/*
- * Copyright (C) 1995-1998, Index Data I/S
+ * Copyright (C) 1995-2000, Index Data
* All rights reserved.
* Sebastian Hammer, Adam Dickmeiss
*
* $Log: zrpn.c,v $
- * Revision 1.79 1998-06-22 11:35:09 adam
+ * 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
*/
#include <stdio.h>
#include <assert.h>
-#ifdef WINDOWS
+#ifdef WIN32
#include <io.h>
#else
#include <unistd.h>
static const char **rpn_char_map_handler (void *vp, const char **from, int len)
{
- struct rpn_char_map_info *p = vp;
+ struct rpn_char_map_info *p = (struct rpn_char_map_info *) vp;
return zebra_maps_input (p->zm, p->reg_type, from, len);
}
static void rpn_char_map_prepare (ZebraHandle zh, int reg_type,
struct rpn_char_map_info *map_info)
{
- map_info->zm = zh->zebra_maps;
+ map_info->zm = zh->service->zebra_maps;
map_info->reg_type = reg_type;
- dict_grep_cmap (zh->dict, map_info, rpn_char_map_handler);
+ dict_grep_cmap (zh->service->dict, map_info, rpn_char_map_handler);
}
typedef struct {
#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;
{
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->service->zebra_maps,
+ reg_type, &src);
+ if (!cp)
+ *dst++ = *src++;
+ else
+ while (*cp)
+ *dst++ = *cp++;
}
*dst = '\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,
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;
}
+/* 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)
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)
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)
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[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 (!isalnum (*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 (!isalnum (*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[256];
attr_init (&relation, zapt, 2);
relation_value = attr_find (&relation, NULL);
- logf (LOG_DEBUG, "relation value=%d", relation_value);
+ logf (LOG_DEBUG, "string relation value=%d", relation_value);
switch (relation_value)
{
case 1:
- if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
- term_dst))
+ if (!term_100 (zh->service->zebra_maps, reg_type,
+ term_sub, term_component,
+ space_split, term_dst))
return 0;
- term_value = atoi (term_tmp);
logf (LOG_DEBUG, "Relation <");
- gen_regular_rel (term_tmp, term_value-1, 1);
+
+ *term_tmp++ = '(';
+ for (i = 0; term_component[i]; )
+ {
+ int j = 0;
+
+ if (i)
+ *term_tmp++ = '|';
+ while (j < i)
+ string_rel_add_char (&term_tmp, term_component, &j);
+
+ *term_tmp++ = '[';
+
+ *term_tmp++ = '^';
+ string_rel_add_char (&term_tmp, term_component, &i);
+ *term_tmp++ = '-';
+
+ *term_tmp++ = ']';
+ *term_tmp++ = '.';
+ *term_tmp++ = '*';
+ }
+ *term_tmp++ = ')';
+ *term_tmp = '\0';
break;
case 2:
- if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
- term_dst))
+ if (!term_100 (zh->service->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++ = '|';
+ }
+ 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->service->zebra_maps, reg_type,
+ term_sub, term_component, space_split, term_dst))
return 0;
- term_value = atoi (term_tmp);
logf (LOG_DEBUG, "Relation >");
- gen_regular_rel (term_tmp, term_value+1, 0);
+
+ *term_tmp++ = '(';
+ for (i = 0; term_component[i];)
+ {
+ int j = 0;
+
+ while (j < i)
+ string_rel_add_char (&term_tmp, term_component, &j);
+ *term_tmp++ = '[';
+
+ *term_tmp++ = '^';
+ *term_tmp++ = '-';
+ string_rel_add_char (&term_tmp, term_component, &i);
+
+ *term_tmp++ = ']';
+ *term_tmp++ = '.';
+ *term_tmp++ = '*';
+
+ *term_tmp++ = '|';
+ }
+ for (i = 0; term_component[i];)
+ string_rel_add_char (&term_tmp, term_component, &i);
+ *term_tmp++ = '.';
+ *term_tmp++ = '+';
+ *term_tmp++ = ')';
+ *term_tmp = '\0';
break;
+ case 4:
+ if (!term_100 (zh->service->zebra_maps, reg_type, term_sub,
+ term_component, space_split, term_dst))
+ return 0;
+ logf (LOG_DEBUG, "Relation >=");
+
+ *term_tmp++ = '(';
+ for (i = 0; term_component[i];)
+ {
+ int j = 0;
+
+ if (i)
+ *term_tmp++ = '|';
+ while (j < i)
+ string_rel_add_char (&term_tmp, term_component, &j);
+ *term_tmp++ = '[';
+
+ if (term_component[i+1])
+ {
+ *term_tmp++ = '^';
+ *term_tmp++ = '-';
+ string_rel_add_char (&term_tmp, term_component, &i);
+ }
