X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Fzrpn.c;h=7fefb9bab549dd067b09e6ce1111c1ef94872596;hp=812d78fb9e008957659fc21ee6b1a065019a4dd1;hb=2e4e9c6def27f1e1463dcb6f205fab6a98054f38;hpb=69da23537c6bb71ab948e079708bf8ea090de73f diff --git a/index/zrpn.c b/index/zrpn.c index 812d78f..7fefb9b 100644 --- a/index/zrpn.c +++ b/index/zrpn.c @@ -1,362 +1,26 @@ -/* - * Copyright (C) 1995-2000, Index Data - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: zrpn.c,v $ - * Revision 1.104 2000-04-05 09:49:35 adam - * On Unix, zebra/z'mbol uses automake. - * - * Revision 1.103 2000/03/20 19:08:36 adam - * Added remote record import using Z39.50 extended services and Segment - * Requests. - * - * Revision 1.102 2000/03/15 15:00:31 adam - * First work on threaded version. - * - * Revision 1.101 2000/03/02 14:35:03 adam - * Fixed proximity handling. - * - * Revision 1.100 1999/12/28 15:48:12 adam - * Minor Fix. - * - * Revision 1.99 1999/12/23 09:03:32 adam - * Changed behaviour of trunc=105 so that * is regular .* and ! is regular . - * - * Revision 1.98 1999/11/30 13:48:04 adam - * Improved installation. Updated for inclusion of YAZ header files. - * - * Revision 1.97 1999/10/14 14:33:50 adam - * Added truncation 5=106. - * - * Revision 1.96 1999/09/23 10:05:05 adam - * Implemented structure=105 searching. - * - * Revision 1.95 1999/09/07 07:19:21 adam - * Work on character mapping. Implemented replace rules. - * - * Revision 1.94 1999/07/20 13:59:18 adam - * Fixed bug that occurred when phrases had 0 hits. - * - * Revision 1.93 1999/06/17 14:38:40 adam - * Bug fix: Scan SEGV'ed when getting unknown use attribute. - * - * 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 - * Server produces diagnostic "Unsupported Attribute Set" when appropriate. - * - * Revision 1.68 1997/10/27 14:33:06 adam - * Moved towards generic character mapping depending on "structure" - * field in abstract syntax file. Fixed a few memory leaks. Fixed - * bug with negative integers when doing searches with relational - * operators. - * - * Revision 1.67 1997/09/29 09:06:10 adam - * Removed one static var in order to make this module thread safe. - * - * Revision 1.66 1997/09/25 14:58:03 adam - * Windows NT port. - * - * Revision 1.65 1997/09/22 12:39:06 adam - * Added get_pos method for the ranked result sets. - * - * Revision 1.64 1997/09/18 08:59:20 adam - * Extra generic handle for the character mapping routines. - * - * Revision 1.63 1997/09/17 12:19:18 adam - * Zebra version corresponds to YAZ version 1.4. - * Changed Zebra server so that it doesn't depend on global common_resource. - * - * Revision 1.62 1997/09/05 15:30:09 adam - * Changed prototype for chr_map_input - added const. - * Added support for C++, headers uses extern "C" for public definitions. - * - * Revision 1.61 1997/02/10 10:21:14 adam - * Bug fix: in search terms character (^) wasn't observed. - * - * Revision 1.60 1997/01/31 11:10:34 adam - * Bug fix: Leading and trailing white space weren't removed in scan tokens. - * - * Revision 1.59 1997/01/17 11:31:46 adam - * Bug fix: complete phrase search didn't work. - * - * Revision 1.58 1996/12/23 15:30:45 adam - * Work on truncation. - * Bug fix: result sets weren't deleted after server shut down. - * - * Revision 1.57 1996/11/11 13:38:02 adam - * Added proximity support in search. - * - * Revision 1.56 1996/11/08 11:10:32 adam - * Buffers used during file match got bigger. - * Compressed ISAM support everywhere. - * Bug fixes regarding masking characters in queries. - * Redesigned Regexp-2 queries. - * - * Revision 1.55 1996/11/04 14:07:44 adam - * Moved truncation code to trunc.c. - * - * Revision 1.54 1996/10/29 14:09:52 adam - * Use of cisam system - enabled if setting isamc is 1. - * - * Revision 1.53 1996/06/26 09:21:43 adam - * Bug fix: local attribute set wasn't obeyed in scan. - * - * Revision 1.52 1996/06/17 14:26:20 adam - * Function gen_regular_rel changed to handle negative numbers. - * - * Revision 1.51 1996/06/11 10:54:15 quinn - * Relevance work - * - * Revision 1.50 1996/06/07 08:51:53 adam - * Bug fix: Character mapping was broken (introducued by last revision). - * - * 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 - * Implemented detection of database availability. - * - * Revision 1.31 1995/10/17 18:02:10 adam - * New feature: databases. Implemented as prefix to words in dictionary. - * - * Revision 1.30 1995/10/16 09:32:38 adam - * More work on relational op. - * - * Revision 1.29 1995/10/13 16:01:49 adam - * Work on relations. - * - * Revision 1.28 1995/10/13 12:26:43 adam - * Optimization of truncation. - * - * Revision 1.27 1995/10/12 17:07:22 adam - * Truncation works. - * - * Revision 1.26 1995/10/12 12:40:54 adam - * Bug fixes in rpn_prox. - * - * Revision 1.25 1995/10/10 13:59:24 adam - * Function rset_open changed its wflag parameter to general flags. - * - * Revision 1.24 1995/10/09 16:18:37 adam - * Function dict_lookup_grep got extra client data parameter. - * - * Revision 1.23 1995/10/06 16:33:37 adam - * Use attribute mappings. - * - * Revision 1.22 1995/10/06 15:07:39 adam - * Structure 'local-number' handled. - * - * Revision 1.21 1995/10/06 13:52:06 adam - * Bug fixes. Handler may abort further scanning. - * - * Revision 1.20 1995/10/06 11:06:33 adam - * Scan entries include 'occurrences' now. - * - * Revision 1.19 1995/10/06 10:43:56 adam - * Scan added. 'occurrences' in scan entries not set yet. - * - * Revision 1.18 1995/10/04 16:57:20 adam - * Key input and merge sort in one pass. - * - * Revision 1.17 1995/10/04 12:55:17 adam - * Bug fix in ranked search. Use=Any keys inserted. - * - * Revision 1.16 1995/10/02 16:24:40 adam - * Use attribute actually used in search requests. - * - * Revision 1.15 1995/10/02 15:18:52 adam - * New member in recRetrieveCtrl: diagnostic. - * - * Revision 1.14 1995/09/28 12:10:32 adam - * Bug fixes. Field prefix used in queries. - * - * Revision 1.13 1995/09/18 14:17:50 adam - * Minor changes. - * - * Revision 1.12 1995/09/15 14:45:21 adam - * Retrieve control. - * Work on truncation. - * - * Revision 1.11 1995/09/14 11:53:27 adam - * First work on regular expressions/truncations. - * - * Revision 1.10 1995/09/11 15:23:26 adam - * More work on relevance search. - * - * Revision 1.9 1995/09/11 13:09:35 adam - * More work on relevance feedback. - * - * Revision 1.8 1995/09/08 14:52:27 adam - * Minor changes. Dictionary is lower case now. - * - * Revision 1.7 1995/09/07 13:58:36 adam - * New parameter: result-set file descriptor (RSFD) to support multiple - * positions within the same result-set. - * Boolean operators: and, or, not implemented. - * Result-set references. - * - * Revision 1.6 1995/09/06 16:11:18 adam - * Option: only one word key per file. - * - * Revision 1.5 1995/09/06 10:33:04 adam - * More work on present. Some log messages removed. - * - * Revision 1.4 1995/09/05 15:28:40 adam - * More work on search engine. - * - * Revision 1.3 1995/09/04 15:20:22 adam - * Minor changes. - * - * Revision 1.2 1995/09/04 12:33:43 adam - * Various cleanup. YAZ util used instead. - * - * Revision 1.1 1995/09/04 09:10:40 adam - * More work on index add/del/update. - * Merge sort implemented. - * Initial work on z39 server. - * - */ +/* $Id: zrpn.c,v 1.142 2004-08-04 08:35:23 adam Exp $ + Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004 + Index Data Aps + +This file is part of the Zebra server. + +Zebra is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free +Software Foundation; either version 2, or (at your option) any later +version. + +Zebra is distributed in the hope that it will be useful, but WITHOUT ANY +WARRANTY; without even the implied warranty of MERCHANTABILITY or +FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +for more details. + +You should have received a copy of the GNU General Public License +along with Zebra; see