-/*
- * Copyright (C) 1995-1998, Index Data I/S
- * All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: zrpn.c,v $
- * 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 1.133 2003-04-15 20:48:04 adam Exp $
+ Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003
+ 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 <stdio.h>
#include <assert.h>
-#ifdef WINDOWS
+#ifdef WIN32
#include <io.h>
#else
#include <unistd.h>
#endif
#include <ctype.h>
-#include "zserver.h"
+#include "index.h"
+#include <zebra_xpath.h>
#include <charmap.h>
#include <rstemp.h>
#include <rsnull.h>
#include <rsbool.h>
+#include <rsbetween.h>
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 = vp;
- return zebra_maps_input (p->zm, p->reg_type, from, len);
+ struct rpn_char_map_info *p = (struct rpn_char_map_info *) vp;
+ 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->zebra_maps;
+ map_info->zm = reg->zebra_maps;
map_info->reg_type = reg_type;
- dict_grep_cmap (zh->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)
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);
}
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)
{
#ifdef TERM_COUNT
int *term_no;
#endif
- ISAM_P *isam_p_buf;
+ ISAMS_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->zebra_maps, reg_type, &src);
- while (*cp)
- *dst++ = *cp++;
+ const char *cp = zebra_maps_output (zh->reg->zebra_maps,
+ reg_type, &src);
+ if (!cp && len < IT_MAX_WORD-1)
+ dst[len++] = *src++;
+ else
+ 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,
struct grep_info *p)
{
- char term_tmp[1024];
if (p->isam_p_indx == p->isam_p_size)
{
- ISAM_P *new_isam_p_buf;
+ ISAMS_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 = xmalloc (sizeof(*new_isam_p_buf) *
- p->isam_p_size);
+ new_isam_p_buf = (ISAMS_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,
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,
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)++;
}
static int grep_handle (char *name, const char *info, void *p)
{
- add_isam_p (name, info, p);
+ add_isam_p (name, info, (struct grep_info *) p);
return 0;
}
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,
char *dst_term)
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;
{
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++;
return i;
}
+/* term_101: handle term, where trunc=Process # */
static int term_101 (ZebraMaps zebra_maps, int reg_type,
const char **src, char *dst, int space_split,
char *dst_term)
break;
while (s1 < s0)
{
- if (!isalnum (*s1))
+ if (strchr(REGEX_CHARS, *s1))
dst[i++] = '\\';
dst_term[j++] = *s1;
dst[i++] = *s1++;
return i;
}
-
+/* term_103: handle term, where trunc=re-2 (regular expressions) */
static int term_103 (ZebraMaps zebra_maps, int reg_type, const char **src,
char *dst, int *errors, int space_split,
char *dst_term)
break;
while (s1 < s0)
{
- if (!isalnum (*s1))
+ if (strchr(REGEX_CHARS, *s1))
dst[i++] = '\\';
dst_term[j++] = *s1;
dst[i++] = *s1++;
return i;
}
+/* term_103: handle term, where trunc=re-1 (regular expressions) */
static int term_102 (ZebraMaps zebra_maps, int reg_type, const char **src,
char *dst, int space_split, char *dst_term)
{
dst_term);
}
+
+/* term_104: handle term, where trunc=Process # and ! */
+static int term_104 (ZebraMaps zebra_maps, int reg_type,
+ 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;
+ s0 = *src;
+ while (*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 == '#')
+ {
+ dst[i++] = '.';
+ dst_term[j++] = *s0++;
+ }
+ {
+ s1 = s0;
+ map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0));
+ if (space_split && **map == *CHR_SPACE)
+ break;
+ while (s1 < s0)
+ {
+ if (strchr(REGEX_CHARS, *s1))
+ dst[i++] = '\\';
+ dst_term[j++] = *s1;
+ dst[i++] = *s1++;
+ }
+ }
+ }
+ dst[i] = '\0';
+ dst_term[j++] = '\0';
+ *src = s0;
+ return i;
+}
+
+/* term_105/106: handle term, where trunc=Process * and ! and right trunc */
+static int term_105 (ZebraMaps zebra_maps, int reg_type,
+ const char **src, char *dst, int space_split,
+ char *dst_term, int right_truncate)
+{
+ const char *s0, *s1;
+ const char **map;
+ int i = 0;
+ int j = 0;
+
+ if (!term_pre (zebra_maps, reg_type, src, "*!", "*!"))
