1 /* $Id: zsets.c,v 1.72 2004-11-29 21:55:28 adam Exp $
2 Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004
5 This file is part of the Zebra server.
7 Zebra is free software; you can redistribute it and/or modify it under
8 the terms of the GNU General Public License as published by the Free
9 Software Foundation; either version 2, or (at your option) any later
12 Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with Zebra; see the file LICENSE.zebra. If not, write to the
19 Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
35 #define SORT_IDX_ENTRYSIZE 64
36 #define ZSET_SORT_MAX_LEVEL 3
38 struct zebra_set_term_entry {
49 NMEM rset_nmem; /* for creating the rsets in */
54 struct zset_sort_info *sort_info;
55 struct zebra_set_term_entry *term_entries;
57 struct zebra_set *next;
61 struct zset_sort_entry {
64 char buf[ZSET_SORT_MAX_LEVEL][SORT_IDX_ENTRYSIZE];
67 struct zset_sort_info {
70 struct zset_sort_entry *all_entries;
71 struct zset_sort_entry **entries;
74 ZebraSet resultSetAddRPN (ZebraHandle zh, NMEM m,
75 Z_RPNQuery *rpn, int num_bases,
86 zebraSet = resultSetAdd (zh, setname, 1);
92 zebraSet->rset_nmem=nmem_create();
94 zebraSet->num_bases = num_bases;
96 nmem_malloc (zebraSet->nmem, num_bases * sizeof(*zebraSet->basenames));
97 for (i = 0; i<num_bases; i++)
98 zebraSet->basenames[i] = nmem_strdup (zebraSet->nmem, basenames[i]);
101 zebraSet->rset = rpn_search (zh, zebraSet->nmem, zebraSet->rset_nmem,
102 rpn, zebraSet->num_bases,
103 zebraSet->basenames, zebraSet->name,
105 zh->hits = zebraSet->hits;
108 zebraSet->locked = 0;
112 void resultSetAddTerm (ZebraHandle zh, ZebraSet s, int reg_type,
113 const char *db, int set,
114 int use, const char *term)
116 assert(zh); /* compiler shut up */
118 s->nmem = nmem_create ();
119 if (!s->term_entries)
122 s->term_entries_max = 1000;
124 nmem_malloc (s->nmem, s->term_entries_max *
125 sizeof(*s->term_entries));
126 for (i = 0; i < s->term_entries_max; i++)
127 s->term_entries[i].term = 0;
129 if (s->hits < s->term_entries_max)
131 s->term_entries[s->hits].reg_type = reg_type;
132 s->term_entries[s->hits].db = nmem_strdup (s->nmem, db);
133 s->term_entries[s->hits].set = set;
134 s->term_entries[s->hits].use = use;
135 s->term_entries[s->hits].term = nmem_strdup (s->nmem, term);
140 ZebraSet resultSetAdd (ZebraHandle zh, const char *name, int ov)
145 for (s = zh->sets; s; s = s->next)
146 if (!strcmp (s->name, name))
150 yaz_log(YLOG_DEBUG, "updating result set %s", name);
151 if (!ov || s->locked)
154 rset_delete (s->rset);
156 nmem_destroy (s->rset_nmem);
158 nmem_destroy (s->nmem);
162 const char *sort_max_str = zebra_get_resource(zh, "sortmax", "1000");
164 yaz_log(YLOG_DEBUG, "adding result set %s", name);
165 s = (ZebraSet) xmalloc (sizeof(*s));
168 s->name = (char *) xmalloc (strlen(name)+1);
169 strcpy (s->name, name);
171 s->sort_info = (struct zset_sort_info *)
172 xmalloc (sizeof(*s->sort_info));
173 s->sort_info->max_entries = atoi(sort_max_str);
174 if (s->sort_info->max_entries < 2)
175 s->sort_info->max_entries = 2;
177 s->sort_info->entries = (struct zset_sort_entry **)
