-/* $Id: zrpn.c,v 1.220 2006-06-22 23:06:06 adam Exp $
+/* $Id: zrpn.c,v 1.229 2006-09-08 18:24:53 adam Exp $
Copyright (C) 1995-2006
Index Data ApS
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.
+along with this program; if not, write to the Free Software
+Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+
*/
#include <stdio.h>
static int log_level_set = 0;
static int log_level_rpn = 0;
+#define TERMSET_DISABLE 1
+
static const char **rpn_char_map_handler(void *vp, const char **from, int len)
{
struct rpn_char_map_info *p = (struct rpn_char_map_info *) vp;
assert(*info == sizeof(*p->isam_p_buf));
memcpy(p->isam_p_buf + p->isam_p_indx, info+1, sizeof(*p->isam_p_buf));
-#if 1
if (p->termset)
{
const char *db;
const char *index_name;
int len = key_SU_decode (&ord, (const unsigned char *) name);
- zebra_term_untrans (p->zh, p->reg_type, term_tmp, name+len+1);
+ zebra_term_untrans (p->zh, p->reg_type, term_tmp, name+len);
yaz_log(log_level_rpn, "grep: %d %c %s", ord, name[len], term_tmp);
zebraExplain_lookup_ord(p->zh->reg->zei,
ord, 0 /* index_type */, &db, &index_name);
resultSetAddTerm(p->zh, p->termset, name[len], db,
index_name, term_tmp);
}
-#endif
(p->isam_p_indx)++;
}
break;
case 3:
case 102:
- case 103:
case -1:
if (!**term_sub)
return 1;
strcat(term_tmp, term_component);
strcat(term_tmp, ")");
break;
+ case 103:
+ yaz_log(log_level_rpn, "Relation always matches");
+ /* skip to end of term (we don't care what it is) */
+ while (**term_sub != '\0')
+ (*term_sub)++;
+ break;
default:
*error_code = YAZ_BIB1_UNSUPP_RELATION_ATTRIBUTE;
return 0;
AttrType hits_limit_attr;
int term_ref_id_int;
- attr_init_APT(&hits_limit_attr, zapt, 9);
+ attr_init_APT(&hits_limit_attr, zapt, 11);
*hits_limit_value = attr_find(&hits_limit_attr, NULL);
attr_init_APT(&term_ref_id_attr, zapt, 10);
const char *term_ref_id_str = 0;
*rset = 0;
- term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str,
- stream);
+ term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str, stream);
grep_info->isam_p_indx = 0;
res = string_term(zh, zapt, term_sub, attributeSet, stream, grep_info,
reg_type, complete_flag, num_bases, basenames,
char ord_buf[32];
int ord_len, i;
- termp = *term_sub;
+ termp = *term_sub; /* start of term for each database */
if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no]))
{
return ZEBRA_FAIL;
}
- if (zebra_apt_get_ord(zh, zapt, zinfo_index_category_index,
- reg_type, xpath_use,
+ if (zebra_apt_get_ord(zh, zapt, reg_type, xpath_use,
attributeSet, &ord) != ZEBRA_OK)
continue;
+ bases_ok++;
+
*ol = ord_list_append(stream, *ol, ord);
-
- if (prefix_len)
- term_dict[prefix_len++] = '|';
- else
- term_dict[prefix_len++] = '(';
-
ord_len = key_SU_encode (ord, ord_buf);
+
+ term_dict[prefix_len++] = '(';
for (i = 0; i<ord_len; i++)
{
- term_dict[prefix_len++] = 1;
+ term_dict[prefix_len++] = 1; /* our internal regexp escape char */
term_dict[prefix_len++] = ord_buf[i];
}
-
- bases_ok++;
-
term_dict[prefix_len++] = ')';
term_dict[prefix_len] = '\0';
j = prefix_len;
{
case -1: /* not specified */
case 100: /* do not truncate */
- if (!string_relation (zh, zapt, &termp, term_dict,
- attributeSet,
- reg_type, space_split, term_dst,