+ return 0;
+ s0 = *src;
+ while (*s0)
+ {
+ if (*s0 == '*')
+ {
+ dst[i++] = '.';
+ dst[i++] = '*';
+ dst_term[j++] = *s0++;
+ }
+ else if (*s0 == '!')
+ {
+ dst[i++] = '.';
+ dst_term[j++] = *s0++;
+ }
+ {
+ s1 = s0;
+ map = zebra_maps_input (zebra_maps, reg_type, &s0, strlen(s0));
+ if (space_split && **map == *CHR_SPACE)
+ break;
+ while (s1 < s0)
+ {
+ if (strchr(REGEX_CHARS, *s1))
+ dst[i++] = '\\';
+ dst_term[j++] = *s1;
+ dst[i++] = *s1++;
+ }
+ }
+ }
+ if (right_truncate)
+ {
+ dst[i++] = '.';
+ dst[i++] = '*';
+ }
+ dst[i] = '\0';
+
+ dst_term[j++] = '\0';
+ *src = s0;
+ return i;
+}
+
+
/* gen_regular_rel - generate regular expression from relation
* val: border value (inclusive)
* islt: 1 if <=; 0 if >=.
}
else
{
- strcpy (dst, "((-");
+ strcpy (dst, "(-(");
islt = 0;
}
val = -val;
dst[dst_p] = '\0';
if (islt)
{
- for (i=1; i<pos; i++)
- strcat (dst, "[0-9]?");
+ /* match everything less than 10^(pos-1) */
+ strcat (dst, "0*");
+ for (i=1; i<pos; i++)
+ strcat (dst, "[0-9]?");
}
else
{
+ /* match everything greater than 10^pos */
for (i = 0; i <= pos; i++)
strcat (dst, "[0-9]");
strcat (dst, "[0-9]*");
strcat (dst, "))");
}
-static int relational_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
- const char **term_sub,
- char *term_dict,
- oid_value attributeSet,
- struct grep_info *grep_info,
- int *max_pos,
- int reg_type,
- char *term_dst)
+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[2*IT_MAX_WORD+20];
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 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
- term_dst))
+ if (!term_100 (zh->reg->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++ = '*';
+
+ if ((term_tmp - term_dict) > IT_MAX_WORD)
+ break;
+ }
+ *term_tmp++ = ')';
+ *term_tmp = '\0';
break;
case 2:
- if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
- term_dst))
+ if (!term_100 (zh->reg->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 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
- term_dst))
- 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++ = '|';
+
+ if ((term_tmp - term_dict) > IT_MAX_WORD)
+ break;
+ }
+ 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 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
- term_dst))
+ if (!term_100 (zh->reg->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++ = '|';
+
+ if ((term_tmp - term_dict) > IT_MAX_WORD)
+ break;
+ }
+ 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->reg->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++ = '*';
+
+ if ((term_tmp - term_dict) > IT_MAX_WORD)
+ break;
+ }
+ *term_tmp++ = ')';
+ *term_tmp = '\0';
break;
+ case 3:
default:
- return 0;
+ logf (LOG_DEBUG, "Relation =");
+ if (!term_100 (zh->reg->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 (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 field_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)
+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);
+
+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)
{
- char term_dict[2*IT_MAX_WORD+2];
+ 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;
AttrType truncation;
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);
+ rpn_char_map_prepare (zh->reg, reg_type, &rcmi);
attr_init (&use, zapt, 1);
- use_value = attr_find (&use, &curAttributeSet);
- logf (LOG_DEBUG, "field_term, use value %d", use_value);
+ 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)
- zh->errCode = 114;
- else
- zh->errCode = 121;
- return -1;
- }
- if (zebraExplain_curDatabase (zh->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;
}
- for (local_attr = attp.local_attributes; local_attr;
- local_attr = local_attr->next)
+ if (use_value == -2) /* string attribute (assume IDXPATH/any) */
{
- 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
+ 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)))
+ {
+ 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 */
