-/* $Id: zrpn.c,v 1.192 2005-05-24 11:35:42 adam Exp $
+/* $Id: zrpn.c,v 1.209 2006-02-20 12:38:42 adam Exp $
Copyright (C) 1995-2005
Index Data ApS
#include <assert.h>
#ifdef WIN32
#include <io.h>
-#else
+#endif
+#if HAVE_UNISTD_H
#include <unistd.h>
#endif
#include <ctype.h>
Z_AttributesPlusTerm *zapt;
} AttrType;
-
static int log_level_set = 0;
static int log_level_rpn = 0;
}
static int attr_find_ex(AttrType *src, oid_value *attributeSetP,
- const char **string_value)
+ const char **string_value)
{
int num_attributes;
ZebraSet termset;
};
-static void term_untrans(ZebraHandle zh, int reg_type,
- char *dst, const char *src)
+void zebra_term_untrans(ZebraHandle zh, int reg_type,
+ char *dst, const char *src)
{
int len = 0;
while (*src)
{
const char *cp = zebra_maps_output(zh->reg->zebra_maps,
reg_type, &src);
- if (!cp && len < IT_MAX_WORD-1)
- dst[len++] = *src++;
+ if (!cp)
+ {
+ if (len < IT_MAX_WORD-1)
+ dst[len++] = *src;
+ src++;
+ }
else
while (*cp && len < IT_MAX_WORD-1)
dst[len++] = *cp++;
const char *db;
int set, use;
char term_tmp[IT_MAX_WORD];
- int su_code = 0;
- int len = key_SU_decode (&su_code, name);
+ int ord = 0;
+ int len = key_SU_decode (&ord, (const unsigned char *) name);
- term_untrans (p->zh, p->reg_type, term_tmp, name+len+1);
- yaz_log(log_level_rpn, "grep: %d %c %s", su_code, name[len], term_tmp);
+ zebra_term_untrans (p->zh, p->reg_type, term_tmp, name+len+1);
+ yaz_log(log_level_rpn, "grep: %d %c %s", ord, name[len], term_tmp);
zebraExplain_lookup_ord (p->zh->reg->zei,
- su_code, &db, &set, &use);
+ ord, 0 /* index_type */, &db, &set, &use);
yaz_log(log_level_rpn, "grep: set=%d use=%d db=%s", set, use, db);
resultSetAddTerm(p->zh, p->termset, name[len], db,
struct grep_info *grep_info,
int reg_type, int complete_flag,
int num_bases, char **basenames,
- char *term_dst, int xpath_use);
+ char *term_dst, int xpath_use,
+ struct ord_list **ol);
+
+static ZEBRA_RES term_limits_APT(ZebraHandle zh,
+ Z_AttributesPlusTerm *zapt,
+ zint *hits_limit_value,
+ const char **term_ref_id_str,
+ NMEM nmem)
+{
+ AttrType term_ref_id_attr;
+ AttrType hits_limit_attr;
+ int term_ref_id_int;
+
+ attr_init(&hits_limit_attr, zapt, 9);
+ *hits_limit_value = attr_find(&hits_limit_attr, NULL);
+
+ attr_init(&term_ref_id_attr, zapt, 10);
+ term_ref_id_int = attr_find_ex(&term_ref_id_attr, NULL, term_ref_id_str);
+ if (term_ref_id_int >= 0)
+ {
+ char *res = nmem_malloc(nmem, 20);
+ sprintf(res, "%d", term_ref_id_int);
+ *term_ref_id_str = res;
+ }
+
+ /* no limit given ? */
+ if (*hits_limit_value == -1)
+ {
+ if (*term_ref_id_str)
+ {
+ /* use global if term_ref is present */
+ *hits_limit_value = zh->approx_limit;
+ }
+ else
+ {
+ /* no counting if term_ref is not present */
+ *hits_limit_value = 0;
+ }
+ }
+ else if (*hits_limit_value == 0)
+ {
+ /* 0 is the same as global limit */
+ *hits_limit_value = zh->approx_limit;
+ }
+ yaz_log(YLOG_DEBUG, "term_limits_APT ref_id=%s limit=" ZINT_FORMAT,
+ *term_ref_id_str ? *term_ref_id_str : "none",
+ *hits_limit_value);
+ return ZEBRA_OK;
+}
static ZEBRA_RES term_trunc(ZebraHandle zh,
Z_AttributesPlusTerm *zapt,
struct rset_key_control *kc)
{
ZEBRA_RES res;
+ struct ord_list *ol;
+ zint hits_limit_value;
+ const char *term_ref_id_str = 0;
