo *zi, int reg_type, - char **dstp, const char *src) -{ - char *dst = odr_malloc (zi->odr, strlen(src)*2+1); - *dstp = dst; +static void count_set (RSET r, int *count) +{ + int psysno = 0; + int kno = 0; + struct it_key key; + RSFD rfd; + int term_index; - while (*src) + logf (LOG_DEBUG, "count_set"); + + *count = 0; + rfd = rset_open (r, RSETF_READ); + while (rset_read (r, rfd, &key, &term_index)) { - const char *cp = zebra_maps_output (zi->zebra_maps, reg_type, &src); - while (*cp) - *dst++ = *cp++; + if (key.sysno != psysno) + { + psysno = key.sysno; + (*count)++; + } + kno++; } - *dst = '\0'; + rset_close (r, rfd); + logf (LOG_DEBUG, "%d keys, %d records", kno, *count); } -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) +void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, + oid_value attributeset, + int num_bases, char **basenames, + int *position, int *num_entries, ZebraScanEntry **list, + int *is_partial) { int i; int pos = *position; @@ -1506,28 +2223,29 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, AttrType use; int use_value; struct scan_info *scan_info_array; - struct scan_entry *glist; + ZebraScanEntry *glist; int ords[32], ord_no = 0; int ptr[32]; - int reg_type; + unsigned reg_id; char *search_type = NULL; + char *rank_type = NULL; int complete_flag; - - logf (LOG_DEBUG, "scan, position = %d, num = %d", pos, num); + int sort_flag; if (attributeset == VAL_NONE) attributeset = VAL_BIB1; + + logf (LOG_DEBUG, "position = %d, num = %d", pos, num); attr_init (&use, zapt, 1); use_value = attr_find (&use, &attributeset); - logf (LOG_DEBUG, "use value %d", use_value); - if (zebra_maps_attr (zi->zebra_maps, zapt, ®_type, &search_type, - &complete_flag)) + if (zebra_maps_attr (zh->zebra_maps, zapt, ®_id, &search_type, + &rank_type, &complete_flag, &sort_flag)) { - zi->errCode = 113; - return zi->errCode; + zh->errCode = 113; + return ; } if (use_value == -1) @@ -1538,36 +2256,42 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, attent attp; data1_local_attribute *local_attr; - if ((r=att_getentbyatt (zi, &attp, attributeset, use_value))) + if ((r=att_getentbyatt (zh, &attp, attributeset, use_value))) { logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d", attributeset, use_value); if (r == -1) - zi->errCode = 114; + zh->errCode = 114; else - zi->errCode = 121; + zh->errCode = 121; } - if (zebTargetInfo_curDatabase (zi->zti, basenames[base_no])) + if (zebraExplain_curDatabase (zh->zei, basenames[base_no])) { - zi->errString = basenames[base_no]; - return zi->errCode = 109; /* Database unavailable */ + zh->errString = basenames[base_no]; + zh->errCode = 109; /* Database unavailable */ + return; } for (local_attr = attp.local_attributes; local_attr && ord_no < 32; local_attr = local_attr->next) { int ord; - ord = zebTargetInfo_lookupSU (zi->zti, attp.attset_ordinal, - local_attr->local); + ord = zebraExplain_lookupSU (zh->zei, attp.attset_ordinal, + local_attr->local); if (ord > 0) ords[ord_no++] = ord; } } if (ord_no == 0) - return zi->errCode = 113; + { + zh->errCode = 113; + return; + } + /* prepare dictionary scanning */ before = pos-1; after = 1+num-pos; - scan_info_array = odr_malloc (zi->odr, ord_no * sizeof(*scan_info_array)); + scan_info_array = (struct scan_info *) + odr_malloc (stream, ord_no * sizeof(*scan_info_array)); for (i = 0; i < ord_no; i++) { int j, prefix_len = 0; @@ -1575,31 +2299,35 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, struct scan_info *scan_info = scan_info_array + i; struct rpn_char_map_info rcmi; - rpn_char_map_prepare (zi, reg_type, &rcmi); + rpn_char_map_prepare (zh, reg_id, &rcmi); scan_info->before = before; scan_info->after = after; - scan_info->odr = zi->odr; + scan_info->odr = stream; - scan_info->list = odr_malloc (zi->odr, (before+after)* - sizeof(*scan_info->list)); + scan_info->list = (struct scan_info_entry *) + odr_malloc (stream, (before+after) * sizeof(*scan_info->list)); for (j = 0; jlist[j].term = NULL; - termz[prefix_len++] = ords[i]; - termz[prefix_len++] = reg_type; + + prefix_len += key_SU_code (ords[i], termz + prefix_len); + termz[prefix_len++] = reg_id; termz[prefix_len] = 0; strcpy (scan_info->prefix, termz); - trans_scan_term (zi, zapt, termz+prefix_len, reg_type); + trans_scan_term (zh, zapt, termz+prefix_len, reg_id); - dict_scan (zi->dict, termz, &before_tmp, &after_tmp, scan_info, + dict_scan (zh->dict, termz, &before_tmp, &after_tmp, scan_info, scan_handle); } - glist = odr_malloc (zi->odr, (before+after)*sizeof(*glist)); + glist = (ZebraScanEntry *) + odr_malloc (stream, (before+after)*sizeof(*glist)); + + /* consider terms after main term */ for (i = 0; i < ord_no; i++) ptr[i] = before; - *status = BEND_SCAN_SUCCESS; + *is_partial = 0; for (i = 0; imem, reg_id, + &glist[i+before].term, mterm); + rset = rset_trunc (zh, &scan_info_array[j0].list[ptr[j0]].isam_p, 1, + glist[i+before].term, strlen(glist[i+before].term), + NULL); ptr[j0]++; for (j = j0+1; jmem, reg_id, + &glist[before-1-i].term, mterm); rset = rset_trunc - (zi, &scan_info_array[j0].list[before-1-ptr[j0]].isam_p, 1); + (zh, &scan_info_array[j0].list[before-1-ptr[j0]].isam_p, 1, + glist[before-1-i].term, strlen(glist[before-1-i].term), + NULL); ptr[j0]++; @@ -1693,8 +2430,10 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, rset_bool_parms bool_parms; RSET rset2; - rset2 = rset_trunc (zi, - &scan_info_array[j].list[before-1-ptr[j]].isam_p, 1); + rset2 = rset_trunc (zh, + &scan_info_array[j].list[before-1-ptr[j]].isam_p, 1, + glist[before-1-i].term, + strlen(glist[before-1-i].term), NULL); bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare_it; @@ -1712,7 +2451,7 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, i = before-i; if (i) { - *status = BEND_SCAN_PARTIAL; + *is_partial = 1; *position -= i; *num_entries -= i; } @@ -1720,8 +2459,7 @@ int rpn_scan (ZServerInfo *zi, Z_AttributesPlusTerm *zapt, logf (LOG_DEBUG, "position = %d, num_entries = %d", *position, *num_entries); - if (zi->errCode) - logf (LOG_DEBUG, "scan error: %d", zi->errCode); - return zi->errCode; + if (zh->errCode) + logf (LOG_DEBUG, "scan error: %d", zh->errCode); }