+ 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;
+ }
+ }
+ 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->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; i<ord_len; i++)
- {
- term_dict[prefix_len++] = 1;
- term_dict[prefix_len++] = ord_buf[i];
- }
+
+ ord_len = key_SU_encode (ord, ord_buf);
+ for (i = 0; i<ord_len; i++)
+ {
+ term_dict[prefix_len++] = 1;
+ term_dict[prefix_len++] = ord_buf[i];
+ }
}
if (!prefix_len)
{
+ 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++] = ')';
+ 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 (!relational_term (zh, zapt, &termp, term_dict,
- attributeSet, grep_info, &max_pos, reg_type,
- term_dst))
- {
- j = prefix_len;
- switch (truncation_value)
- {
- case -1: /* not specified */
- case 100: /* do not truncate */
- term_dict[j++] = '(';
- if (!term_100 (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=none:%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: /* left truncation */
- case 3: /* left&right truncation */
- 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;
- }
+ 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_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 %d", r);
+ break;
+ case 1: /* right truncation */
+ term_dict[j++] = '(';
+ 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->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->reg->zebra_maps, reg_type,
+ &termp, term_dict + j, space_split, term_dst))
+ return 0;
+ strcat (term_dict, ")");
+ 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->reg->zebra_maps, reg_type,
+ &termp, term_dict + j, space_split, term_dst))
+ return 0;
+ strcat (term_dict, ".*)");
+ 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->reg->zebra_maps, reg_type,
+ &termp, term_dict + j, space_split, term_dst))
+ return 0;
+ strcat (term_dict, ")");
+ 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->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->reg->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->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->reg->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;
+ case 104: /* process # and ! in term */
+ term_dict[j++] = '(';
+ 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->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->reg->zebra_maps, reg_type,
+ &termp, term_dict + j, space_split, term_dst, 1))
+ return 0;
+ strcat (term_dict, ")");
+ 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->reg->zebra_maps, reg_type,
+ &termp, term_dict + j, space_split, term_dst, 0))
+ return 0;
+ strcat (term_dict, ")");
+ 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;
}
}
*term_sub = termp;
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->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_proximity (ZebraHandle zh, RSET rset1, RSET rset2,
- 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 rsfd1, rsfd2;
- int more1, more2;
- struct it_key buf1, buf2;
- RSFD rsfd_result;
- RSET result;
- rset_temp_parms parms;
- int term_index;
-
- rsfd1 = rset_open (rset1, RSETF_READ);
- more1 = rset_read (rset1, rsfd1, &buf1, &term_index);
-
- rsfd2 = rset_open (rset2, RSETF_READ);
- more2 = rset_read (rset2, rsfd2, &buf2, &term_index);
+ char termz0[IT_MAX_WORD];
- 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);
-
- logf (LOG_DEBUG, "rpn_proximity excl=%d ord=%d rel=%d dis=%d",
- exclusion, ordered, relation, distance);
- while (more1 && more2)
- {
- int cmp = key_compare_it (&buf1, &buf2);
- if (cmp < -1)
- more1 = rset_read (rset1, rsfd1, &buf1, &term_index);
- else if (cmp > 1)
- more2 = rset_read (rset2, rsfd2, &buf2, &term_index);
- else
+ if (zapt_term_to_utf8(zh, zapt, termz0))
+ return -1; /* error */
+ else
+ {
+ const char **map;
+ const char *cp = (const char *) termz0;
+ const char *cp_end = cp + strlen(cp);
+ const char *src;
+ int i = 0;
+ const char *space_map = NULL;
+ int len;
+
+ while ((len = (cp_end - cp)) > 0)