the file LICENSE.zebra. If not, write to the +Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA +02111-1307, USA. +*/ + + #include #include #ifdef WIN32 @@ -366,57 +30,67 @@ #endif #include -#include "zserver.h" +#include "index.h" +#include #include #include #include #include +#include +#include struct rpn_char_map_info { ZebraMaps zm; int reg_type; }; +typedef struct { + int type; + int major; + int minor; + Z_AttributesPlusTerm *zapt; +} AttrType; + + static const char **rpn_char_map_handler (void *vp, const char **from, int len) { struct rpn_char_map_info *p = (struct rpn_char_map_info *) vp; - return zebra_maps_input (p->zm, p->reg_type, from, len); + const char **out = zebra_maps_input (p->zm, p->reg_type, from, len); +#if 0 + if (out && *out) + { + const char *outp = *out; + yaz_log (LOG_LOG, "---"); + while (*outp) + { + yaz_log (LOG_LOG, "%02X", *outp); + outp++; + } + } +#endif + return out; } -static void rpn_char_map_prepare (ZebraHandle zh, int reg_type, +static void rpn_char_map_prepare (struct zebra_register *reg, int reg_type, struct rpn_char_map_info *map_info) { - map_info->zm = zh->service->zebra_maps; + map_info->zm = reg->zebra_maps; map_info->reg_type = reg_type; - dict_grep_cmap (zh->service->dict, map_info, rpn_char_map_handler); + dict_grep_cmap (reg->dict, map_info, rpn_char_map_handler); } -typedef struct { - int type; - int major; - int minor; - Z_AttributesPlusTerm *zapt; -} AttrType; - -static int attr_find (AttrType *src, oid_value *attributeSetP) +static int attr_find_ex (AttrType *src, oid_value *attributeSetP, + const char **string_value) { 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) @@ -433,19 +107,34 @@ static int attr_find (AttrType *src, oid_value *attributeSetP) return *element->value.numeric; break; case Z_AttributeValue_complex: - if (src->minor >= element->value.complex->num_list || - element->value.complex->list[src->minor]->which != - Z_StringOrNumeric_numeric) - break; - ++(src->minor); + if (src->minor >= element->value.complex->num_list) + break; if (element->attributeSet && attributeSetP) { oident *attrset; - + attrset = oid_getentbyoid (element->attributeSet); *attributeSetP = attrset->value; } - return *element->value.complex->list[src->minor-1]->u.numeric; + if (element->value.complex->list[src->minor]->which == + Z_StringOrNumeric_numeric) + { + ++(src->minor); + return + *element->value.complex->list[src->minor-1]->u.numeric; + } + else if (element->value.complex->list[src->minor]->which == + Z_StringOrNumeric_string) + { + if (!string_value) + break; + ++(src->minor); + *string_value = + element->value.complex->list[src->minor-1]->u.string; + return -2; + } + else + break; default: assert (0); } @@ -455,6 +144,11 @@ static int attr_find (AttrType *src, oid_value *attributeSetP) return -1; } +static int attr_find (AttrType *src, oid_value *attributeSetP) +{ + return attr_find_ex (src, attributeSetP, 0); +} + static void attr_init (AttrType *src, Z_AttributesPlusTerm *zapt, int type) { @@ -470,27 +164,29 @@ struct grep_info { #ifdef TERM_COUNT int *term_no; #endif - ISAMS_P *isam_p_buf; + ISAMC_P *isam_p_buf; int isam_p_size; int isam_p_indx; ZebraHandle zh; int reg_type; + ZebraSet termset; }; static void term_untrans (ZebraHandle zh, int reg_type, char *dst, const char *src) { + int len = 0; while (*src) { - const char *cp = zebra_maps_output (zh->service->zebra_maps, + const char *cp = zebra_maps_output (zh->reg->zebra_maps, reg_type, &src); - if (!cp) - *dst++ = *src++; + if (!cp && len < IT_MAX_WORD-1) + dst[len++] = *src++; else - while (*cp) - *dst++ = *cp++; + while (*cp && len < IT_MAX_WORD-1) + dst[len++] = *cp++; } - *dst = '\0'; + dst[len] = '\0'; } static void add_isam_p (const char *name, const char *info, @@ -498,12 +194,12 @@ static void add_isam_p (const char *name, const char *info, { if (p->isam_p_indx == p->isam_p_size) { - ISAMS_P *new_isam_p_buf; + ISAMC_P *new_isam_p_buf; #ifdef TERM_COUNT int *new_term_no; #endif p->isam_p_size = 2*p->isam_p_size + 100; - new_isam_p_buf = (ISAMS_P *) xmalloc (sizeof(*new_isam_p_buf) * + new_isam_p_buf = (ISAMC_P *) xmalloc (sizeof(*new_isam_p_buf) * p->isam_p_size); if (p->isam_p_buf) { @@ -528,9 +224,24 @@ static void add_isam_p (const char *name, const char *info, 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); +#if 1 + if (p->termset) + { + const char *db; + int set, use; + char term_tmp[IT_MAX_WORD]; + int su_code = 0; + int len = key_SU_decode (&su_code, name); + + term_untrans (p->zh, p->reg_type, term_tmp, name+len+1); + logf (LOG_LOG, "grep: %d %c %s", su_code, name[len], term_tmp); + zebraExplain_lookup_ord (p->zh->reg->zei, + su_code, &db, &set, &use); + logf (LOG_LOG, "grep: set=%d use=%d db=%s", set, use, db); + + resultSetAddTerm (p->zh, p->termset, name[len], db, + set, use, term_tmp); + } #endif (p->isam_p_indx)++; } @@ -564,6 +275,8 @@ static int term_pre (ZebraMaps zebra_maps, int reg_type, const char **src, return *s0; } +#define REGEX_CHARS " []()|.*+?!" + /* term_100: handle term, where trunc=none (no operators at all) */ static int term_100 (ZebraMaps zebra_maps, int reg_type, const char **src, char *dst, int space_split, @@ -574,6 +287,9 @@ static int term_100 (ZebraMaps zebra_maps, int reg_type, int i = 0; int j = 0; + const char *space_start = 0; + const char *space_end = 0; + if (!term_pre (zebra_maps, reg_type, src, NULL, NULL)) return 0; s0 = *src; @@ -581,11 +297,36 @@ static int term_100 (ZebraMaps zebra_maps, int reg_type, { s1 = s0; map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0)); - if (space_split && **map == *CHR_SPACE) - break; + if (space_split) + { + if (**map == *CHR_SPACE) + break; + } + else /* complete subfield only. */ + { + if (**map == *CHR_SPACE) + { /* save space mapping for later .. */ + space_start = s1; + space_end = s0; + continue; + } + else if (space_start) + { /* reload last space */ + while (space_start < space_end) + { + if (strchr (REGEX_CHARS, *space_start)) + dst[i++] = '\\'; + dst_term[j++] = *space_start; + dst[i++] = *space_start++; + } + /* and reset */ + space_start = space_end = 0; + } + } + /* add non-space char */ while (s1 < s0) { - if (!isalnum (*s1) && *s1 != '-') + if (strchr(REGEX_CHARS, *s1)) dst[i++] = '\\'; dst_term[j++] = *s1; dst[i++] = *s1++; @@ -626,7 +367,7 @@ static int term_101 (ZebraMaps zebra_maps, int reg_type, break; while (s1 < s0) { - if (!isalnum (*s1)) + if (strchr(REGEX_CHARS, *s1)) dst[i++] = '\\'; dst_term[j++] = *s1; dst[i++] = *s1++; @@ -675,7 +416,7 @@ static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src, break; while (s1 < s0) { - if (!isalnum (*s1)) + if (strchr(REGEX_CHARS, *s1)) dst[i++] = '\\'; dst_term[j++] = *s1; dst[i++] = *s1++; @@ -707,18 +448,43 @@ static int term_104 (ZebraMaps zebra_maps, int reg_type, int i = 0; int j = 0; - if (!term_pre (zebra_maps, reg_type, src, "#!", "#!")) + if (!term_pre (zebra_maps, reg_type, src, "?*#", "?