-/* $Id: zsets.c,v 1.87 2005-06-07 11:36:38 adam Exp $
+/* $Id: zsets.c,v 1.101 2006-05-10 08:13:26 adam Exp $
Copyright (C) 1995-2005
Index Data ApS
#endif
#include "index.h"
+#include "rank.h"
#include <yaz/diagbib1.h>
#include <rset.h>
int num_bases;
char **basenames;
Z_RPNQuery *rpn;
+ Z_SortKeySpecList *sortSpec;
struct zset_sort_info *sort_info;
struct zebra_set_term_entry *term_entries;
int term_entries_max;
zint cache_position; /* last position */
RSFD cache_rfd; /* rfd (NULL if not existing) */
zint cache_psysno; /* sysno for last position */
+ zint approx_limit; /* limit before we do approx */
};
struct zset_sort_entry {
for (i = 0; sort_sequence->specs[i]; i++)
;
sort_sequence->num_specs = i;
+ rset->hits_limit = sset->approx_limit;
if (!i)
{
res = resultSetRank (zh, sset, rset, rset_nmem);
zh->hits = 0;
- zebraSet = resultSetAdd (zh, setname, 1);
+ zebraSet = resultSetAdd(zh, setname, 1);
if (!zebraSet)
return ZEBRA_FAIL;
zebraSet->locked = 1;
zebraSet->basenames =
nmem_malloc (zebraSet->nmem, num_bases * sizeof(*zebraSet->basenames));
for (i = 0; i<num_bases; i++)
- zebraSet->basenames[i] = nmem_strdup (zebraSet->nmem, basenames[i]);
+ zebraSet->basenames[i] = nmem_strdup(zebraSet->nmem, basenames[i]);
res = resultSetSearch(zh, zebraSet->nmem, zebraSet->rset_nmem,
rpn, zebraSet);
(s->hits)++;
}
-ZebraSet resultSetAdd (ZebraHandle zh, const char *name, int ov)
+ZebraSet resultSetAdd(ZebraHandle zh, const char *name, int ov)
{
ZebraSet s;
int i;
s->rset_nmem = 0;
s->nmem = 0;
s->rpn = 0;
+ s->sortSpec = 0;
s->cache_position = 0;
s->cache_rfd = 0;
+ s->approx_limit = zh->approx_limit;
return s;
}
-ZebraSet resultSetGet (ZebraHandle zh, const char *name)
+ZebraSet resultSetGet(ZebraHandle zh, const char *name)
{
ZebraSet s;
if (!s->rset_nmem)
s->rset_nmem=nmem_create();
resultSetSearch(zh, nmem, s->rset_nmem, s->rpn, s);
+ if (s->rset && s->sortSpec)
+ {
+ int sort_status;
+ yaz_log(log_level_resultsets, "resort %s", name);
+ resultSetSortSingle (zh, nmem, s, s->rset, s->sortSpec,
+ &sort_status);
+ }
nmem_destroy (nmem);
}
return s;
}
}
-void resultSetDestroy (ZebraHandle zh, int num, char **names,int *statuses)
+void resultSetDestroy(ZebraHandle zh, int num, char **names,int *statuses)
{
ZebraSet * ss = &zh->sets;
int i;
RSET rset;
int i;
struct zset_sort_info *sort_info;
+ size_t sysno_mem_index = 0;
+
+ if (zh->m_staticrank)
+ sysno_mem_index = 1;
if (!log_level_set)
loglevels();
}
while (num_i < num && rset_read (rfd, &key, 0))
{
- zint this_sys = key.mem[0];
+ zint this_sys = key.mem[sysno_mem_index];
if (this_sys != psysno)
{
psysno = this_sys;
new_entry->score = score;
}
+static Z_RPNQuery *copy_RPNQuery(Z_RPNQuery *src, NMEM nmem)
+{
+ Z_RPNQuery *dst = 0;
+ ODR encode = odr_createmem(ODR_ENCODE);
+ ODR decode = odr_createmem(ODR_DECODE);
+
+ if (z_RPNQuery(encode, &src, 0, 0))
+ {
+ int len;