+ else
+ {
+ string_rel_add_char (&term_tmp, term_component, &i);
+ *term_tmp++ = '-';
+ }
+ *term_tmp++ = ']';
+ *term_tmp++ = '.';
+ *term_tmp++ = '*';
+ }
+ *term_tmp++ = ')';
+ *term_tmp = '\0';
+ break;
+ case 3:
default:
- return 0;
+ logf (LOG_DEBUG, "Relation =");
+ if (!term_100 (zh->service->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)
{
- char term_dict[2*IT_MAX_WORD+2];
+ char term_dict[2*IT_MAX_WORD+4000];
int j, r, base_no;
AttrType truncation;
int truncation_value;
rpn_char_map_prepare (zh, reg_type, &rcmi);
attr_init (&use, zapt, 1);
use_value = attr_find (&use, &curAttributeSet);
- logf (LOG_DEBUG, "field_term, use value %d", use_value);
+ logf (LOG_DEBUG, "string_term, use value %d", use_value);
attr_init (&truncation, zapt, 5);
truncation_value = attr_find (&truncation, NULL);
logf (LOG_DEBUG, "truncation value %d", truncation_value);
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->zei, basenames[base_no]))
+ if (zebraExplain_curDatabase (zh->service->zei, basenames[base_no]))
{
zh->errCode = 109; /* Database unavailable */
zh->errString = basenames[base_no];
char ord_buf[32];
int i, ord_len;
- ord = zebraExplain_lookupSU (zh->zei, attp.attset_ordinal,
+ ord = zebraExplain_lookupSU (zh->service->zei, attp.attset_ordinal,
local_attr->local);
if (ord < 0)
continue;
}
if (!prefix_len)
{
- zh->errCode = 114;
+ char val_str[32];
+ sprintf (val_str, "%d", use_value);
+ zh->errCode = 114;
+ zh->errString = nmem_strdup (stream, val_str);
return -1;
}
term_dict[prefix_len++] = ')';
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_DEBUG, "dict_lookup_grep: %s", term_dict+prefix_len);
+ r = dict_lookup_grep (zh->service->dict, term_dict, 0,
+ grep_info, &max_pos, 0, grep_handle);
+ if (r)
+ logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r);
+ break;
+ case 1: /* right truncation */
+ term_dict[j++] = '(';
+ if (!term_100 (zh->service->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,
+ &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,
+ &termp, term_dict + j, space_split, term_dst))
+ return 0;
+ strcat (term_dict, ")");
+ dict_lookup_grep (zh->service->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,
+ &termp, term_dict + j, space_split, term_dst))
+ return 0;
+ strcat (term_dict, ".*)");
+ dict_lookup_grep (zh->service->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,
+ &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,
+ &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,
+ &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,
+ &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->service->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,
+ &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->service->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,
+ &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,
+ &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,
+ &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,
+ &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,
+ &max_pos, 0, grep_handle);
+ if (r)
+ logf (LOG_WARN, "dict_lookup_grep err, trunc=*/!: %d", r);
+ break;
}
}
*term_sub = termp;
while ((len = (cp_end - cp)) > 0)
{
- map = zebra_maps_input (zh->zebra_maps, reg_type, &cp, len);
+ map = zebra_maps_input (zh->service->zebra_maps, reg_type, &cp, len);
if (**map == *CHR_SPACE)
space_map = *map;
else
termz[i] = '\0';
}
-static RSET rpn_proximity (ZebraHandle zh, RSET rset1, RSET rset2,
- int ordered,
- int exclusion, int relation, int distance)
-{
- 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);
-
- 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
- {
- 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
- {
- 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);
- }
- }
- rset_close (result, rsfd_result);
- rset_close (rset1, rsfd1);
- rset_close (rset2, rsfd2);
- return result;
-}
-
-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 term_index;
const char *flags = NULL;
- rsfd = xmalloc (sizeof(*rsfd)*rset_no);
- more = xmalloc (sizeof(*more)*rset_no);
- buf = xmalloc (sizeof(*buf)*rset_no);
+ rsfd = (RSFD *) xmalloc (sizeof(*rsfd)*rset_no);