{
- int sysno = buf1.sysno;
- int seqno[500];
- int n = 0;
-
- seqno[n++] = buf1.seqno;
- while ((more1 = rset_read (rset1, rsfd1, &buf1, &term_index)) &&
- sysno == buf1.sysno)
- if (n < 500)
- seqno[n++] = buf1.seqno;
- do
+ map = zebra_maps_input (zh->reg->zebra_maps, reg_type, &cp, len);
+ if (**map == *CHR_SPACE)
+ space_map = *map;
+ else
{
- for (i = 0; i<n; i++)
- {
- int diff = buf2.seqno - seqno[i];
- int excl = exclusion;
- if (!ordered && diff < 0)
- diff = -diff;
- switch (relation)
- {
- case 1: /* < */
- if (diff < distance)
- excl = !excl;
- break;
- case 2: /* <= */
- if (diff <= distance)
- excl = !excl;
- break;
- case 3: /* == */
- if (diff == distance)
- excl = !excl;
- break;
- case 4: /* >= */
- if (diff >= distance)
- excl = !excl;
- break;
- case 5: /* > */
- if (diff > distance)
- excl = !excl;
- break;
- case 6: /* != */
- if (diff != distance)
- excl = !excl;
- break;
- }
- if (excl)
- rset_write (result, rsfd_result, &buf2);
- }
- } while ((more2 = rset_read (rset2, rsfd2, &buf2, &term_index)) &&
- sysno == buf2.sysno);
+ 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';
}
- rset_close (result, rsfd_result);
- rset_close (rset1, rsfd1);
- rset_close (rset2, rsfd2);
- return result;
+ return 0;
}
-static RSET rpn_prox (ZebraHandle zh, RSET *rset, int rset_no)
+static RSET rpn_prox (ZebraHandle zh, RSET *rset, int rset_no,
+ int ordered, int exclusion, int relation, int distance)
{
int i;
RSFD *rsfd;
int length_prox_term = 0;
int min_nn = 10000000;
int term_index;
+ int term_type = Z_Term_characterString;
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);
+ *prox_term = '\0';
for (i = 0; i<rset_no; i++)
{
int j;
- buf[i] = 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; j<rset[i]->no_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 (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;
+ term_type = rset[i]->rset_terms[j]->type;
+
+ /* only if all term types are of type characterString .. */
+ /* the resulting term is of that type */
+ if (term_type != Z_Term_characterString)
+ term_type = Z_Term_general;
}
}
+ for (i = 0; i<rset_no; i++)
+ {
+ buf[i] = 0;
+ rsfd[i] = 0;
+ }
+ for (i = 0; i<rset_no; i++)
+ {
+ 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;
+ }
if (i != rset_no)
{
+ /* at least one is empty ... return null set */
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 = rset_term_create (prox_term, length_prox_term,
+ flags, term_type);
parms.rset_term->nn = 0;
result = rset_create (rset_kind_null, &parms);
}
- else
+ 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, -1, flags);
+ parms.rset_term = rset_term_create (prox_term, length_prox_term,
+ flags, term_type);
parms.rset_term->nn = min_nn;
+ parms.cmp = key_compare_it;
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<rset_no; i++)
more[0] = rset_read (*rset, *rsfd, *buf, &term_index);
}
}
-
- for (i = 0; i<rset_no; i++)
+ rset_close (result, rsfd_result);
+ }
+ else if (rset_no == 2)
+ {
+ /* generic proximity case (two input sets only) ... */
+ rset_temp_parms parms;
+ RSFD rsfd_result;
+
+ yaz_log (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, term_type);
+ parms.rset_term->nn = min_nn;
+ parms.cmp = key_compare_it;
+ 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[0] && more[1])
{
- rset_close (rset[i], rsfd[i]);
- xfree (buf[i]);
+ 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; i<n; i++)
+ {
+ int diff = buf[1]->seqno - 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);
}
+ else
+ {
+ rset_null_parms parms;
+
+ parms.rset_term = rset_term_create (prox_term, length_prox_term,
+ flags, term_type);
+ parms.rset_term->nn = 0;
+ result = rset_create (rset_kind_null, &parms);
+ }
+ for (i = 0; i<rset_no; i++)
+ {
+ if (rsfd[i])
+ rset_close (rset[i], rsfd[i]);
+ xfree (buf[i]);
+ }
xfree (buf);
xfree (more);
xfree (rsfd);
return result;
}
+