+ char *buf = odr_getbuf(encode, &len, 0);
+
+ if (buf)
+ {
+ odr_setbuf(decode, buf, len, 0);
+ z_RPNQuery(decode, &dst, 0, 0);
+ }
+ }
+ nmem_transfer(nmem, decode->mem);
+ odr_destroy(encode);
+ odr_destroy(decode);
+ return dst;
+}
+
+static Z_SortKeySpecList *copy_SortKeySpecList(Z_SortKeySpecList *src, NMEM nmem)
+{
+ Z_SortKeySpecList *dst = 0;
+ ODR encode = odr_createmem(ODR_ENCODE);
+ ODR decode = odr_createmem(ODR_DECODE);
+
+ if (z_SortKeySpecList(encode, &src, 0, 0))
+ {
+ int len;
+ char *buf = odr_getbuf(encode, &len, 0);
+
+ if (buf)
+ {
+ odr_setbuf(decode, buf, len, 0);
+ z_SortKeySpecList(decode, &dst, 0, 0);
+ }
+ }
+ nmem_transfer(nmem, decode->mem);
+ odr_destroy(encode);
+ odr_destroy(decode);
+ return dst;
+}
+
+ZebraSet resultSetClone(ZebraHandle zh, const char *setname,
+ ZebraSet rset)
+{
+ ZebraSet nset;
+ int i;
+
+ nset = resultSetAdd(zh, setname, 1);
+ if (!nset)
+ return 0;
+
+ nset->nmem = nmem_create();
+
+ nset->num_bases = rset->num_bases;
+ nset->basenames =
+ nmem_malloc (nset->nmem, nset->num_bases * sizeof(*rset->basenames));
+ for (i = 0; i<rset->num_bases; i++)
+ nset->basenames[i] = nmem_strdup(nset->nmem, rset->basenames[i]);
+
+ if (rset->rset)
+ nset->rset = rset_dup(rset->rset);
+ if (rset->rpn)
+ nset->rpn = copy_RPNQuery(rset->rpn, nset->nmem);
+ return nset;
+}
+
ZEBRA_RES resultSetSort(ZebraHandle zh, NMEM nmem,
int num_input_setnames, const char **input_setnames,
const char *output_setname,
return ZEBRA_FAIL;
}
if (strcmp (output_setname, input_setnames[0]))
- {
- rset = rset_dup (rset);
- sset = resultSetAdd (zh, output_setname, 1);
- sset->rset = rset;
- }
+ sset = resultSetClone(zh, output_setname, sset);
+ sset->sortSpec = copy_SortKeySpecList(sort_sequence, sset->nmem);
return resultSetSortSingle (zh, nmem, sset, rset, sort_sequence,
sort_status);
}
TERMID termid;
TERMID *terms;
int numTerms = 0;
+ size_t sysno_mem_index = 0;
+
+ if (zh->m_staticrank)
+ sysno_mem_index = 1;
+
assert(nmem); /* compiler shut up about unused param */
sset->sort_info->num_entries = 0;
rfd = rset_open (rset, RSETF_READ);
while (rset_read (rfd, &key, &termid))
{
- zint this_sys = key.mem[0];
+ zint this_sys = key.mem[sysno_mem_index];
if (log_level_searchhits)
key_logdump_txt(log_level_searchhits, &key, termid->name);
kno++;
ZebraRankClass rank_class;
struct zset_sort_info *sort_info;
const char *rank_handler_name = res_get_def(zh->res, "rank", "rank-1");
+ size_t sysno_mem_index = 0;
+
+ if (zh->m_staticrank)
+ sysno_mem_index = 1;
if (!log_level_set)
loglevels();
terms = (TERMID *) nmem_malloc(nmem, sizeof(*terms)*n);
rset_getterms(rset, terms, n, &numTerms);
+
rank_class = zebraRankLookup(zh, rank_handler_name);
if (!rank_class)
{
RSFD rfd = rset_open(rset, RSETF_READ);
struct rank_control *rc = rank_class->control;
double score;
+ zint count = 0;
void *handle =
(*rc->begin) (zh->reg, rank_class->class_handle, rset, nmem,