- &relation_error))
+ if (!string_relation(zh, zapt, &termp, term_dict,
+ attributeSet,
+ reg_type, space_split, term_dst,
+ &relation_error))
{
if (relation_error)
{
grep_info->zh = zh;
grep_info->reg_type = reg_type;
grep_info->termset = 0;
-
if (!zapt)
return ZEBRA_OK;
attr_init_APT(&termset, zapt, 8);
attr_find_ex(&termset, NULL, &termset_value_string);
if (termset_value_numeric != -1)
{
+#if TERMSET_DISABLE
+ zebra_setError(zh, YAZ_BIB1_UNSUPP_SEARCH, "termset");
+ return ZEBRA_FAIL;
+#else
char resname[32];
const char *termset_name = 0;
if (termset_value_numeric != -2)
zebra_setError(zh, YAZ_BIB1_ILLEGAL_RESULT_SET_NAME, termset_name);
return ZEBRA_FAIL;
}
+#endif
}
return ZEBRA_OK;
}
/**
\brief Create result set(s) for list of terms
\param zh Zebra Handle
+ \param zapt Attributes Plust Term (RPN leaf)
\param termz term as used in query but converted to UTF-8
\param attributeSet default attribute set
\param stream memory for result
\param xpath_use use attribute for X-Path (-1 for no X-path)
\param num_bases number of databases
\param basenames array of databases
- \param rset_mem memory for result sets
+ \param rset_nmem memory for result sets
\param result_sets output result set for each term in list (output)
- \param number number of output result sets
+ \param num_result_sets number of output result sets
\param kc rset key control to be used for created result sets
*/
static ZEBRA_RES term_list_trunc(ZebraHandle zh,
struct grep_info grep_info;
const char *termp = termz;
int alloc_sets = 0;
- int empty_term = *termz ? 0 : 1;
- empty_term = 0;
*num_result_sets = 0;
*term_dst = 0;
if (grep_info_prepare(zh, zapt, &grep_info, reg_type) == ZEBRA_FAIL)
break;
(*num_result_sets)++;
- if (empty_term)
- break;
if (!*termp)
break;
}
return ZEBRA_OK;
}
-
-static ZEBRA_RES always_term(ZebraHandle zh, Z_AttributesPlusTerm *zapt,
- oid_value attributeSet, NMEM stream,
- struct grep_info *grep_info,
- int reg_type, int complete_flag,
- int num_bases, char **basenames,
- const char *xpath_use,
- struct ord_list **ol)
+static ZEBRA_RES rpn_search_APT_position(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ oid_value attributeSet,
+ int reg_type,
+ int num_bases, char **basenames,
+ NMEM rset_nmem,
+ RSET *rset,
+ struct rset_key_control *kc)
{
- char term_dict[2*IT_MAX_WORD+4000];
- int r, base_no;
- struct rpn_char_map_info rcmi;
-
- int bases_ok = 0; /* no of databases with OK attribute */
+ RSET *f_set;
+ int base_no;
+ int position_value;
+ int num_sets = 0;
+ AttrType position;
- *ol = ord_list_create(stream);
+ attr_init_APT(&position, zapt, 3);
+ position_value = attr_find(&position, NULL);
+ switch(position_value)
+ {
+ case 3:
+ case -1:
+ return ZEBRA_OK;
+ case 1:
+ case 2:
+ break;
+ default:
+ zebra_setError_zint(zh, YAZ_BIB1_UNSUPP_POSITION_ATTRIBUTE,
+ position_value);
+ return ZEBRA_FAIL;
+ }
- rpn_char_map_prepare (zh->reg, reg_type, &rcmi);
+ if (!zebra_maps_is_first_in_field(zh->reg->zebra_maps, reg_type))
+ {
+ zebra_setError_zint(zh, YAZ_BIB1_UNSUPP_POSITION_ATTRIBUTE,
+ position_value);
+ return ZEBRA_FAIL;
+ }
+ if (!zh->reg->isamb && !zh->reg->isamc)
+ {
+ zebra_setError_zint(zh, YAZ_BIB1_UNSUPP_POSITION_ATTRIBUTE,