+char *normalize_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt,
+ const char *termz, NMEM stream, unsigned reg_id)
+{
+ WRBUF wrbuf = 0;
+ AttrType truncation;
+ int truncation_value;
+ char *ex_list = 0;
+
+ attr_init (&truncation, zapt, 5);
+ truncation_value = attr_find (&truncation, NULL);
+
+ switch (truncation_value)
+ {
+ default:
+ ex_list = "";
+ break;
+ case 101:
+ ex_list = "#";
+ break;
+ case 102:
+ case 103:
+ ex_list = 0;
+ break;
+ case 104:
+ ex_list = "!#";
+ break;
+ case 105:
+ ex_list = "!*";
+ break;
+ }
+ if (ex_list)
+ wrbuf = zebra_replace(zh->reg->zebra_maps, reg_id, ex_list,
+ termz, strlen(termz));
+ if (!wrbuf)
+ return nmem_strdup(stream, termz);
+ else
+ {
+ char *buf = (char*) nmem_malloc (stream, wrbuf_len(wrbuf)+1);
+ memcpy (buf, wrbuf_buf(wrbuf), wrbuf_len(wrbuf));
+ buf[wrbuf_len(wrbuf)] = '\0';
+ return buf;
+ }
+}
+
+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,
+ 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, rset_no = 0;
+ struct grep_info grep_info;
+ char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type);
const char *termp = termz;
+
+ *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);
+ 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;
+ if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset)))
+ break;
+ }
+ grep_info_delete (&grep_info);
+ if (rset_no == 0)
+ {
+ rset_null_parms parms;
+
+ 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<rset_no; i++)
+ rset_delete (rset[i]);
+ return result;
+}
+
+static RSET rpn_search_APT_or_list (ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const char *termz_org,
+ oid_value attributeSet,
+ NMEM stream,
+ int reg_type, int complete_flag,
+ 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_or_list termp=%s", termp);
+ 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;
+ if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset)))
+ break;
+ }
+ grep_info_delete (&grep_info);
+ if (rset_no == 0)
+ {
+ rset_null_parms parms;
+
+ parms.rset_term = rset_term_create (termz, -1, rank_type,
+ zapt->term->which);
+ return rset_create (rset_kind_null, &parms);
+ }
+ result = rset[0];
+ for (i = 1; i<rset_no; i++)
+ {
+ rset_bool_parms bool_parms;
+
+ bool_parms.rset_l = result;
+ bool_parms.rset_r = rset[i];
+ bool_parms.key_size = sizeof(struct it_key);
+ bool_parms.cmp = key_compare_it;
+ result = rset_create (rset_kind_or, &bool_parms);
+ }
+ return result;
+}
+static RSET rpn_search_APT_and_list (ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const char *termz_org,
+ oid_value attributeSet,
+ NMEM stream,
+ int reg_type, int complete_flag,
+ 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, rset_no = 0;
+ struct grep_info grep_info;
+ char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type);
+ const char *termp = termz;
+
+ 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 = field_term (zh, zapt, &termp, attributeSet, &grep_info,
- reg_type, complete_flag, num_bases, basenames,
- term_dst);
- if (r < 1)
+ logf (LOG_DEBUG, "APT_and_list termp=%s", termp);
+ 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 >= sizeof(rset)/sizeof(*rset))
+ 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);
- for (i = 0; i<rset_no; i++)
- rset_delete (rset[i]);
- return result;
+ result = rset[0];
+ for (i = 1; i<rset_no; i++)
+ {
+ rset_bool_parms bool_parms;
+
+ bool_parms.rset_l = result;
+ bool_parms.rset_r = rset[i];
+ bool_parms.key_size = sizeof(struct it_key);
+ bool_parms.cmp = key_compare_it;
+ result = rset_create (rset_kind_and, &bool_parms);
+ }
+ return result;
+}
+
+static int numeric_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
+ const char **term_sub,
+ char *term_dict,
+ oid_value attributeSet,
+ struct grep_info *grep_info,
+ int *max_pos,
+ int reg_type,
+ char *term_dst)
+{
+ AttrType relation;
+ int relation_value;
+ int term_value;
+ int r;
+ char *term_tmp = term_dict + strlen(term_dict);
+