terms, numTerms);
- zint psysno = 0;
+ zint psysno = 0; /* previous doc id / sys no */
+ zint pstaticrank = 0; /* previous static rank */
+ int stop_flag = 0;
while (rset_read(rfd, &key, &termid))
{
- zint this_sys = key.mem[0];
+ zint this_sys = key.mem[sysno_mem_index];
+
zint seqno = key.mem[key.len-1];
kno++;
if (log_level_searchhits)
key_logdump_txt(log_level_searchhits, &key, termid->name);
- if (this_sys != psysno)
- {
- if (rfd->counted_items >= rset->hits_limit)
+ if (this_sys != psysno)
+ { /* new record .. */
+ if (rfd->counted_items > rset->hits_limit)
break;
if (psysno)
- {
- score = (*rc->calc) (handle, psysno);
+ { /* only if we did have a previous record */
+ score = (*rc->calc) (handle, psysno, pstaticrank,
+ &stop_flag);
+ /* insert the hit. A=Ascending */
resultSetInsertRank (zh, sort_info, psysno, score, 'A');
+ count++;
+ if (stop_flag)
+ break;
}
psysno = this_sys;
+ if (zh->m_staticrank)
+ pstaticrank = key.mem[0];
}
(*rc->add) (handle, CAST_ZINT_TO_INT(seqno), termid);
}
+ /* no more items */
if (psysno)
- {
- score = (*rc->calc)(handle, psysno);
+ { /* we had - at least - one record */
+ score = (*rc->calc)(handle, psysno, pstaticrank, &stop_flag);
+ /* insert the hit. A=Ascending */
resultSetInsertRank(zh, sort_info, psysno, score, 'A');
+ count++;
}
(*rc->end) (zh->reg, handle);
rset_close (rfd);
if (approx_array)
approx_array[no] = rset->hits_approx;
#if 0
- yaz_log(YLOG_LOG, "rset=%p term=%s count=" ZINT_FORMAT,
- rset, rset->term->name, rset->hits_count);
+ yaz_log(YLOG_LOG, "rset=%p term=%s limit=" ZINT_FORMAT
+ " count=" ZINT_FORMAT,
+ rset, rset->term->name, rset->hits_limit, rset->hits_count);
#endif
no++;
}
ZEBRA_RES zebra_result_set_term_info(ZebraHandle zh, const char *setname,
int no, zint *count, int *approx,
- char *termbuf, size_t *termlen)
+ char *termbuf, size_t *termlen,
+ const char **term_ref_id)
{
ZebraSet sset = resultSetGet(zh, setname);
if (sset)
}
termbuf[*termlen] = '\0';
}
+ if (term_ref_id)
+ *term_ref_id = term_array[no]->ref_id;
xfree(term_array);
xfree(hits_array);
zint sysno, zebra_snippets *snippets)
{
ZebraSet sset = resultSetGet(zh, setname);
- yaz_log(YLOG_LOG, "zebra_get_hit_vector setname=%s zysno=" ZINT_FORMAT,
+ yaz_log(YLOG_DEBUG, "zebra_get_hit_vector setname=%s zysno=" ZINT_FORMAT,
setname, sysno);
if (!sset)
return ZEBRA_FAIL;
if (termid)
{
struct ord_list *ol;
- key_logdump_txt(YLOG_LOG, &key, termid->name);
for (ol = termid->ol; ol; ol = ol->next)
{
- yaz_log(YLOG_LOG, " ord=%d", ol->ord);
zebra_snippets_append(snippets, key.mem[key.len-1],
ol->ord, termid->name);
}
rset_delete(rset_comb);
nmem_destroy(nmem);
+ kc->dec(kc);
}
return ZEBRA_OK;
}
+/*
+ * Local variables:
+ * c-basic-offset: 4
+ * indent-tabs-mode: nil
+ * End:
+ * vim: shiftwidth=4 tabstop=8 expandtab
+ */
+
projects / idzebra-moved-to-github.git / blobdiff