*#")) return 0; s0 = *src; while (*s0) { - if (*s0 == '#') + if (*s0 == '?') + { + dst_term[j++] = *s0++; + if (*s0 >= '0' && *s0 <= '9') + { + int limit = 0; + while (*s0 >= '0' && *s0 <= '9') + { + limit = limit * 10 + (*s0 - '0'); + dst_term[j++] = *s0++; + } + if (limit > 20) + limit = 20; + while (--limit >= 0) + { + dst[i++] = '.'; + dst[i++] = '?'; + } + } + else + { + dst[i++] = '.'; + dst[i++] = '*'; + } + } + else if (*s0 == '*') { dst[i++] = '.'; dst[i++] = '*'; dst_term[j++] = *s0++; } - else if (*s0 == '!') + else if (*s0 == '#') { dst[i++] = '.'; dst_term[j++] = *s0++; @@ -730,7 +496,7 @@ static int term_104 (ZebraMaps zebra_maps, int reg_type, break; while (s1 < s0) { - if (!isalnum (*s1)) + if (strchr(REGEX_CHARS, *s1)) dst[i++] = '\\'; dst_term[j++] = *s1; dst[i++] = *s1++; @@ -776,7 +542,7 @@ static int term_105 (ZebraMaps zebra_maps, int reg_type, break; while (s1 < s0) { - if (!isalnum (*s1)) + if (strchr(REGEX_CHARS, *s1)) dst[i++] = '\\'; dst_term[j++] = *s1; dst[i++] = *s1++; @@ -934,7 +700,7 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, int relation_value; int i; char *term_tmp = term_dict + strlen(term_dict); - char term_component[256]; + char term_component[2*IT_MAX_WORD+20]; attr_init (&relation, zapt, 2); relation_value = attr_find (&relation, NULL); @@ -943,7 +709,7 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, switch (relation_value) { case 1: - if (!term_100 (zh->service->zebra_maps, reg_type, + if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, term_component, space_split, term_dst)) return 0; @@ -968,12 +734,15 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, *term_tmp++ = ']'; *term_tmp++ = '.'; *term_tmp++ = '*'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; } *term_tmp++ = ')'; *term_tmp = '\0'; break; case 2: - if (!term_100 (zh->service->zebra_maps, reg_type, + if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, term_component, space_split, term_dst)) return 0; @@ -997,6 +766,9 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, *term_tmp++ = '*'; *term_tmp++ = '|'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; } for (i = 0; term_component[i]; ) string_rel_add_char (&term_tmp, term_component, &i); @@ -1004,7 +776,7 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, *term_tmp = '\0'; break; case 5: - if (!term_100 (zh->service->zebra_maps, reg_type, + if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, term_component, space_split, term_dst)) return 0; logf (LOG_DEBUG, "Relation >"); @@ -1027,6 +799,9 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, *term_tmp++ = '*'; *term_tmp++ = '|'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; } for (i = 0; term_component[i];) string_rel_add_char (&term_tmp, term_component, &i); @@ -1036,7 +811,7 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, *term_tmp = '\0'; break; case 4: - if (!term_100 (zh->service->zebra_maps, reg_type, term_sub, + if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, term_component, space_split, term_dst)) return 0; logf (LOG_DEBUG, "Relation >="); @@ -1066,6 +841,9 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, *term_tmp++ = ']'; *term_tmp++ = '.'; *term_tmp++ = '*'; + + if ((term_tmp - term_dict) > IT_MAX_WORD) + break; } *term_tmp++ = ')'; *term_tmp = '\0'; @@ -1073,7 +851,7 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, case 3: default: logf (LOG_DEBUG, "Relation ="); - if (!term_100 (zh->service->zebra_maps, reg_type, term_sub, + if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, term_component, space_split, term_dst)) return 0; strcat (term_tmp, "("); @@ -1084,12 +862,44 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, } 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) + 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, int xpath_use); + +static RSET term_trunc (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, + const char *rank_type, int xpath_use) +{ + int r; + grep_info->isam_p_indx = 0; + r = string_term (zh, zapt, term_sub, attributeSet, stream, grep_info, + reg_type, complete_flag, num_bases, basenames, + term_dst, xpath_use); + if (r < 1) + return 0; + logf (LOG_DEBUG, "term: %s", term_dst); + return rset_trunc (zh, grep_info->isam_p_buf, + grep_info->isam_p_indx, term_dst, + strlen(term_dst), rank_type, 1 /* preserve pos */, + zapt->term->which); +} + + +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, int xpath_use) { char term_dict[2*IT_MAX_WORD+4000]; int j, r, base_no; @@ -1097,93 +907,130 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, int truncation_value; AttrType use; int use_value; + const char *use_string = 0; oid_value curAttributeSet = attributeSet; const char *termp; struct rpn_char_map_info rcmi; int space_split = complete_flag ? 0 : 1; - rpn_char_map_prepare (zh, reg_type, &rcmi); + int bases_ok = 0; /* no of databases with OK attribute */ + int errCode = 0; /* err code (if any is not OK) */ + char *errString = 0; /* addinfo */ + + rpn_char_map_prepare (zh->reg, reg_type, &rcmi); attr_init (&use, zapt, 1); - use_value = attr_find (&use, &curAttributeSet); + use_value = attr_find_ex (&use, &curAttributeSet, &use_string); 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); - if (use_value == -1) + if (use_value == -1) /* no attribute - assumy "any" */ use_value = 1016; - for (base_no = 0; base_no < num_bases; base_no++) { attent attp; + data1_local_attribute id_xpath_attr; 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) - { - char val_str[32]; - sprintf (val_str, "%d", use_value); - zh->errCode = 114; - zh->errString = nmem_strdup (stream, val_str); - } - else - { - 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 (zebraExplain_curDatabase (zh->service->zei, basenames[base_no])) + if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) { zh->errCode = 109; /* Database unavailable */ zh->errString = basenames[base_no]; return -1; } + if (xpath_use > 0 && use_value == -2) + { + use_value = xpath_use; + attp.local_attributes = &id_xpath_attr; + attp.attset_ordinal = VAL_IDXPATH; + id_xpath_attr.next = 0; + id_xpath_attr.local = use_value; + } + else if (curAttributeSet == VAL_IDXPATH) + { + attp.local_attributes = &id_xpath_attr; + attp.attset_ordinal = VAL_IDXPATH; + id_xpath_attr.next = 0; + id_xpath_attr.local = use_value; + } + else + { + if ((r=att_getentbyatt (zh, &attp, curAttributeSet, use_value, + use_string))) + { + logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d", + curAttributeSet, use_value, r); + if (r == -1) + { + /* set was found, but value wasn't defined */ + errCode = 114; + if (use_string) + errString = nmem_strdup(stream, use_string); + else + { + char val_str[32]; + sprintf (val_str, "%d", use_value); + errString = nmem_strdup (stream, val_str); + } + } + else + { + int oid[OID_SIZE]; + struct oident oident; + + oident.proto = PROTO_Z3950; + oident.oclass = CLASS_ATTSET; + oident.value = curAttributeSet; + oid_ent_to_oid (&oident, oid); + + errCode = 121; + errString = nmem_strdup (stream, oident.desc); + } + continue; + } + } 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->service->zei, attp.attset_ordinal, - local_attr->local); + char ord_buf[32]; + int i, ord_len; + + ord = zebraExplain_lookupSU (zh->reg->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; - zh->errString = nmem_strdup (stream, val_str); - return -1; +#if 1 + bases_ok++; +#else + char val_str[32]; + sprintf (val_str, "%d", use_value); + errCode = 114; + errString = nmem_strdup (stream, val_str); +#endif + continue; } - term_dict[prefix_len++] = ')'; + bases_ok++; /* this has OK attributes */ + + 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]); @@ -1197,60 +1044,60 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, 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->service->dict, term_dict, 0, + logf (LOG_LOG, "dict_lookup_grep: %s", term_dict+prefix_len); + r = dict_lookup_grep (zh->reg->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_WARN, "dict_lookup_grep fail %d", r); break; case 1: /* right truncation */ term_dict[j++] = '('; - if (!term_100 (zh->service->zebra_maps, reg_type, + if (!term_100 (zh->reg->zebra_maps, reg_type, &termp, term_dict + j, space_split, term_dst)) return 0; strcat (term_dict, ".*)"); - dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, + dict_lookup_grep (zh->reg->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->service->zebra_maps, reg_type, + if (!term_100 (zh->reg->zebra_maps, reg_type, &termp, term_dict + j, space_split, term_dst)) return 0; strcat (term_dict, ")"); - dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, + dict_lookup_grep (zh->reg->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->service->zebra_maps, reg_type, + if (!term_100 (zh->reg->zebra_maps, reg_type, &termp, term_dict + j, space_split, term_dst)) return 0; strcat (term_dict, ".