+ attr_init (&relation, zapt, 2);
+ relation_value = attr_find (&relation, NULL);
+
+ logf (LOG_DEBUG, "numeric relation value=%d", relation_value);
+
+ if (!term_100 (zh->reg->zebra_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->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_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, 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, reg_type, &rcmi);
+ attr_init (&use, zapt, 1);
+ use_value = attr_find_ex (&use, &curAttributeSet, &use_string);
+
+ if (use_value == -1)
+ 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 (use_value == -2) /* string attribute (assume IDXPATH/any) */
+ {
+ 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)))
+ {
+ 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->errString = nmem_strdup (stream, val_str);
+ zh->errCode = 114;
+ }
+ else
+ zh->errCode = 121;
+ return -1;
+ }
+ }
+ if (zebraExplain_curDatabase (zh->reg->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->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_encode (ord, ord_buf);
+ for (i = 0; i<ord_len; i++)
+ {
+ term_dict[prefix_len++] = 1;
+ term_dict[prefix_len++] = ord_buf[i];
+ }
+ }
+ if (!prefix_len)
+ {
+ 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++] = ')';
+ 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_or_list (ZebraHandle zh,
- Z_AttributesPlusTerm *zapt,
+static RSET rpn_search_APT_numeric (ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
const char *termz,
- oid_value attributeSet,
+ 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];
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_or_list termp=%s", termp);
- grep_info.isam_p_indx = 0;
- r = field_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 >= 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<rset_no; i++)
- {
- rset_bool_parms bool_parms;
-
- bool_parms.rset_l = result;
- bool_parms.rset_r = rset[i];
- bool_parms.key_size = sizeof(struct it_key);
- bool_parms.cmp = key_compare_it;
- result = rset_create (rset_kind_or, &bool_parms);
- }
- return result;
-}
-
-static RSET rpn_search_APT_and_list (ZebraHandle zh,
- Z_AttributesPlusTerm *zapt,
- const char *termz,
- oid_value attributeSet,
- 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;
-
+ if (grep_info_prepare (zh, zapt, &grep_info, reg_type, stream))
+ return 0;
while (1)
{
- logf (LOG_DEBUG, "APT_and_list termp=%s", termp);
+ logf (LOG_DEBUG, "APT_numeric termp=%s", termp);
grep_info.isam_p_indx = 0;
- r = field_term (zh, zapt, &termp, attributeSet, &grep_info,
- reg_type, complete_flag, num_bases, basenames,
- term_dst);
+ r = numeric_term (zh, zapt, &termp, attributeSet, &grep_info,
+ reg_type, complete_flag, num_bases, basenames,
+ 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 >= sizeof(rset)/sizeof(*rset))
+ 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];
return result;
}
-
static RSET rpn_search_APT_local (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
const char *termz,
oid_value attributeSet,
+ NMEM stream,
const char *rank_type)
{
RSET result;
struct it_key key;
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->res, "setTmpDir");
result = rset_create (rset_kind_temp, &parms);
return result;
}
+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;
+ char termz[20];
+
+ 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; i<sort_sequence->num_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;
+ sprintf (termz, "%d", i);
+
+ 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;
+
+ sks->which = Z_SortKeySpec_null;
+ sks->u.null = odr_nullval ();
+ sort_sequence->specs[i] = sks;
+
+ parms.rset_term = rset_term_create (termz, -1, rank_type,
+ zapt->term->which);
+ return rset_create (rset_kind_null, &parms);
+}
+
+/* pop - moved to xpath.c */
+#if 0
+
+struct xpath_predicate {
+ int which;
+ union {
+#define XPATH_PREDICATE_RELATION 1
+ struct {
+ char *name;