+ more = (int *) xmalloc (sizeof(*more)*rset_no);
+ buf = (struct it_key **) xmalloc (sizeof(*buf)*rset_no);
+ *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;
}
}
+ 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);
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);
parms.rset_term->nn = min_nn;
parms.key_size = sizeof (struct it_key);
- parms.temp_path = res_get (zh->res, "setTmpDir");
+ parms.temp_path = res_get (zh->service->res, "setTmpDir");
result = rset_create (rset_kind_temp, &parms);
rsfd_result = rset_open (result, RSETF_WRITE);
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;
+
+ 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);
+
+ 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);
+ 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->service->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 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,
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;
+ char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type);
+ const char *termp = termz;
#ifdef TERM_COUNT
grep_info.term_no = 0;
{
logf (LOG_DEBUG, "APT_phrase termp=%s", termp);
grep_info.isam_p_indx = 0;
- r = field_term (zh, zapt, &termp, attributeSet, &grep_info,
+ r = string_term (zh, zapt, &termp, attributeSet, stream, &grep_info,
reg_type, complete_flag, num_bases, basenames,
term_dst);
if (r < 1)
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
}
else if (rset_no == 1)
return (rset[0]);
- result = rpn_prox (zh, rset, rset_no);
+ 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,
+ const char *termz_org,
oid_value attributeSet,
+ NMEM stream,
int reg_type, int complete_flag,
const char *rank_type,
int num_bases, char **basenames)
{
char term_dst[IT_MAX_WORD+1];
- const char *termp = termz;
RSET rset[60], result;
int i, r, rset_no = 0;
struct grep_info grep_info;
-
+ char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type);
+ const char *termp = termz;
#ifdef TERM_COUNT
grep_info.term_no = 0;
#endif
{
logf (LOG_DEBUG, "APT_or_list termp=%s", termp);
grep_info.isam_p_indx = 0;
- r = field_term (zh, zapt, &termp, attributeSet, &grep_info,
+ r = string_term (zh, zapt, &termp, attributeSet, stream, &grep_info,
reg_type, complete_flag, num_bases, basenames,
term_dst);
if (r < 1)
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
static RSET rpn_search_APT_and_list (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,
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;
+ char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type);
+ const char *termp = termz;
#ifdef TERM_COUNT
grep_info.term_no = 0;
{
logf (LOG_DEBUG, "APT_and_list termp=%s", termp);
grep_info.isam_p_indx = 0;
- r = field_term (zh, zapt, &termp, attributeSet, &grep_info,
+ r = string_term (zh, zapt, &termp, attributeSet, stream, &grep_info,
reg_type, complete_flag, num_bases, basenames,
term_dst);
if (r < 1)
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
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->service->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->service->dict, term_dict, 0, grep_info, max_pos,
+ 0, grep_handle);
+ if (r)
+ logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r);
+ logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx);
+ return 1;
+}
+
+static int numeric_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
+ const char **term_sub,
+ oid_value attributeSet, struct grep_info *grep_info,
+ int reg_type, int complete_flag,
+ int num_bases, char **basenames,
+ char *term_dst)
+{
+ char term_dict[2*IT_MAX_WORD+2];
+ int r, base_no;
+ AttrType use;
+ int use_value;
+ oid_value curAttributeSet = attributeSet;
+ const char *termp;
+ struct rpn_char_map_info rcmi;
+
+ rpn_char_map_prepare (zh, reg_type, &rcmi);
+ attr_init (&use, zapt, 1);
+ use_value = attr_find (&use, &curAttributeSet);
+ logf (LOG_DEBUG, "numeric_term, use value %d", use_value);
+
+ if (use_value == -1)
+ use_value = 1016;
+
+ for (base_no = 0; base_no < num_bases; base_no++)
+ {
+ attent attp;
+ data1_local_attribute *local_attr;
+ int max_pos, prefix_len = 0;
+
+ termp = *term_sub;
+ if ((r=att_getentbyatt (zh, &attp, curAttributeSet, use_value)))