*rset = 0;
+
+ 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,
- term_dst, xpath_use);
+ term_dst, xpath_use, &ol);
if (res != ZEBRA_OK)
return res;
if (!*term_sub) /* no more terms ? */
grep_info->isam_p_indx, term_dst,
strlen(term_dst), rank_type, 1 /* preserve pos */,
zapt->term->which, rset_nmem,
- kc, kc->scope);
+ kc, kc->scope, ol, reg_type, hits_limit_value,
+ term_ref_id_str);
if (!*rset)
return ZEBRA_FAIL;
return ZEBRA_OK;
}
-static char *nmem_strdup_i(NMEM nmem, int v)
-{
- char val_str[64];
- sprintf(val_str, "%d", v);
- return nmem_strdup(nmem, val_str);
-}
-
static ZEBRA_RES 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_dst, int xpath_use,
+ struct ord_list **ol)
{
char term_dict[2*IT_MAX_WORD+4000];
int j, r, base_no;
int space_split = complete_flag ? 0 : 1;
int bases_ok = 0; /* no of databases with OK attribute */
- int errCode = 0; /* err code (if any is not OK) */
- char *errString = 0; /* addinfo */
+
+ *ol = ord_list_create(stream);
rpn_char_map_prepare (zh->reg, reg_type, &rcmi);
attr_init(&use, zapt, 1);
}
else if (use_string &&
(ord = zebraExplain_lookup_attr_str(zh->reg->zei,
+ reg_type,
use_string)) >= 0)
{
/* we have a match for a raw string attribute */
term_dict[prefix_len++] = ord_buf[i];
}
attp.local_attributes = 0; /* no more attributes */
+ *ol = ord_list_append(stream, *ol, ord);
}
else
{
if (r == -1)
{
/* set was found, but value wasn't defined */
- errCode = YAZ_BIB1_UNSUPP_USE_ATTRIBUTE;
if (use_string)
- errString = nmem_strdup(stream, use_string);
+ zebra_setError(zh,
+ YAZ_BIB1_UNSUPP_USE_ATTRIBUTE,
+ use_string);
else
- errString = nmem_strdup_i (stream, use_value);
+ zebra_setError_zint(zh, YAZ_BIB1_UNSUPP_USE_ATTRIBUTE,
+ use_value);
}
else
{
oident.value = curAttributeSet;
oid_ent_to_oid (&oident, oid);
- errCode = YAZ_BIB1_UNSUPP_ATTRIBUTE_SET;
- errString = nmem_strdup(stream, oident.desc);
+ zebra_setError(zh,
+ YAZ_BIB1_UNSUPP_ATTRIBUTE_SET,
+ oident.desc);
+
}
continue;
}
int i, ord_len;
ord = zebraExplain_lookup_attr_su(zh->reg->zei,
+ reg_type,
attp.attset_ordinal,
local_attr->local);
if (ord < 0)
continue;
+ *ol = ord_list_append(stream, *ol, ord);
if (prefix_len)
term_dict[prefix_len++] = '|';
else
attr_ok = 1;
term_dict[prefix_len++] = ')';
- term_dict[prefix_len++] = 1;
- term_dict[prefix_len++] = reg_type;
- yaz_log(log_level_rpn, "reg_type = %d", term_dict[prefix_len-1]);
term_dict[prefix_len] = '\0';
j = prefix_len;
switch (truncation_value)
}
}
if (!bases_ok)
- {
- zebra_setError(zh, errCode, errString);
return ZEBRA_FAIL;
- }
*term_sub = termp;
yaz_log(YLOG_DEBUG, "%d positions", grep_info->isam_p_indx);
return ZEBRA_OK;
case Z_Term_general:
if (zh->iconv_to_utf8 != 0)
{
- char *inbuf = term->u.general->buf;
+ char *inbuf = (char *) term->u.general->buf;
size_t inleft = term->u.general->len;
char *outbuf = termz;
size_t outleft = IT_MAX_WORD-1;
return ZEBRA_OK;
}
-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
return ZEBRA_OK;
}
-
+/**
+ \brief Create result set(s) for list of terms
+ \param zh Zebra Handle
+ \param termz term as used in query but converted to UTF-8
+ \param attributeSet default attribute set
+ \param stream memory for result
+ \param reg_type register type ('w', 'p',..)