+ char *op;
+ char *value;
+ } relation;
+#define XPATH_PREDICATE_BOOLEAN 2
+ struct {
+ const char *op;
+ struct xpath_predicate *left;
+ struct xpath_predicate *right;
+ } boolean;
+ } u;
+};
+
+struct xpath_location_step {
+ char *part;
+ struct xpath_predicate *predicate;
+};
+
+#endif
+
+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; i<ord_len; i++)
+ {
+ term_dict[prefix_len++] = 1;
+ term_dict[prefix_len++] = ord_buf[i];
+ }
+ term_dict[prefix_len++] = ')';
+ term_dict[prefix_len++] = 1;
+ term_dict[prefix_len++] = reg_type;
+
+ strcpy (term_dict+prefix_len, term);
+
+ grep_info.isam_p_indx = 0;
+ r = dict_lookup_grep (zh->reg->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<xpath_len; i++)
+ {
+ yaz_log (LOG_LOG, "XPATH %d %s", i, xpath[i].part);
+
+ }
+
+ curAttributeSet = VAL_IDXPATH;
+
+ /*
+ //a -> 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])
+ {
+ char predicate_str[128];
+
+ strcpy (predicate_str,
+ xpath[level].predicate->u.relation.name+1);
+ if (xpath[level].predicate->u.relation.value)
+ {
+ strcat (predicate_str, "=");
+ strcat (predicate_str,
+ xpath[level].predicate->u.relation.value);
+ }
+ rset_attr = xpath_trunc (
+ zh, stream, '0', predicate_str, 3, curAttributeSet);
+ }
+ 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,
+ oid_value attributeSet, NMEM stream,
+ Z_SortKeySpecList *sort_sequence,
int num_bases, char **basenames)
{
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->zebra_maps, zapt, ®_id, &search_type,
- &rank_type, &complete_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)
+ 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)
{
- zh->errCode = 124;
- return NULL;
+ xpath_use = 1016;
+ if (xpath[xpath_len-1].part[0] == '@')
+ xpath_use = 1015;
}
- trans_term (zh, zapt, termz);
if (!strcmp (search_type, "phrase"))
{
- return rpn_search_APT_phrase (zh, zapt, termz, attributeSet,
+ 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,
+ 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,
+ 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,
+ rset = rpn_search_APT_local (zh, zapt, termz, attributeSet, stream,
rank_type);
}
- zh->errCode = 118;
- return NULL;
+ else if (!strcmp (search_type, "numeric"))
+ {
+ rset = rpn_search_APT_numeric (zh, zapt, termz, attributeSet, stream,
+ reg_id, complete_flag, rank_type,
+ xpath_use,
+ num_bases, basenames);
+ }
+ 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,
- oid_value attributeSet, ODR stream,
+ oid_value attributeSet, NMEM stream,
+ Z_SortKeySpecList *sort_sequence,
int num_bases, char **basenames)
{
RSET r = NULL;
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;
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)
{
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 = odr_malloc (stream, 16);
+ 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 (stream, 16);
- zh->errCode = 132;
- zh->errString = val;
- sprintf (val, "%d", *zop->u.prox->proximityUnitCode);
- return NULL;
- }
-#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),
- *zop->u.prox->relationType,
- *zop->u.prox->distance);
+ else
+ {
+ RSET rsets[2];
+
+ rsets[0] = bool_parms.rset_l;
+ rsets[1] = bool_parms.rset_r;
+
+ 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]);
+ }
break;
default:
zh->errCode = 110;
{
logf (LOG_DEBUG, "rpn_search_APT");
r = rpn_search_APT (zh, zs->u.simple->u.attributesPlusTerm,
- attributeSet, num_bases, basenames);
+ attributeSet, stream, sort_sequence,
+ num_bases, basenames);
}
else if (zs->u.simple->which == Z_Operand_resultSetId)
{
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;
}
-void rpn_search (ZebraHandle zh, ODR stream,
+
+RSET rpn_search (ZebraHandle zh, NMEM nmem,
Z_RPNQuery *rpn, int num_bases, char **basenames,
- const char *setname)
+ const char *setname,
+ ZebraSet sset)
{
RSET rset;
oident *attrset;
oid_value attributeSet;
-
- zlog_rpn (rpn);