+ {
+ logf (LOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d",
+ curAttributeSet, use_value, r);
+ if (r == -1)
+ zh->errCode = 114;
+ else
+ zh->errCode = 121;
+ return -1;
+ }
+ if (zebraExplain_curDatabase (zh->service->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->service->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];
+ }
+ }
+ if (!prefix_len)
+ {
+ zh->errCode = 114;
+ return -1;
+ }
+ term_dict[prefix_len++] = ')';
+ term_dict[prefix_len++] = 1;
+ term_dict[prefix_len++] = reg_type;
+ logf (LOG_DEBUG, "reg_type = %d", term_dict[prefix_len-1]);
+ term_dict[prefix_len] = '\0';
+ if (!numeric_relation (zh, zapt, &termp, term_dict,
+ attributeSet, grep_info, &max_pos, reg_type,
+ term_dst))
+ return 0;
+ }
+ *term_sub = termp;
+ logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx);
+ return 1;
+}
+
+static RSET rpn_search_APT_numeric (ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ const char *termz,
+ oid_value attributeSet,
+ NMEM stream,
+ int reg_type, int complete_flag,
+ const char *rank_type,
+ int num_bases, char **basenames)
+{
+ char term_dst[IT_MAX_WORD+1];
+ const char *termp = termz;
+ RSET rset[60], result;
+ int i, r, rset_no = 0;
+ struct grep_info grep_info;
+
+#ifdef TERM_COUNT
+ grep_info.term_no = 0;
+#endif
+ grep_info.isam_p_size = 0;
+ grep_info.isam_p_buf = NULL;
+ grep_info.zh = zh;
+ grep_info.reg_type = reg_type;
+
+ while (1)
+ {
+ logf (LOG_DEBUG, "APT_numeric termp=%s", termp);
+ grep_info.isam_p_indx = 0;
+ r = numeric_term (zh, zapt, &termp, attributeSet, &grep_info,
+ reg_type, complete_flag, num_bases, basenames,
+ term_dst);
+ if (r < 1)
+ break;
+ logf (LOG_DEBUG, "term: %s", term_dst);
+ rset[rset_no] = rset_trunc (zh, grep_info.isam_p_buf,
+ grep_info.isam_p_indx, term_dst,
+ strlen(term_dst), rank_type);
+ assert (rset[rset_no]);
+ if (++rset_no >= (int) (sizeof(rset)/sizeof(*rset)))
+ break;
+ }
+#ifdef TERM_COUNT
+ xfree(grep_info.term_no);
+#endif
+ xfree (grep_info.isam_p_buf);
+ if (rset_no == 0)
+ {
+ rset_null_parms parms;
+
+ parms.rset_term = rset_term_create (term_dst, -1, rank_type);
+ return rset_create (rset_kind_null, &parms);
+ }
+ result = rset[0];
+ for (i = 1; 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 RSET rpn_search_APT_local (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
const char *termz,
oid_value attributeSet,
+ NMEM stream,
const char *rank_type)
{
RSET result;
parms.rset_term = rset_term_create (termz, -1, rank_type);
parms.key_size = sizeof (struct it_key);
- parms.temp_path = res_get (zh->res, "setTmpDir");
+ parms.temp_path = res_get (zh->service->res, "setTmpDir");
result = rset_create (rset_kind_temp, &parms);
rsfd = rset_open (result, RSETF_WRITE);
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;
+
+ 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;
+
+ oe.proto = PROTO_Z3950;
+ oe.oclass = CLASS_ATTSET;
+ oe.value = attributeSet;
+ if (!oid_ent_to_oid (&oe, oid))
+ return 0;
+
+ sks = (Z_SortKeySpec *) nmem_malloc (stream, sizeof(*sks));
+ sks->sortElement = (Z_SortElement *)
+ nmem_malloc (stream, sizeof(*sks->sortElement));
+ sks->sortElement->which = Z_SortElement_generic;
+ sk = sks->sortElement->u.generic = (Z_SortKey *)
+ nmem_malloc (stream, sizeof(*sk));
+ sk->which = Z_SortKey_sortAttributes;
+ sk->u.sortAttributes = (Z_SortAttributes *)
+ nmem_malloc (stream, sizeof(*sk->u.sortAttributes));
+
+ sk->u.sortAttributes->id = oid;
+ sk->u.sortAttributes->list = (Z_AttributeList *)
+ nmem_malloc (stream, sizeof(*sk->u.sortAttributes->list));
+ sk->u.sortAttributes->list->num_attributes = 1;
+ sk->u.sortAttributes->list->attributes = (Z_AttributeElement **)
+ nmem_malloc (stream, sizeof(*sk->u.sortAttributes->list->attributes));
+ ae = *sk->u.sortAttributes->list->attributes = (Z_AttributeElement *)
+ nmem_malloc (stream, sizeof(**sk->u.sortAttributes->list->attributes));
+ ae->attributeSet = 0;
+ ae->attributeType = (int *)
+ nmem_malloc (stream, sizeof(*ae->attributeType));
+ *ae->attributeType = 1;
+ ae->which = Z_AttributeValue_numeric;
+ ae->value.numeric = (int *)
+ nmem_malloc (stream, sizeof(*ae->value.numeric));