*)"); - dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, + dict_lookup_grep (zh->reg->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->service->zebra_maps, reg_type, + if (!term_101 (zh->reg->zebra_maps, reg_type, &termp, term_dict + j, space_split, term_dst)) return 0; strcat (term_dict, ")"); - r = dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, + r = dict_lookup_grep (zh->reg->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->service->zebra_maps, reg_type, + if (!term_102 (zh->reg->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->service->dict, term_dict, 0, grep_info, + r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, &max_pos, 0, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep err, trunc=regular: %d", @@ -1259,12 +1106,12 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, case 103: /* Regexp-2 */ r = 1; term_dict[j++] = '('; - if (!term_103 (zh->service->zebra_maps, reg_type, + if (!term_103 (zh->reg->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->service->dict, term_dict, r, grep_info, + r = dict_lookup_grep (zh->reg->dict, term_dict, r, grep_info, &max_pos, 2, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep err, trunc=eregular: %d", @@ -1272,309 +1119,140 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, break; case 104: /* process # and ! in term */ term_dict[j++] = '('; - if (!term_104 (zh->service->zebra_maps, reg_type, + if (!term_104 (zh->reg->zebra_maps, reg_type, &termp, term_dict + j, space_split, term_dst)) return 0; strcat (term_dict, ")"); - r = dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, + r = dict_lookup_grep (zh->reg->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 105: /* process * and ! in term */ term_dict[j++] = '('; - if (!term_105 (zh->service->zebra_maps, reg_type, + if (!term_105 (zh->reg->zebra_maps, reg_type, &termp, term_dict + j, space_split, term_dst, 1)) return 0; strcat (term_dict, ")"); - r = dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, + r = dict_lookup_grep (zh->reg->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 106: /* process * and ! in term */ term_dict[j++] = '('; - if (!term_105 (zh->service->zebra_maps, reg_type, + if (!term_105 (zh->reg->zebra_maps, reg_type, &termp, term_dict + j, space_split, term_dst, 0)) return 0; strcat (term_dict, ")"); - r = dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, + r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, &max_pos, 0, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep err, trunc=*/!: %d", r); break; } } + if (!bases_ok) + { + zh->errCode = errCode; + zh->errString = errString; + return -1; + } *term_sub = termp; logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); return 1; } -static void trans_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - char *termz) + +/* convert APT search term to UTF8 */ +static int zapt_term_to_utf8 (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + char *termz) { size_t sizez; Z_Term *term = zapt->term; - sizez = term->u.general->len; - 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 (ZebraHandle zh, Z_AttributesPlusTerm *zapt, - char *termz, int reg_type) -{ - Z_Term *term = zapt->term; - const 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; - const char *space_map = NULL; - int len; - - while ((len = (cp_end - cp)) > 0) + switch (term->which) { - map = zebra_maps_input (zh->service->zebra_maps, reg_type, &cp, len); - if (**map == *CHR_SPACE) - space_map = *map; + case Z_Term_general: + if (zh->iconv_to_utf8 != 0) + { + char *inbuf = term->u.general->buf; + size_t inleft = term->u.general->len; + char *outbuf = termz; + size_t outleft = IT_MAX_WORD-1; + size_t ret; + + ret = yaz_iconv(zh->iconv_to_utf8, &inbuf, &inleft, + &outbuf, &outleft); + if (ret == (size_t)(-1)) + { + ret = yaz_iconv(zh->iconv_to_utf8, 0, 0, 0, 0); + zh->errCode = 125; + return -1; + } + *outbuf = 0; + } else { - if (i && space_map) - for (src = space_map; *src; src++) - termz[i++] = *src; - space_map = NULL; - for (src = *map; *src; src++) - termz[i++] = *src; + sizez = term->u.general->len; + if (sizez > IT_MAX_WORD-1) + sizez = IT_MAX_WORD-1; + memcpy (termz, term->u.general->buf, sizez); + termz[sizez] = '\0'; } + break; + case Z_Term_characterString: + sizez = strlen(term->u.characterString); + if (sizez > IT_MAX_WORD-1) + sizez = IT_MAX_WORD-1; + memcpy (termz, term->u.characterString, sizez); + termz[sizez] = '\0'; + break; + default: + zh->errCode = 124; + return -1; } - termz[i] = '\0'; + return 0; } -static RSET rpn_prox (ZebraHandle zh, RSET *rset, int rset_no, - int ordered, int exclusion, int relation, int distance) +/* convert APT SCAN term to internal cmap */ +static int trans_scan_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, + char *termz, int reg_type) { - int i; - RSFD *rsfd; - int *more; - struct it_key **buf; - RSET result; - char prox_term[1024]; - int length_prox_term = 0; - int min_nn = 10000000; - int term_index; - const char *flags = NULL; - - rsfd = (RSFD *) xmalloc (sizeof(*rsfd)*rset_no); - more = (int *) xmalloc (sizeof(*more)*rset_no); - buf = (struct it_key **) xmalloc (sizeof(*buf)*rset_no); - - *prox_term = '\0'; - for (i = 0; ino_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 (lterm + length_prox_term < sizeof(prox_term)-1) - { - 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; - } - } - for (i = 0; inn = 0; - result = rset_create (rset_kind_null, &parms); - } - else if (ordered && relation == 3 && exclusion == 0 && distance == 1) - { - /* special proximity case = phrase search ... */ - rset_temp_parms parms; - RSFD rsfd_result; - - parms.rset_term = rset_term_create (prox_term, length_prox_term, - flags); - parms.rset_term->nn = min_nn; - parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zh->service->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); - } - } - rset_close (result, rsfd_result); - } - else if (rset_no == 2) - { - /* generic proximity case (two input sets only) ... */ - rset_temp_parms parms; - RSFD rsfd_result; - - logf (LOG_LOG, "generic prox, dist = %d, relation = %d, ordered =%d, exclusion=%d", - distance, relation, ordered, exclusion); - parms.rset_term = rset_term_create (prox_term, length_prox_term, - flags); - parms.rset_term->nn = min_nn; - parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zh->service->res, "setTmpDir"); - result = rset_create (rset_kind_temp, &parms); - rsfd_result = rset_open (result, RSETF_WRITE); + char termz0[IT_MAX_WORD]; - while (more[0] && more[1]) - { - int cmp = key_compare_it (buf[0], buf[1]); - if (cmp < -1) - more[0] = rset_read (rset[0], rsfd[0], buf[0], &term_index); - else if (cmp > 1) - more[1] = rset_read (rset[1], rsfd[1], buf[1], &term_index); - else - { - int sysno = buf[0]->sysno; - int seqno[500]; - int n = 0; - - seqno[n++] = buf[0]->seqno; - while ((more[0] = rset_read (rset[0], rsfd[0], buf[0], - &term_index)) && - sysno == buf[0]->sysno) - if (n < 500) - seqno[n++] = buf[0]->seqno; - do - { - for (i = 0; iseqno - seqno[i]; - int excl = exclusion; - if (!ordered && diff < 0) - diff = -diff; - switch (relation) - { - case 1: /* < */ - if (diff < distance && diff >= 0) - excl = !excl; - break; - case 2: /* <= */ - if (diff <= distance && diff >= 0) - excl = !excl; - break; - case 3: /* == */ - if (diff == distance && diff >= 0) - excl = !excl; - break; - case 4: /* >= */ - if (diff >= distance && diff >= 0) - excl = !excl; - break; - case 5: /* > */ - if (diff > distance && diff >= 0) - excl = !excl; - break; - case 6: /* != */ - if (diff != distance && diff >= 0) - excl = !excl; - break; - } - if (excl) - { - rset_write (result, rsfd_result, buf[1]); - break; - } - } - } while ((more[1] = rset_read (rset[1], rsfd[1], buf[1], - &term_index)) && - sysno == buf[1]->sysno); - } - } - rset_close (result, rsfd_result); - } + if (zapt_term_to_utf8(zh, zapt, termz0)) + return -1; /* error */ else { - rset_null_parms parms; - - parms.rset_term = rset_term_create (prox_term, length_prox_term, - flags); - parms.rset_term->nn = 0; - result = rset_create (rset_kind_null, &parms); - } - for (i = 0; i 0) + { + map = zebra_maps_input (zh->reg->zebra_maps, reg_type, &cp, len); + if (**map == *CHR_SPACE) + space_map = *map; + else + { + if (i && space_map) + for (src = space_map; *src; src++) + termz[i++] = *src; + space_map = NULL; + for (src = *map; *src; src++) + termz[i++] = *src; + } + } + termz[i] = '\0'; } - xfree (buf); - xfree (more); - xfree (rsfd); - return result; + return 0; } - char *normalize_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, const char *termz, NMEM stream, unsigned reg_id) { @@ -1606,7 +1284,7 @@ char *normalize_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, break; } if (ex_list) - wrbuf = zebra_replace(zh->service->zebra_maps, reg_id, ex_list, + wrbuf = zebra_replace(zh->reg->zebra_maps, reg_id, ex_list, termz, strlen(termz)); if (!wrbuf) return nmem_strdup(stream, termz); @@ -1619,63 +1297,124 @@ char *normalize_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt, } } +static void grep_info_delete (struct grep_info *grep_info) +{ +#ifdef TERM_COUNT + xfree(grep_info->term_no); +#endif + xfree (grep_info->isam_p_buf); +} + +static int grep_info_prepare (ZebraHandle zh, + Z_AttributesPlusTerm *zapt, + struct grep_info *grep_info, + int reg_type, + NMEM stream) +{ + AttrType termset; + int termset_value_numeric; + const char *termset_value_string; + +#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; + grep_info->termset = 0; + + if (!zapt) + return 0; + attr_init (&termset, zapt, 8); + termset_value_numeric = + attr_find_ex (&termset, NULL, &termset_value_string); + if (termset_value_numeric != -1) + { + char resname[32]; + const char *termset_name = 0; + if (termset_value_numeric != -2) + { + + sprintf (resname, "%d", termset_value_numeric); + termset_name = resname; + } + else + termset_name = termset_value_string; + logf (LOG_LOG, "creating termset set %s", termset_name); + grep_info->termset = resultSetAdd (zh, termset_name, 1); + if (!grep_info->termset) + { + zh->errCode = 128; + zh->errString = nmem_strdup (stream, termset_name); + return -1; + } + } + return 0; +} + + static RSET rpn_search_APT_phrase (ZebraHandle zh, Z_AttributesPlusTerm *zapt, const char *termz_org, oid_value attributeSet, NMEM stream, int reg_type, int complete_flag, - const char *rank_type, + const char *rank_type, int xpath_use, int num_bases, char **basenames) { char term_dst[IT_MAX_WORD+1]; RSET rset[60], result; - int i, r, rset_no = 0; + int rset_no = 0; struct grep_info grep_info; char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type); const char *termp = 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; - + *term_dst = 0; + if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream)) + return 0; while (1) { logf (LOG_DEBUG, "APT_phrase termp=%s", termp); - grep_info.isam_p_indx = 0; - r = string_term (zh, zapt, &termp, attributeSet, stream, &grep_info, - reg_type, complete_flag, num_bases, basenames, - term_dst); - if (r < 1) + rset[rset_no] = term_trunc (zh, zapt, &termp, attributeSet, + stream, &grep_info, + reg_type, complete_flag, + num_bases, basenames, + term_dst, rank_type, + xpath_use); + if (!rset[rset_no]) 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); + grep_info_delete (&grep_info); if (rset_no == 0) { rset_null_parms parms; - parms.rset_term = rset_term_create (term_dst, -1, rank_type); + parms.rset_term = rset_term_create (termz, -1, rank_type, + zapt->term->which); return rset_create (rset_kind_null, &parms); } else if (rset_no == 1) return (rset[0]); - result = rpn_prox (zh, rset, rset_no, 1, 0, 3, 1); - 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); + grep_info_delete (&grep_info); if (rset_no == 0) { rset_null_parms parms; - parms.rset_term = rset_term_create (term_dst, -1, rank_type); + parms.rset_term = rset_term_create (termz, -1, rank_type, + zapt->term->which); return rset_create (rset_kind_null, &parms); } result = rset[0]; @@ -1739,6 +1469,7 @@ static RSET rpn_search_APT_or_list (ZebraHandle zh, bool_parms.rset_r = rset[i]; bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare_it; + bool_parms.log_item = key_logdump_txt; result = rset_create (rset_kind_or, &bool_parms); } return result; @@ -1750,50 +1481,41 @@ static RSET rpn_search_APT_and_list (ZebraHandle zh, oid_value attributeSet, NMEM stream, int reg_type, int complete_flag, - const char *rank_type, + const char *rank_type, + int xpath_use, int num_bases, char **basenames) { char term_dst[IT_MAX_WORD+1]; RSET rset[60], result; - int i, r, rset_no = 0; + int i, rset_no = 0; struct grep_info grep_info; char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type); const char *termp = 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; - + if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream)) + return 0; while (1) { logf (LOG_DEBUG, "APT_and_list termp=%s", termp); - grep_info.isam_p_indx = 0; - r = string_term (zh, zapt, &termp, attributeSet, stream, &grep_info, - reg_type, complete_flag, num_bases, basenames, - term_dst); - if (r < 1) + rset[rset_no] = term_trunc (zh, zapt, &termp, attributeSet, + stream, &grep_info, + reg_type, complete_flag, + num_bases, basenames, + term_dst, rank_type, + xpath_use); + if (!rset[rset_no]) 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); + grep_info_delete (&grep_info); if (rset_no == 0) { rset_null_parms parms; - parms.rset_term = rset_term_create (term_dst, -1, rank_type); + parms.rset_term = rset_term_create (termz, -1, rank_type, + zapt->term->which); return rset_create (rset_kind_null, &parms); } result = rset[0]; @@ -1805,6 +1527,7 @@ static RSET rpn_search_APT_and_list (ZebraHandle zh, bool_parms.rset_r = rset[i]; bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare_it; + bool_parms.log_item = key_logdump_txt; result = rset_create (rset_kind_and, &bool_parms); } return result; @@ -1830,7 +1553,7 @@ static int numeric_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, logf (LOG_DEBUG, "numeric relation value=%d", relation_value); - if (!term_100 (zh->service->zebra_maps, reg_type, term_sub, term_tmp, 1, + if (!term_100 (zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1, term_dst)) return 0; term_value = atoi (term_tmp); @@ -1858,7 +1581,7 @@ static int numeric_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt, sprintf (term_tmp, "(0*%d)", term_value); } logf (LOG_DEBUG, "dict_lookup_grep: %s", term_tmp); - r = dict_lookup_grep (zh->service->dict, term_dict, 0, grep_info, max_pos, + r = dict_lookup_grep (zh->reg->dict, term_dict, 0, grep_info, max_pos, 0, grep_handle); if (r) logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r); @@ -1871,20 +1594,24 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, oid_value attributeSet, struct grep_info *grep_info, int reg_type, int complete_flag, int num_bases, char **basenames, - char *term_dst) + char *term_dst, int xpath_use, NMEM stream) { char term_dict[2*IT_MAX_WORD+2]; int r, base_no; AttrType use; int use_value; + const char *use_string = 0; oid_value curAttributeSet = attributeSet; const char *termp; struct rpn_char_map_info rcmi; - rpn_char_map_prepare (zh, reg_type, &rcmi); + int bases_ok = 0; /* no of databases with OK attribute */ + int errCode = 0; /* err code (if any is not OK) */ + char *errString = 0; /* addinfo */ + + rpn_char_map_prepare (zh->reg, reg_type, &rcmi); attr_init (&use, zapt, 1); - use_value = attr_find (&use, &curAttributeSet); - logf (LOG_DEBUG, "numeric_term, use value %d", use_value); + use_value = attr_find_ex (&use, &curAttributeSet, &use_string); if (use_value == -1) use_value = 1016; @@ -1892,21 +1619,46 