+ 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; i<sort_sequence->num_specs; i++)
+ sort_sequence->specs[i] = 0;
+
attrset = oid_getentbyoid (rpn->attributeSetId);
attributeSet = attrset->value;
- rset = rpn_search_structure (zh, rpn->RPNStructure, attributeSet, stream,
- num_bases, basenames);
+ rset = rpn_search_structure (zh, rpn->RPNStructure, attributeSet,
+ nmem, sort_sequence, num_bases, basenames);
if (!rset)
- return;
+ return 0;
- resultSetAdd (zh, setname, 1, rset, &zh->hits);
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
+ {
+ 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 {
char *term;
- ISAM_P isam_p;
+ ISAMS_P isam_p;
};
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))
return 1;
- if (pos > 0)
- idx = scan_info->after - pos + scan_info->before;
+ if (pos > 0) 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));
+ assert (*info == sizeof(ISAMS_P));
+ memcpy (&scan_info->list[idx].isam_p, info+1, sizeof(ISAMS_P));
return 0;
}
-static void scan_term_untrans (ZebraHandle zh, ODR stream, int reg_type,
+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);
+ char term_src[IT_MAX_WORD];
+ char term_dst[IT_MAX_WORD];
- *dst = odr_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)
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;
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;
- zlog_scan (zapt, attributeset);
- 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);
- if (zebra_maps_attr (zh->zebra_maps, zapt, ®_id, &search_type,
- &rank_type, &complete_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;
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);
+ zh->errCode = 114;
+ zh->errString = odr_strdup (stream, val_str);
+ }
else
zh->errCode = 121;
+ *num_entries = 0;
+ return;
}
- if (zebraExplain_curDatabase (zh->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;
}
for (local_attr = attp.local_attributes; local_attr && ord_no < 32;
{
int ord;
- ord = zebraExplain_lookupSU (zh->zei, attp.attset_ordinal,
+ ord = zebraExplain_lookupSU (zh->reg->zei, attp.attset_ordinal,
local_attr->local);
if (ord > 0)
ords[ord_no++] = ord;
}
if (ord_no == 0)
{
+ *num_entries = 0;
zh->errCode = 113;
return;
}
+ /* prepare dictionary scanning */
before = pos-1;
after = 1+num-pos;
- scan_info_array = odr_malloc (stream, 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;
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;
scan_info->odr = stream;
- scan_info->list = odr_malloc (stream, (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; j<before+after; j++)
scan_info->list[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->dict, termz, &before_tmp, &after_tmp, scan_info,
- scan_handle);
+ dict_scan (zh->reg->dict, termz, &before_tmp, &after_tmp,
+ scan_info, scan_handle);
}
- glist = odr_malloc (stream, (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;
}
if (j0 == -1)
break;
- scan_term_untrans (zh, stream, reg_id,
+ scan_term_untrans (zh, stream->mem, 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);
+ NULL, 0, zapt->term->which);
ptr[j0]++;
for (j = j0+1; j<ord_no; j++)
rset2 =
rset_trunc (zh, &scan_info_array[j].list[ptr[j]].isam_p, 1,
glist[i+before].term,
- strlen(glist[i+before].term), NULL);
+ strlen(glist[i+before].term), NULL, 0,
+ zapt->term->which);
bool_parms.key_size = sizeof(struct it_key);
bool_parms.cmp = key_compare_it;
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.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);
}
*is_partial = 1;
}
+ /* consider terms before main term */
for (i = 0; i<ord_no; i++)
ptr[i] = 0;
if (j0 == -1)
break;
- scan_term_untrans (zh, stream, reg_id,
+ scan_term_untrans (zh, stream->mem, reg_id,
&glist[before-1-i].term, mterm);
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]++;
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;
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.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);
}