+ position_value);
+ return ZEBRA_FAIL;
+ }
+ f_set = xmalloc(sizeof(RSET) * num_bases);
for (base_no = 0; base_no < num_bases; base_no++)
{
int ord = -1;
- int regex_range = 0;
- int max_pos, prefix_len = 0;
char ord_buf[32];
- int ord_len, i;
+ char term_dict[100];
+ int ord_len;
+ char *val;
+ ISAM_P isam_p;
if (zebraExplain_curDatabase (zh->reg->zei, basenames[base_no]))
{
return ZEBRA_FAIL;
}
- if (zebra_apt_get_ord(zh, zapt, zinfo_index_category_alwaysmatches,
- reg_type, xpath_use,
+ if (zebra_apt_get_ord(zh, zapt, reg_type, 0,
attributeSet, &ord) != ZEBRA_OK)
continue;
- *ol = ord_list_append(stream, *ol, ord);
-
- 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];
- }
+ memcpy(term_dict, ord_buf, ord_len);
+ strcpy(term_dict+ord_len, FIRST_IN_FIELD_STR);
+ val = dict_lookup(zh->reg->dict, term_dict);
+ if (!val)
+ continue;
+ assert(*val == sizeof(ISAM_P));
+ memcpy(&isam_p, val+1, sizeof(isam_p));
- term_dict[prefix_len++] = ')';
- term_dict[prefix_len] = '\0';
- bases_ok++;
-
- r = dict_lookup_grep(zh->reg->dict, term_dict, regex_range,
- grep_info, &max_pos,
- ord_len /* number of "exact" chars */,
- grep_handle);
+ if (zh->reg->isamb)
+ f_set[num_sets++] = rsisamb_create(rset_nmem, kc, kc->scope,
+ zh->reg->isamb, isam_p, 0);
+ else if (zh->reg->isamc)
+ f_set[num_sets++] = rsisamc_create(rset_nmem, kc, kc->scope,
+ zh->reg->isamc, isam_p, 0);
}
- if (!bases_ok)
- return ZEBRA_FAIL;
+ if (num_sets)
+ {
+ *rset = rset_create_or(rset_nmem, kc, kc->scope,
+ 0 /* termid */, num_sets, f_set);
+ }
+ xfree(f_set);
return ZEBRA_OK;
}
-
-static ZEBRA_RES rpn_search_APT_alwaysmatches(ZebraHandle zh,
- Z_AttributesPlusTerm *zapt,
- const char *termz_org,
- oid_value attributeSet,
- NMEM stream,
- int reg_type, int complete_flag,
- const char *rank_type,
- const char *xpath_use,
- int num_bases, char **basenames,
- NMEM rset_nmem,
- RSET *rset,
- struct rset_key_control *kc)
-{
- const char *term_dst = "always";
- struct grep_info grep_info;
- zint hits_limit_value;
- const char *term_ref_id_str = 0;
- ZEBRA_RES res;
- struct ord_list *ol;
-
- term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str,
- stream);
- if (grep_info_prepare(zh, zapt, &grep_info, reg_type) == ZEBRA_FAIL)
- return ZEBRA_FAIL;
-
- grep_info.isam_p_indx = 0;
-
- res = always_term(zh, zapt, attributeSet, stream, &grep_info,
- reg_type, complete_flag, num_bases, basenames,
- xpath_use, &ol);
- if (res == ZEBRA_OK)
- {
- *rset = 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, rset_nmem,
- kc, kc->scope, ol, reg_type, hits_limit_value,
- term_ref_id_str);
- if (!*rset)
- res = ZEBRA_FAIL;
- }
- grep_info_delete (&grep_info);
- return res;
-}
-
+
static ZEBRA_RES rpn_search_APT_phrase(ZebraHandle zh,
Z_AttributesPlusTerm *zapt,
const char *termz_org,
num_bases, basenames,
rset_nmem,
&result_sets, &num_result_sets, kc);
+
if (res != ZEBRA_OK)
return res;
+
+ if (num_result_sets > 0)
+ {
+ RSET first_set = 0;
+ res = rpn_search_APT_position(zh, zapt, attributeSet,
+ reg_type,
+ num_bases, basenames,
+ rset_nmem, &first_set,
+ kc);
+ if (res != ZEBRA_OK)
+ return res;
+ if (first_set)
+ {
+ RSET *nsets = nmem_malloc(stream,
+ sizeof(RSET) * (num_result_sets+1));