+ \param complete_flag whether it's phrases or not
+ \param rank_type term flags for ranking
+ \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 result_sets output result set for each term in list (output)
+ \param number 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,
Z_AttributesPlusTerm *zapt,
- const char *termz_org,
+ const char *termz,
oid_value attributeSet,
NMEM stream,
int reg_type, int complete_flag,
{
char term_dst[IT_MAX_WORD+1];
struct grep_info grep_info;
- char *termz = normalize_term(zh, zapt, termz_org, stream, reg_type);
const char *termp = termz;
int alloc_sets = 0;
struct rpn_char_map_info rcmi;
int bases_ok = 0; /* no of databases with OK attribute */
- int errCode = 0; /* err code (if any is not OK) */
- char *errString = 0; /* addinfo */
rpn_char_map_prepare (zh->reg, reg_type, &rcmi);
attr_init(&use, zapt, 1);
else
{
if ((r = att_getentbyatt (zh, &attp, curAttributeSet, use_value,
- use_string)))
+ use_string)))
{
yaz_log(YLOG_DEBUG, "att_getentbyatt fail. set=%d use=%d r=%d",
curAttributeSet, use_value, r);
if (r == -1)
{
- errCode = YAZ_BIB1_UNSUPP_USE_ATTRIBUTE;
if (use_string)
- errString = nmem_strdup(stream, use_string);
+ zebra_setError(zh,
+ YAZ_BIB1_UNSUPP_USE_ATTRIBUTE,
+ use_string);
else
- errString = nmem_strdup_i (stream, use_value);
+ zebra_setError_zint(zh, YAZ_BIB1_UNSUPP_USE_ATTRIBUTE,
+ use_value);
}
else
- errCode = YAZ_BIB1_UNSUPP_ATTRIBUTE_SET;
+ zebra_setError(zh, YAZ_BIB1_UNSUPP_ATTRIBUTE_SET, 0);
continue;
}
}
int i, ord_len;
ord = zebraExplain_lookup_attr_su(zh->reg->zei,
+ reg_type,
attp.attset_ordinal,
local_attr->local);
if (ord < 0)
continue;
}
bases_ok++;
- term_dict[prefix_len++] = ')';
- term_dict[prefix_len++] = 1;
- term_dict[prefix_len++] = reg_type;
- yaz_log(YLOG_DEBUG, "reg_type = %d", term_dict[prefix_len-1]);
+ term_dict[prefix_len++] = ')';
term_dict[prefix_len] = '\0';
if (!numeric_relation(zh, zapt, &termp, term_dict,
attributeSet, grep_info, &max_pos, reg_type,
}
}
if (!bases_ok)
- {
- zebra_setError(zh, errCode, errString);
return ZEBRA_FAIL;
- }
*term_sub = termp;
yaz_log(YLOG_DEBUG, "%d positions", grep_info->isam_p_indx);
return ZEBRA_OK;
}
+
static ZEBRA_RES rpn_search_APT_numeric(ZebraHandle zh,
Z_AttributesPlusTerm *zapt,
const char *termz,
ZEBRA_RES res;
struct grep_info grep_info;
int alloc_sets = 0;
+ zint hits_limit_value;
+ const char *term_ref_id_str = 0;
+
+ term_limits_APT(zh, zapt, &hits_limit_value, &term_ref_id_str, stream);
yaz_log(log_level_rpn, "APT_numeric t='%s'", termz);
if (grep_info_prepare(zh, zapt, &grep_info, reg_type) == ZEBRA_FAIL)
strlen(term_dst), rank_type,
0 /* preserve position */,
zapt->term->which, rset_nmem,
- kc, kc->scope);
+ kc, kc->scope, 0, reg_type,
+ hits_limit_value,
+ term_ref_id_str);