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, for (base_no = 0; base_no < num_bases; base_no++) { attent attp; + data1_local_attribute id_xpath_attr; data1_local_attribute *local_attr; int max_pos, prefix_len = 0; termp = *term_sub; - if ((r=att_getentbyatt (zh, &attp, curAttributeSet, use_value))) + if (use_value == -2) /* string attribute (assume IDXPATH/any) */ { - 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; + use_value = xpath_use; + attp.local_attributes = &id_xpath_attr; + attp.attset_ordinal = VAL_IDXPATH; + id_xpath_attr.next = 0; + id_xpath_attr.local = use_value; + } + else if (curAttributeSet == VAL_IDXPATH) + { + attp.local_attributes = &id_xpath_attr; + attp.attset_ordinal = VAL_IDXPATH; + id_xpath_attr.next = 0; + id_xpath_attr.local = use_value; } - if (zebraExplain_curDatabase (zh->service->zei, basenames[base_no])) + else + { + if ((r=att_getentbyatt (zh, &attp, curAttributeSet, use_value, + use_string))) + { + logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d", + curAttributeSet, use_value, r); + if (r == -1) + { + char val_str[32]; + sprintf (val_str, "%d", use_value); + errString = nmem_strdup (stream, val_str); + errCode = 114; + } + else + errCode = 121; + continue; + } + } + if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) { zh->errCode = 109; /* Database unavailable */ zh->errString = basenames[base_no]; @@ -1919,7 +1671,7 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, char ord_buf[32]; int i, ord_len; - ord = zebraExplain_lookupSU (zh->service->zei, attp.attset_ordinal, + ord = zebraExplain_lookupSU (zh->reg->zei, attp.attset_ordinal, local_attr->local); if (ord < 0) continue; @@ -1928,7 +1680,7 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, else term_dict[prefix_len++] = '('; - ord_len = key_SU_code (ord, ord_buf); + ord_len = key_SU_encode (ord, ord_buf); for (i = 0; ierrCode = 114; - return -1; + char val_str[32]; + sprintf (val_str, "%d", use_value); + errCode = 114; + errString = nmem_strdup (stream, val_str); + continue; } + bases_ok++; term_dict[prefix_len++] = ')'; term_dict[prefix_len++] = 1; term_dict[prefix_len++] = reg_type; @@ -1950,6 +1706,12 @@ static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt, term_dst)) return 0; } + if (!bases_ok) + { + zh->errCode = errCode; + zh->errString = errString; + return -1; + } *term_sub = termp; logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx); return 1; @@ -1961,7 +1723,7 @@ static RSET rpn_search_APT_numeric (ZebraHandle zh, oid_value attributeSet, NMEM stream, int reg_type, int complete_flag, - const char *rank_type, + const char *rank_type, int xpath_use, int num_bases, char **basenames) { char term_dst[IT_MAX_WORD+1]; @@ -1970,40 +1732,35 @@ static RSET rpn_search_APT_numeric (ZebraHandle zh, 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; - + if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream)) + return 0; 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); + term_dst, xpath_use, + stream); 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); + strlen(term_dst), rank_type, + 0 /* preserve position */, + zapt->term->which); 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); + grep_info_delete (&grep_info); if (rset_no == 0) { rset_null_parms parms; - parms.rset_term = rset_term_create (term_dst, -1, rank_type); + parms.rset_term = rset_term_create (term_dst, -1, rank_type, + zapt->term->which); return rset_create (rset_kind_null, &parms); } result = rset[0]; @@ -2015,6 +1772,7 @@ static RSET rpn_search_APT_numeric (ZebraHandle zh, bool_parms.rset_r = rset[i]; bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare_it; + bool_parms.log_item = key_logdump_txt; result = rset_create (rset_kind_and, &bool_parms); } return result; @@ -2029,18 +1787,30 @@ static RSET rpn_search_APT_local (ZebraHandle zh, Z_AttributesPlusTerm *zapt, RSET result; RSFD rsfd; struct it_key key; + int sys; rset_temp_parms parms; - parms.rset_term = rset_term_create (termz, -1, rank_type); + parms.rset_term = rset_term_create (termz, -1, rank_type, + zapt->term->which); + parms.cmp = key_compare_it; parms.key_size = sizeof (struct it_key); - parms.temp_path = res_get (zh->service->res, "setTmpDir"); + parms.temp_path = res_get (zh->res, "setTmpDir"); result = rset_create (rset_kind_temp, &parms); rsfd = rset_open (result, RSETF_WRITE); - key.sysno = atoi (termz); + sys = atoi(termz); + if (sys <= 0) + sys = 1; +#if IT_KEY_NEW + key.mem[0] = sys; + key.mem[1] = 1; + key.len = 2; +#else + key.sysno = sys; key.seqno = 1; if (key.sysno <= 0) key.sysno = 1; +#endif rset_write (result, rsfd, &key); rset_close (result, rsfd); return result; @@ -2062,6 +1832,7 @@ static RSET rpn_sort_spec (ZebraHandle zh, Z_AttributesPlusTerm *zapt, Z_AttributeElement *ae; int oid[OID_SIZE]; oident oe; + char termz[20]; attr_init (&sort_relation_type, zapt, 7); sort_relation_value = attr_find (&sort_relation_type, &attributeSet); @@ -2085,6 +1856,7 @@ static RSET rpn_sort_spec (ZebraHandle zh, Z_AttributesPlusTerm *zapt, zapt->term->u.general->len); if (i >= sort_sequence->num_specs) i = 0; + sprintf (termz, "%d", i); oe.proto = PROTO_Z3950; oe.oclass = CLASS_ATTSET; @@ -2122,30 +1894,256 @@ static RSET rpn_sort_spec (ZebraHandle zh, Z_AttributesPlusTerm *zapt, sks->sortRelation = (int *) nmem_malloc (stream, sizeof(*sks->sortRelation)); if (sort_relation_value == 1) - *sks->sortRelation = Z_SortRelation_ascending; + *sks->sortRelation = Z_SortKeySpec_ascending; else if (sort_relation_value == 2) - *sks->sortRelation = Z_SortRelation_descending; + *sks->sortRelation = Z_SortKeySpec_descending; else - *sks->sortRelation = Z_SortRelation_ascending; + *sks->sortRelation = Z_SortKeySpec_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); + parms.rset_term = rset_term_create (termz, -1, rank_type, + zapt->term->which); return rset_create (rset_kind_null, &parms); } +static int parse_xpath(ZebraHandle zh, Z_AttributesPlusTerm *zapt, + oid_value attributeSet, + struct xpath_location_step *xpath, int max, NMEM mem) +{ + oid_value curAttributeSet = attributeSet; + AttrType use; + const char *use_string = 0; + + attr_init (&use, zapt, 1); + attr_find_ex (&use, &curAttributeSet, &use_string); + + if (!use_string || *use_string != '/') + return -1; + + return zebra_parse_xpath_str(use_string, xpath, max, mem); +} + + + +static RSET xpath_trunc(ZebraHandle zh, NMEM stream, + int reg_type, const char *term, int use, + oid_value curAttributeSet) +{ + RSET rset; + struct grep_info grep_info; + char term_dict[2048]; + char ord_buf[32]; + int prefix_len = 0; + int ord = zebraExplain_lookupSU (zh->reg->zei, curAttributeSet, use); + int ord_len, i, r, max_pos; + int term_type = Z_Term_characterString; + const char *flags = "void"; + + if (grep_info_prepare (zh, 0 /* zapt */, &grep_info, '0', stream)) + { + rset_null_parms parms; + + parms.rset_term = rset_term_create (term, strlen(term), + flags, term_type); + parms.rset_term->nn = 0; + return rset_create (rset_kind_null, &parms); + } + + if (ord < 0) + { + rset_null_parms parms; + + parms.rset_term = rset_term_create (term, strlen(term), + flags, term_type); + parms.rset_term->nn = 0; + return rset_create (rset_kind_null, &parms); + } + if (prefix_len) + term_dict[prefix_len++] = '|'; + else + term_dict[prefix_len++] = '('; + + ord_len = key_SU_encode (ord, ord_buf); + for (i = 0; ireg->dict, term_dict, 0, + &grep_info, &max_pos, 0, grep_handle); + yaz_log (LOG_LOG, "%s %d positions", term, + grep_info.isam_p_indx); + rset = rset_trunc (zh, grep_info.isam_p_buf, + grep_info.isam_p_indx, term, strlen(term), + flags, 1, term_type); + grep_info_delete (&grep_info); + return rset; +} + +static RSET rpn_search_xpath (ZebraHandle zh, + oid_value attributeSet, + int num_bases, char **basenames, + NMEM stream, const char *rank_type, RSET rset, + int xpath_len, struct xpath_location_step *xpath) +{ + oid_value curAttributeSet = attributeSet; + int base_no; + int i; + + if (xpath_len < 0) + return rset; + + yaz_log (LOG_LOG, "len=%d", xpath_len); + for (i = 0; i a/.