+ nsets[0] = first_set;
+ memcpy(nsets+1, result_sets, sizeof(RSET) * num_result_sets);
+ result_sets = nsets;
+ num_result_sets++;
+ }
+ }
if (num_result_sets == 0)
*rset = rset_create_null(rset_nmem, kc, 0);
else if (num_result_sets == 1)
{
RSET *result_sets = 0;
int num_result_sets = 0;
+ int i;
ZEBRA_RES res =
term_list_trunc(zh, zapt, termz_org, attributeSet,
stream, reg_type, complete_flag,
&result_sets, &num_result_sets, kc);
if (res != ZEBRA_OK)
return res;
+
+ for (i = 0; i<num_result_sets; i++)
+ {
+ RSET first_set = 0;
+ res = rpn_search_APT_position(zh, zapt, attributeSet,
+ reg_type,
+ num_bases, basenames,
+ rset_nmem, &first_set,
+ kc);
+ if (res != ZEBRA_OK)
+ {
+ for (i = 0; i<num_result_sets; i++)
+ rset_delete(result_sets[i]);
+ return res;
+ }
+
+ if (first_set)
+ {
+ RSET tmp_set[2];
+
+ tmp_set[0] = first_set;
+ tmp_set[1] = result_sets[i];
+
+ result_sets[i] = rset_create_prox(
+ rset_nmem, kc, kc->scope,
+ 2, tmp_set,
+ 1 /* ordered */, 0 /* exclusion */,
+ 3 /* relation */, 1 /* distance */);
+ }
+ }
if (num_result_sets == 0)
*rset = rset_create_null(rset_nmem, kc, 0);
else if (num_result_sets == 1)
{
RSET *result_sets = 0;
int num_result_sets = 0;
+ int i;
ZEBRA_RES res =
term_list_trunc(zh, zapt, termz_org, attributeSet,
stream, reg_type, complete_flag,
kc);
if (res != ZEBRA_OK)
return res;
+ for (i = 0; i<num_result_sets; i++)
+ {
+ RSET first_set = 0;
+ res = rpn_search_APT_position(zh, zapt, attributeSet,
+ reg_type,
+ num_bases, basenames,
+ rset_nmem, &first_set,
+ kc);
+ if (res != ZEBRA_OK)
+ {
+ for (i = 0; i<num_result_sets; i++)
+ rset_delete(result_sets[i]);
+ return res;
+ }
+
+ if (first_set)
+ {
+ RSET tmp_set[2];
+
+ tmp_set[0] = first_set;
+ tmp_set[1] = result_sets[i];
+
+ result_sets[i] = rset_create_prox(
+ rset_nmem, kc, kc->scope,
+ 2, tmp_set,
+ 1 /* ordered */, 0 /* exclusion */,
+ 3 /* relation */, 1 /* distance */);
+ }
+ }
+
+
if (num_result_sets == 0)
*rset = rset_create_null(rset_nmem, kc, 0);
else if (num_result_sets == 1)
yaz_log(log_level_rpn, "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:
yaz_log(log_level_rpn, "Relation <");
+ if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1,
+ term_dst))
+ return 0;
+ term_value = atoi (term_tmp);
gen_regular_rel(term_tmp, term_value-1, 1);
break;
case 2:
yaz_log(log_level_rpn, "Relation <=");
+ if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1,
+ term_dst))
+ return 0;
+ term_value = atoi (term_tmp);
gen_regular_rel(term_tmp, term_value, 1);
break;
case 4:
yaz_log(log_level_rpn, "Relation >=");
+ if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1,
+ term_dst))
+ return 0;
+ term_value = atoi (term_tmp);
gen_regular_rel(term_tmp, term_value, 0);
break;
case 5:
yaz_log(log_level_rpn, "Relation >");
+ if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1,
+ term_dst))
+ return 0;
+ term_value = atoi (term_tmp);
gen_regular_rel(term_tmp, term_value+1, 0);
break;
case -1:
case 3:
yaz_log(log_level_rpn, "Relation =");
+ if (!term_100(zh->reg->zebra_maps, reg_type, term_sub, term_tmp, 1,
+ term_dst))
+ return 0;
+ term_value = atoi (term_tmp);
sprintf(term_tmp, "(0*%d)", term_value);
break;
+ case 103:
+ /* term_tmp untouched.. */
+ while (**term_sub != '\0')
+ (*term_sub)++;
+ break;
default:
*error_code = YAZ_BIB1_UNSUPP_RELATION_ATTRIBUTE;
return 0;
return ZEBRA_FAIL;
}
- if (zebra_apt_get_ord(zh, zapt, zinfo_index_category_index,
- reg_type, xpath_use, attributeSet, &ord)
- != ZEBRA_OK)
+ if (zebra_apt_get_ord(zh, zapt, reg_type, xpath_use,
+ attributeSet, &ord) != ZEBRA_OK)
continue;
+ bases_ok++;
*ol = ord_list_append(stream, *ol, ord);
- if (prefix_len)
- term_dict[prefix_len++] = '|';
- else
- term_dict[prefix_len++] = '(';
-
ord_len = key_SU_encode (ord, ord_buf);
+
+ term_dict[prefix_len++] = '(';
for (i = 0; i < ord_len; i++)
{
term_dict[prefix_len++] = 1;
}
term_dict[prefix_len++] = ')';
term_dict[prefix_len] = '\0';
- bases_ok++;
+
if (!numeric_relation(zh, zapt, &termp, term_dict,
attributeSet, grep_info, &max_pos, reg_type,
term_dst, &relation_error))
if (!result_sets[num_result_sets])
break;
num_result_sets++;
+ if (!*termp)
+ break;
}
grep_info_delete(&grep_info);
- if (termp)
- {
- int i;
- for (i = 0; i<num_result_sets; i++)
- rset_delete(result_sets[i]);
- return ZEBRA_FAIL;
- }
+
+ if (res != ZEBRA_OK)
+ return res;
if (num_result_sets == 0)
*rset = rset_create_null(rset_nmem, kc, 0);
- if (num_result_sets == 1)
+ else if (num_result_sets == 1)
*rset = result_sets[0];
else
- *rset = rset_create_and(rset_nmem, kc, kc->scope,
+ *rset = rset_create_and(rset_nmem, kc, kc->scope,
num_result_sets, result_sets);
if (!*rset)
- return ZEBRA_FAIL;
+ return ZEBRA_FAIL;
return ZEBRA_OK;
}
xpath_use = ZEBRA_XPATH_ATTR_CDATA; /* last step is attribute */
else
xpath_use = ZEBRA_XPATH_CDATA; /* searching for cdata */
+
+ if (1)
+ {
+ AttrType relation;
+ int relation_value;
+
+ attr_init_APT(&relation, zapt, 2);
+ relation_value = attr_find(&relation, NULL);
+
+ if (relation_value == 103) /* alwaysmatches */
+ {
+ *rset = 0; /* signal no "term" set */
+ return rpn_search_xpath(zh, num_bases, basenames,
+ stream, rank_type, *rset,
+ xpath_len, xpath, rset_nmem, rset, kc);
+ }
+ }
}
/* search using one of the various search type strategies
num_bases, basenames, rset_nmem,
rset, kc);
}
- else if (!strcmp(search_type, "always"))
- {
- if (xpath_len >= 0) /* alwaysmatches and X-Path ? */
- {
- *rset = 0; /* signal no "term" set */
- return rpn_search_xpath(zh, num_bases, basenames,
- stream, rank_type, *rset,
- xpath_len, xpath, rset_nmem, rset, kc);
- }
- else
- {
- res = rpn_search_APT_alwaysmatches(zh, zapt, termz,
- attributeSet, stream,
- reg_id, complete_flag,
- rank_type,
- xpath_use,
- num_bases, basenames, rset_nmem,
- rset, kc);
- }
- }
else
{
zebra_setError(zh, YAZ_BIB1_UNSUPP_SEARCH, 0);
else
idx = - pos - 1;
+ /* skip special terms.. of no interest */
+ if (name[len_prefix] < 4)
+ return 1;
+
if (idx < 0)
return 0;
scan_info->list[idx].term = (char *)
*num_entries = 0;
return ZEBRA_FAIL;
}
- if (zebra_apt_get_ord(zh, zapt, zinfo_index_category_index,
- index_type, 0, attributeset, &ord)
+ if (zebra_apt_get_ord(zh, zapt, index_type, 0, attributeset, &ord)
!= ZEBRA_OK)
continue;
ords[ord_no++] = ord;
if (ord_no == 0)
{
*num_entries = 0;
- return ZEBRA_OK;
+ return ZEBRA_FAIL;
}
/* prepare dictionary scanning */
if (num < 1)