if (!result_sets[num_result_sets])
break;
num_result_sets++;
char term_dict[2048];
char ord_buf[32];
int prefix_len = 0;
- int ord = zebraExplain_lookup_attr_su(zh->reg->zei, curAttributeSet, use);
+ int ord = zebraExplain_lookup_attr_su(zh->reg->zei, reg_type,
+ curAttributeSet, use);
int ord_len, i, r, max_pos;
int term_type = Z_Term_characterString;
const char *flags = "void";
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;
rset = rset_trunc(zh, grep_info.isam_p_buf,
grep_info.isam_p_indx, term, strlen(term),
flags, 1, term_type,rset_nmem,
- kc, kc->scope);
+ kc, kc->scope, 0, reg_type, 0 /* hits_limit */,
+ 0 /* term_ref_id_str */);
grep_info_delete(&grep_info);
return rset;
}
if (sort_flag)
return rpn_sort_spec(zh, zapt, attributeSet, stream, sort_sequence,
rank_type, rset_nmem, rset, kc);
+ /* consider if an X-Path query is used */
xpath_len = parse_xpath(zh, zapt, attributeSet, xpath, 10, stream);
if (xpath_len >= 0)
{
- xpath_use = 1016;
- if (xpath[xpath_len-1].part[0] == '@')
- xpath_use = 1015;
- }
-
+ xpath_use = 1016; /* searching for element by default */
+ if (xpath[xpath_len-1].part[0] == '@')
+ xpath_use = 1015; /* last step an attribute .. */
+ }
+
+ /* search using one of the various search type strategies
+ termz is our UTF-8 search term
+ attributeSet is top-level default attribute set
+ stream is ODR for search
+ reg_id is the register type
+ complete_flag is 1 for complete subfield, 0 for incomplete
+ xpath_use is use-attribute to be used for X-Path search, 0 for none
+ */
if (!strcmp(search_type, "phrase"))
{
res = rpn_search_APT_phrase(zh, zapt, termz, attributeSet, stream,
return 0;
}
-static void scan_term_untrans (ZebraHandle zh, NMEM stream, int reg_type,
- char **dst, const char *src)
+void zebra_term_untrans_iconv(ZebraHandle zh, NMEM stream, int reg_type,
+ char **dst, const char *src)
{
char term_src[IT_MAX_WORD];
char term_dst[IT_MAX_WORD];
- term_untrans (zh, reg_type, term_src, src);
+ zebra_term_untrans (zh, reg_type, term_src, src);
if (zh->iconv_from_utf8 != 0)
{
*dst = nmem_strdup(stream, term_src);
}
-static void count_set (RSET r, int *count)
+static void count_set(ZebraHandle zh, RSET rset, zint *count)
{
zint psysno = 0;
- int kno = 0;
struct it_key key;
RSFD rfd;
yaz_log(YLOG_DEBUG, "count_set");
+ rset->hits_limit = zh->approx_limit;
+
*count = 0;
- rfd = rset_open (r, RSETF_READ);
- while (rset_read (rfd, &key,0 /* never mind terms */))
+ rfd = rset_open(rset, RSETF_READ);
+ while (rset_read(rfd, &key,0 /* never mind terms */))
{
if (key.mem[0] != psysno)
{
psysno = key.mem[0];
- (*count)++;
+ if (rfd->counted_items >= rset->hits_limit)
+ break;
}
- kno++;
}
rset_close (rfd);
- yaz_log(YLOG_DEBUG, "%d keys, %d records", kno, *count);
+ *count = rset->hits_count;
}