+ *ae->value.numeric = use_value;
+
+ sks->sortRelation = (int *)
+ nmem_malloc (stream, sizeof(*sks->sortRelation));
+ if (sort_relation_value == 1)
+ *sks->sortRelation = Z_SortRelation_ascending;
+ else if (sort_relation_value == 2)
+ *sks->sortRelation = Z_SortRelation_descending;
+ else
+ *sks->sortRelation = Z_SortRelation_ascending;
+
+ sks->caseSensitivity = (int *)
+ nmem_malloc (stream, sizeof(*sks->caseSensitivity));
+ *sks->caseSensitivity = 0;
+
+#ifdef ASN_COMPILED
+ sks->which = Z_SortKeySpec_null;
+ sks->u.null = odr_nullval ();
+#else
+ sks->missingValueAction = 0;
+#endif
+
+ sort_sequence->specs[i] = sks;
+
+ parms.rset_term = rset_term_create ("", -1, rank_type);
+ return rset_create (rset_kind_null, &parms);
+}
+
+
static RSET rpn_search_APT (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
- oid_value attributeSet,
+ 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;
int complete_flag;
+ int sort_flag;
char termz[IT_MAX_WORD+1];
- zebra_maps_attr (zh->zebra_maps, zapt, ®_id, &search_type,
- &rank_type, &complete_flag);
+ zebra_maps_attr (zh->service->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);
}
trans_term (zh, zapt, termz);
+ if (sort_flag)
+ return rpn_sort_spec (zh, zapt, attributeSet, stream, sort_sequence,
+ rank_type);
+
if (!strcmp (search_type, "phrase"))
{
- return rpn_search_APT_phrase (zh, zapt, termz, attributeSet,
+ return rpn_search_APT_phrase (zh, zapt, termz, attributeSet, stream,
reg_id, complete_flag, rank_type,
num_bases, basenames);
}
else if (!strcmp (search_type, "and-list"))
{
- return rpn_search_APT_and_list (zh, zapt, termz, attributeSet,
+ return rpn_search_APT_and_list (zh, zapt, termz, attributeSet, stream,
reg_id, complete_flag, rank_type,
num_bases, basenames);
}
else if (!strcmp (search_type, "or-list"))
{
- return rpn_search_APT_or_list (zh, zapt, termz, attributeSet,
+ return rpn_search_APT_or_list (zh, zapt, termz, attributeSet, stream,
reg_id, complete_flag, rank_type,
num_bases, basenames);
}
else if (!strcmp (search_type, "local"))
{
- return rpn_search_APT_local (zh, zapt, termz, attributeSet,
+ return rpn_search_APT_local (zh, zapt, termz, attributeSet, stream,
rank_type);
}
+ else if (!strcmp (search_type, "numeric"))
+ {
+ return rpn_search_APT_numeric (zh, zapt, termz, attributeSet, stream,
+ reg_id, complete_flag, rank_type,
+ num_bases, basenames);
+ }
zh->errCode = 118;
return NULL;
}
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)
{
#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);
#else
if (*zop->u.prox->proximityUnitCode != 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->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)
{
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);
- *dst = odr_malloc (stream, strlen(term_dst)+1);
+ *dst = (char *) nmem_malloc (stream, strlen(term_dst)+1);
strcpy (*dst, term_dst);
}
char *search_type = NULL;
char *rank_type = NULL;
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);
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->service->zebra_maps, zapt, ®_id, &search_type,
+ &rank_type, &complete_flag, &sort_flag))
{
zh->errCode = 113;
return ;
zh->errCode = 114;
else
zh->errCode = 121;
+ *num_entries = 0;
+ return;
}
- if (zebraExplain_curDatabase (zh->zei, basenames[base_no]))
+ if (zebraExplain_curDatabase (zh->service->zei, basenames[base_no]))
{
zh->errString = basenames[base_no];
zh->errCode = 109; /* Database unavailable */
{
int ord;
- ord = zebraExplain_lookupSU (zh->zei, attp.attset_ordinal,
+ ord = zebraExplain_lookupSU (zh->service->zei, attp.attset_ordinal,
local_attr->local);
if (ord > 0)
ords[ord_no++] = ord;
/* 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;
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;
trans_scan_term (zh, zapt, termz+prefix_len, reg_id);
- dict_scan (zh->dict, termz, &before_tmp, &after_tmp, scan_info,
- scan_handle);
+ dict_scan (zh->service->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++)
}
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),
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
projects / idzebra-moved-to-github.git / blobdiff