* + //a/b -> b/a/.* + /a -> a/ + /a/b -> b/a/ + + / -> none + + a[@attr=value]/b[@other=othervalue] + + /e/@a val range(e/,range(@a,freetext(w,1015,val),@a),e/) + /a/b val range(b/a/,freetext(w,1016,val),b/a/) + /a/b/@c val range(b/a/,range(@c,freetext(w,1016,val),@c),b/a/) + /a/b[@c=y] val range(b/a/,freetext(w,1016,val),b/a/,@c=y) + /a[@c=y]/b val range(a/,range(b/a/,freetext(w,1016,val),b/a/),a/,@c=y) + /a[@c=x]/b[@c=y] range(a/,range(b/a/,freetext(w,1016,val),b/a/,@c=y),a/,@c=x) + + */ + + dict_grep_cmap (zh->reg->dict, 0, 0); + + for (base_no = 0; base_no < num_bases; base_no++) + { + int level = xpath_len; + int first_path = 1; + + if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) + { + zh->errCode = 109; /* Database unavailable */ + zh->errString = basenames[base_no]; + return rset; + } + while (--level >= 0) + { + char xpath_rev[128]; + int i, len; + rset_between_parms parms; + RSET rset_start_tag = 0, rset_end_tag = 0, rset_attr = 0; + + *xpath_rev = 0; + len = 0; + for (i = level; i >= 1; --i) + { + const char *cp = xpath[i].part; + if (*cp) + { + for (;*cp; cp++) + if (*cp == '*') + { + memcpy (xpath_rev + len, "[^/]*", 5); + len += 5; + } + else if (*cp == ' ') + { + + xpath_rev[len++] = 1; + xpath_rev[len++] = ' '; + } + + else + xpath_rev[len++] = *cp; + xpath_rev[len++] = '/'; + } + else if (i == 1) /* // case */ + { + xpath_rev[len++] = '.'; + xpath_rev[len++] = '*'; + } + } + xpath_rev[len] = 0; + + if (xpath[level].predicate && + xpath[level].predicate->which == XPATH_PREDICATE_RELATION && + xpath[level].predicate->u.relation.name[0]) + { + WRBUF wbuf = wrbuf_alloc(); + wrbuf_puts(wbuf, xpath[level].predicate->u.relation.name+1); + if (xpath[level].predicate->u.relation.value) + { + const char *cp = xpath[level].predicate->u.relation.value; + wrbuf_putc(wbuf, '='); + + while (*cp) + { + if (strchr(REGEX_CHARS, *cp)) + wrbuf_putc(wbuf, '\\'); + wrbuf_putc(wbuf, *cp); + cp++; + } + } + wrbuf_puts(wbuf, ""); + rset_attr = xpath_trunc ( + zh, stream, '0', wrbuf_buf(wbuf), 3, curAttributeSet); + wrbuf_free(wbuf, 1); + } + else + { + if (!first_path) + continue; + } + yaz_log (LOG_LOG, "xpath_rev (%d) = %s", level, xpath_rev); + if (strlen(xpath_rev)) + { + rset_start_tag = xpath_trunc(zh, stream, + '0', xpath_rev, 1, curAttributeSet); + + rset_end_tag = xpath_trunc(zh, stream, + '0', xpath_rev, 2, curAttributeSet); + + parms.key_size = sizeof(struct it_key); + parms.cmp = key_compare_it; + parms.rset_l = rset_start_tag; + parms.rset_m = rset; + parms.rset_r = rset_end_tag; + parms.rset_attr = rset_attr; + parms.printer = key_print_it; + rset = rset_create (rset_kind_between, &parms); + } + first_path = 0; + } + } + + return rset; +} + + + static RSET rpn_search_APT (ZebraHandle zh, Z_AttributesPlusTerm *zapt, oid_value attributeSet, NMEM stream, Z_SortKeySpecList *sort_sequence, @@ -2153,61 +2151,78 @@ static RSET rpn_search_APT (ZebraHandle zh, Z_AttributesPlusTerm *zapt, { unsigned reg_id; char *search_type = NULL; - char *rank_type = NULL; + char rank_type[128]; int complete_flag; int sort_flag; char termz[IT_MAX_WORD+1]; + RSET rset = 0; + int xpath_len; + int xpath_use = 0; + struct xpath_location_step xpath[10]; - zebra_maps_attr (zh->service->zebra_maps, zapt, ®_id, &search_type, - &rank_type, &complete_flag, &sort_flag); + zebra_maps_attr (zh->reg->zebra_maps, zapt, ®_id, &search_type, + rank_type, &complete_flag, &sort_flag); 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 (zapt_term_to_utf8(zh, zapt, termz)) + return 0; if (sort_flag) return rpn_sort_spec (zh, zapt, attributeSet, stream, sort_sequence, rank_type); + xpath_len = parse_xpath(zh, zapt, attributeSet, xpath, 10, stream); + if (xpath_len >= 0) + { + xpath_use = 1016; + if (xpath[xpath_len-1].part[0] == '@') + xpath_use = 1015; + } if (!strcmp (search_type, "phrase")) { - return rpn_search_APT_phrase (zh, zapt, termz, attributeSet, stream, + rset = rpn_search_APT_phrase (zh, zapt, termz, attributeSet, stream, reg_id, complete_flag, rank_type, + xpath_use, num_bases, basenames); } else if (!strcmp (search_type, "and-list")) { - return rpn_search_APT_and_list (zh, zapt, termz, attributeSet, stream, + rset = rpn_search_APT_and_list (zh, zapt, termz, attributeSet, stream, reg_id, complete_flag, rank_type, + xpath_use, num_bases, basenames); } else if (!strcmp (search_type, "or-list")) { - return rpn_search_APT_or_list (zh, zapt, termz, attributeSet, stream, + rset = rpn_search_APT_or_list (zh, zapt, termz, attributeSet, stream, reg_id, complete_flag, rank_type, + xpath_use, num_bases, basenames); } else if (!strcmp (search_type, "local")) { - return rpn_search_APT_local (zh, zapt, termz, attributeSet, stream, + rset = 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, + rset = rpn_search_APT_numeric (zh, zapt, termz, attributeSet, stream, reg_id, complete_flag, rank_type, + xpath_use, num_bases, basenames); } - zh->errCode = 118; - return NULL; + else if (!strcmp (search_type, "always")) + { + rset = 0; + } + else + zh->errCode = 118; + return rpn_search_xpath (zh, attributeSet, num_bases, basenames, + stream, rank_type, rset, xpath_len, xpath); } static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, @@ -2238,6 +2253,7 @@ static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, } bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare_it; + bool_parms.log_item = key_logdump_txt; switch (zop->which) { @@ -2251,21 +2267,11 @@ static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, 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) - { - 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); @@ -2274,31 +2280,26 @@ static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, sprintf (val, "%d", *zop->u.prox->u.known); return NULL; } -#else - if (*zop->u.prox->proximityUnitCode != Z_ProxUnit_word) - { - char *val = (char *) nmem_malloc (stream, 16); - zh->errCode = 132; - zh->errString = val; - sprintf (val, "%d", *zop->u.prox->proximityUnitCode); - return NULL; - } -#endif else { - RSET rsets[2]; - - rsets[0] = bool_parms.rset_l; - rsets[1] = bool_parms.rset_r; + /* new / old prox */ + rset_prox_parms parms; + RSET twosets[2]; - r = rpn_prox (zh, rsets, 2, - *zop->u.prox->ordered, - (!zop->u.prox->exclusion ? 0 : - *zop->u.prox->exclusion), - *zop->u.prox->relationType, - *zop->u.prox->distance); - rset_delete (rsets[0]); - rset_delete (rsets[1]); + twosets[0] = bool_parms.rset_l; + twosets[1] = bool_parms.rset_r; + parms.rset = twosets; + parms.rset_no = 2; + parms.ordered = *zop->u.prox->ordered; + parms.exclusion = (!zop->u.prox->exclusion ? 