ZEBRA_RES rpn_scan(ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt,
}
if (use_string &&
- (ord = zebraExplain_lookup_attr_str(zh->reg->zei,
+ (ord = zebraExplain_lookup_attr_str(zh->reg->zei, reg_id,
use_string)) >= 0)
{
/* we have a match for a raw string attribute */
for (local_attr = attp.local_attributes; local_attr && ord_no < 32;
local_attr = local_attr->next)
{
- ord = zebraExplain_lookup_attr_su(zh->reg->zei,
+ ord = zebraExplain_lookup_attr_su(zh->reg->zei, reg_id,
attp.attset_ordinal,
local_attr->local);
if (ord > 0)
scan_info->list[j].term = NULL;
prefix_len += key_SU_encode (ords[i], termz + prefix_len);
- termz[prefix_len++] = reg_id;
termz[prefix_len] = 0;
strcpy(scan_info->prefix, termz);
if (lo >= 0)
{
/* get result set for first term */
- scan_term_untrans(zh, stream->mem, reg_id,
- &glist[lo].term, mterm);
+ zebra_term_untrans_iconv(zh, stream->mem, reg_id,
+ &glist[lo].term, mterm);
rset = rset_trunc(zh, &scan_info_array[j0].list[ptr[j0]].isam_p, 1,
glist[lo].term, strlen(glist[lo].term),
NULL, 0, zapt->term->which, rset_nmem,
- kc, kc->scope);
+ kc, kc->scope, 0, reg_id, 0 /* hits_limit */,
+ 0 /* term_ref_id_str */);
}
ptr[j0]++; /* move index for this set .. */
/* get result set for remaining scan terms */
glist[lo].term,
strlen(glist[lo].term), NULL, 0,
zapt->term->which,rset_nmem,
- kc, kc->scope);
+ kc, kc->scope, 0, reg_id, 0 /* hits_limit */,
+ 0 /* term_ref_id_str */ );
rset = rsmulti_or_create(rset_nmem, kc,
kc->scope, 0 /* termid */,
2, rsets);
}
if (lo >= 0)
{
+ zint count;
/* merge with limit_set if given */
if (limit_set)
{
2, rsets);
}
/* count it */
- count_set(rset, &glist[lo].occurrences);
+ count_set(zh, rset, &count);
+ glist[lo].occurrences = count;
rset_delete(rset);
}
}
const char *tst;
RSET rset;
int lo = before-1-i; /* offset in result list */
+ zint count;
for (j = 0; j <ord_no; j++)
{
if (j0 == -1)
break;
- scan_term_untrans (zh, stream->mem, reg_id,
- &glist[lo].term, mterm);
+ zebra_term_untrans_iconv(zh, stream->mem, reg_id,
+ &glist[lo].term, mterm);
rset = rset_trunc
(zh, &scan_info_array[j0].list[before-1-ptr[j0]].isam_p, 1,
glist[lo].term, strlen(glist[lo].term),
- NULL, 0, zapt->term->which,rset_nmem,
- kc, kc->scope);
+ NULL, 0, zapt->term->which, rset_nmem,
+ kc, kc->scope, 0, reg_id, 0 /* hits_limit */,
+ 0 /* term_ref_id_str */);
ptr[j0]++;
glist[lo].term,
strlen(glist[lo].term), NULL, 0,
zapt->term->which, rset_nmem,
- kc, kc->scope);
+ kc, kc->scope, 0, reg_id, 0 /* hits_limit */,
+ 0 /* term_ref_id_str */);
rset = rsmulti_or_create(rset_nmem, kc,
kc->scope, 0 /* termid */, 2, rsets);
rset = rsmulti_and_create(rset_nmem, kc,
kc->scope, 2, rsets);
}
- count_set (rset, &glist[lo].occurrences);
+ count_set(zh, rset, &count);
+ glist[lo].occurrences = count;
rset_delete (rset);
}
(*kc->dec)(kc);