0 : + *zop->u.prox->exclusion); + parms.relation = *zop->u.prox->relationType; + parms.distance = *zop->u.prox->distance; + parms.key_size = sizeof(struct it_key); + parms.cmp = key_compare_it; + parms.getseq = key_get_seq; + parms.log_item = key_logdump_txt; + r = rset_create(rset_kind_prox, &parms); } break; default: @@ -2320,18 +2321,26 @@ static RSET rpn_search_structure (ZebraHandle zh, Z_RPNStructure *zs, logf (LOG_DEBUG, "rpn_search_ref"); r = resultSetRef (zh, zs->u.simple->u.resultSetId); if (!r) + { r = rset_create (rset_kind_null, NULL); + zh->errCode = 30; + zh->errString = + nmem_strdup (stream, zs->u.simple->u.resultSetId); + return 0; + } + else + rset_dup(r); } else { zh->errCode = 3; - return NULL; + return 0; } } else { zh->errCode = 3; - return NULL; + return 0; } return r; } @@ -2391,7 +2400,7 @@ RSET rpn_search (ZebraHandle zh, NMEM nmem, struct scan_info_entry { char *term; - ISAMS_P isam_p; + ISAMC_P isam_p; }; struct scan_info { @@ -2415,20 +2424,41 @@ static int scan_handle (char *name, const char *info, int pos, void *client) 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(ISAMS_P)); - memcpy (&scan_info->list[idx].isam_p, info+1, sizeof(ISAMS_P)); + assert (*info == sizeof(ISAMC_P)); + memcpy (&scan_info->list[idx].isam_p, info+1, sizeof(ISAMC_P)); return 0; } static void scan_term_untrans (ZebraHandle zh, NMEM stream, int reg_type, char **dst, const char *src) { - char term_dst[1024]; + char term_src[IT_MAX_WORD]; + char term_dst[IT_MAX_WORD]; - term_untrans (zh, reg_type, term_dst, src); - - *dst = (char *) nmem_malloc (stream, strlen(term_dst)+1); - strcpy (*dst, term_dst); + term_untrans (zh, reg_type, term_src, src); + + if (zh->iconv_from_utf8 != 0) + { + int len; + char *inbuf = term_src; + size_t inleft = strlen(term_src); + char *outbuf = term_dst; + size_t outleft = sizeof(term_dst)-1; + size_t ret; + + ret = yaz_iconv (zh->iconv_from_utf8, &inbuf, &inleft, + &outbuf, &outleft); + if (ret == (size_t)(-1)) + len = 0; + else + len = outbuf - term_dst; + *dst = nmem_malloc (stream, len + 1); + if (len > 0) + memcpy (*dst, term_dst, len); + (*dst)[len] = '\0'; + } + else + *dst = nmem_strdup (stream, term_src); } static void count_set (RSET r, int *count) @@ -2445,11 +2475,19 @@ static void count_set (RSET r, int *count) rfd = rset_open (r, RSETF_READ); while (rset_read (r, rfd, &key, &term_index)) { +#if IT_KEY_NEW + if (key.mem[0] != psysno) + { + psysno = key.mem[0]; + (*count)++; + } +#else if (key.sysno != psysno) { psysno = key.sysno; (*count)++; } +#endif kno++; } rset_close (r, rfd); @@ -2460,7 +2498,7 @@ 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 *is_partial, RSET limit_set, int return_zero) { int i; int pos = *position; @@ -2471,32 +2509,67 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, char termz[IT_MAX_WORD+20]; AttrType use; int use_value; + const char *use_string = 0; struct scan_info *scan_info_array; ZebraScanEntry *glist; int ords[32], ord_no = 0; int ptr[32]; + int bases_ok = 0; /* no of databases with OK attribute */ + int errCode = 0; /* err code (if any is not OK) */ + char *errString = 0; /* addinfo */ + unsigned reg_id; char *search_type = NULL; - char *rank_type = NULL; + char rank_type[128]; int complete_flag; int sort_flag; + *list = 0; if (attributeset == VAL_NONE) attributeset = VAL_BIB1; - logf (LOG_DEBUG, "position = %d, num = %d", pos, num); + if (!limit_set) + { + AttrType termset; + int termset_value_numeric; + const char *termset_value_string; + attr_init (&termset, zapt, 8); + termset_value_numeric = + attr_find_ex (&termset, NULL, &termset_value_string); + if (termset_value_numeric != -1) + { + char resname[32]; + const char *termset_name = 0; + + if (termset_value_numeric != -2) + { + + sprintf (resname, "%d", termset_value_numeric); + termset_name = resname; + } + else + termset_name = termset_value_string; + + limit_set = resultSetRef (zh, termset_name); + } + } + + yaz_log (LOG_DEBUG, "position = %d, num = %d set=%d", + pos, num, attributeset); attr_init (&use, zapt, 1); - use_value = attr_find (&use, &attributeset); + use_value = attr_find_ex (&use, &attributeset, &use_string); - if (zebra_maps_attr (zh->service->zebra_maps, zapt, ®_id, &search_type, - &rank_type, &complete_flag, &sort_flag)) + if (zebra_maps_attr (zh->reg->zebra_maps, zapt, ®_id, &search_type, + rank_type, &complete_flag, &sort_flag)) { + *num_entries = 0; zh->errCode = 113; return ; } + yaz_log (LOG_DEBUG, "use_value = %d", use_value); if (use_value == -1) use_value = 1016; @@ -2506,37 +2579,55 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, attent attp; data1_local_attribute *local_attr; - if ((r=att_getentbyatt (zh, &attp, attributeset, use_value))) + if ((r=att_getentbyatt (zh, &attp, attributeset, use_value, + use_string))) { logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d", attributeset, use_value); if (r == -1) - zh->errCode = 114; + { + char val_str[32]; + sprintf (val_str, "%d", use_value); + errCode = 114; + errString = odr_strdup (stream, val_str); + } else - zh->errCode = 121; - *num_entries = 0; - return; + errCode = 121; + continue; } - if (zebraExplain_curDatabase (zh->service->zei, basenames[base_no])) + if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no])) { zh->errString = basenames[base_no]; zh->errCode = 109; /* Database unavailable */ + *num_entries = 0; return; } + bases_ok++; for (local_attr = attp.local_attributes; local_attr && ord_no < 32; local_attr = local_attr->next) { int ord; - ord = zebraExplain_lookupSU (zh->service->zei, attp.attset_ordinal, + ord = zebraExplain_lookupSU (zh->reg->zei, attp.attset_ordinal, local_attr->local); if (ord > 0) ords[ord_no++] = ord; } } + if (!bases_ok && errCode) + { + zh->errCode = errCode; + zh->errString = errString; + *num_entries = 0; + } if (ord_no == 0) { - zh->errCode = 113; + char val_str[32]; + sprintf (val_str, "%d", use_value); + zh->errCode = 114; + zh->errString = odr_strdup (stream, val_str); + + *num_entries = 0; return; } /* prepare dictionary scanning */ @@ -2551,7 +2642,7 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, struct scan_info *scan_info = scan_info_array + i; struct rpn_char_map_info rcmi; - rpn_char_map_prepare (zh, reg_id, &rcmi); + rpn_char_map_prepare (zh->reg, reg_id, &rcmi); scan_info->before = before; scan_info->after = after; @@ -2562,14 +2653,15 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, for (j = 0; jlist[j].term = NULL; - prefix_len += key_SU_code (ords[i], termz + prefix_len); + prefix_len += key_SU_encode (ords[i], termz + prefix_len); termz[prefix_len++] = reg_id; termz[prefix_len] = 0; strcpy (scan_info->prefix, termz); - trans_scan_term (zh, zapt, termz+prefix_len, reg_id); + if (trans_scan_term (zh, zapt, termz+prefix_len, reg_id)) + return ; - dict_scan (zh->service->dict, termz, &before_tmp, &after_tmp, + dict_scan (zh->reg->dict, termz, &before_tmp, &after_tmp, scan_info, scan_handle); } glist = (ZebraScanEntry *) @@ -2603,7 +2695,7 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, &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); + NULL, 0, zapt->term->which); ptr[j0]++; for (j = j0+1; jterm->which); bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare_it; - bool_parms.rset_l = rset; + bool_parms.log_item = key_logdump_txt; + bool_parms.rset_l = rset; bool_parms.rset_r = rset2; rset = rset_create (rset_kind_or, &bool_parms); @@ -2630,6 +2724,18 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, ptr[j]++; } } + if (limit_set) + { + rset_bool_parms bool_parms; + + bool_parms.key_size = sizeof(struct it_key); + bool_parms.cmp = key_compare_it; + bool_parms.log_item = key_logdump_txt; + bool_parms.rset_l = rset; + bool_parms.rset_r = rset_dup(limit_set); + + rset = rset_create (rset_kind_and, &bool_parms); + } count_set (rset, &glist[i+before].occurrences); rset_delete (rset); } @@ -2669,7 +2775,7 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, rset = rset_trunc (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); + NULL, 0, zapt->term->which); ptr[j0]++; @@ -2685,11 +2791,13 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, 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); + strlen(glist[before-1-i].term), NULL, 0, + zapt->term->which); bool_parms.key_size = sizeof(struct it_key); bool_parms.cmp = key_compare_it; - bool_parms.rset_l = rset; + bool_parms.log_item = key_logdump_txt; + bool_parms.rset_l = rset; bool_parms.rset_r = rset2; rset = rset_create (rset_kind_or, &bool_parms); @@ -2697,6 +2805,18 @@ void rpn_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt, ptr[j]++; } } + if (limit_set) + { + rset_bool_parms bool_parms; + + bool_parms.key_size = sizeof(struct it_key); + bool_parms.cmp = key_compare_it; + bool_parms.log_item = key_logdump_txt; + bool_parms.rset_l = rset; + bool_parms.rset_r = rset_dup(limit_set); + + rset = rset_create (rset_kind_and, &bool_parms); + } count_set (rset, &glist[before-1-i].occurrences); rset_delete (rset); }