178 xmalloc (sizeof(*s->sort_info->entries) *
179 s->sort_info->max_entries);
180 s->sort_info->all_entries = (struct zset_sort_entry *)
181 xmalloc (sizeof(*s->sort_info->all_entries) *
182 s->sort_info->max_entries);
183 for (i = 0; i < s->sort_info->max_entries; i++)
184 s->sort_info->entries[i] = s->sort_info->all_entries + i;
196 ZebraSet resultSetGet (ZebraHandle zh, const char *name)
200 for (s = zh->sets; s; s = s->next)
201 if (!strcmp (s->name, name))
203 if (!s->term_entries && !s->rset && s->rpn)
205 NMEM nmem = nmem_create ();
206 yaz_log(YLOG_LOG, "research %s", name);
208 s->rset_nmem=nmem_create();
210 rpn_search (zh, nmem, s->rset_nmem, s->rpn, s->num_bases,
211 s->basenames, s->name, s);
219 void resultSetInvalidate (ZebraHandle zh)
221 ZebraSet s = zh->sets;
223 for (; s; s = s->next)
226 rset_delete (s->rset);
229 nmem_destroy(s->rset_nmem);
234 void resultSetDestroy (ZebraHandle zh, int num, char **names,int *statuses)
236 ZebraSet * ss = &zh->sets;
240 for (i = 0; i<num; i++)
241 statuses[i] = Z_DeleteStatus_resultSetDidNotExist;
243 zh->errString = NULL;
250 for (i = 0; i<num; i++)
251 if (!strcmp (s->name, names[i]))
254 statuses[i] = Z_DeleteStatus_success;
263 xfree (s->sort_info->all_entries);
264 xfree (s->sort_info->entries);
265 xfree (s->sort_info);
268 nmem_destroy (s->nmem);
270 rset_delete (s->rset);
272 nmem_destroy(s->rset_nmem);
281 ZebraMetaRecord *zebra_meta_records_create (ZebraHandle zh, const char *name,
282 int num, zint *positions)
285 ZebraMetaRecord *sr = 0;
288 struct zset_sort_info *sort_info;
290 if (!(sset = resultSetGet (zh, name)))
292 if (!(rset = sset->rset))
294 if (!sset->term_entries)
296 sr = (ZebraMetaRecord *) xmalloc (sizeof(*sr) * num);
297 for (i = 0; i<num; i++)
304 if (positions[i] <= sset->term_entries_max)
306 sr[i].term = sset->term_entries[positions[i]-1].term;
307 sr[i].db = sset->term_entries[positions[i]-1].db;
313 sr = (ZebraMetaRecord *) xmalloc (sizeof(*sr) * num);
314 for (i = 0; i<num; i++)
321 sort_info = sset->sort_info;
326 for (i = 0; i<num; i++)
328 position = positions[i];
329 if (position > 0 && position <= sort_info->num_entries)
331 yaz_log(YLOG_DEBUG, "got pos=%d (sorted)", position);
332 sr[i].sysno = sort_info->entries[position-1]->sysno;
333 sr[i].score = sort_info->entries[position-1]->score;
337 /* did we really get all entries using sort ? */
338 for (i = 0; i<num; i++)
343 if (i < num) /* nope, get the rest, unsorted - sorry */
352 position = sort_info->num_entries;
353 while (num_i < num && positions[num_i] < position)
355 rfd = rset_open (rset, RSETF_READ);
356 while (num_i < num && rset_read (rfd, &key, 0))
358 zint this_sys = key.mem[0];
359 yaz_log(YLOG_LOG, "RSET READ " ZINT_FORMAT, this_sys);
360 if (this_sys != psysno)
365 /* determine we alreay have this in our set */
366 for (i = sort_info->num_entries; --i >= 0; )
367 if (psysno == sort_info->entries[i]->sysno)
373 assert (num_i < num);
374 if (position == positions[num_i])
376 sr[num_i].sysno = psysno;
377 yaz_log(YLOG_DEBUG, "got pos=%d (unsorted)", position);
378 sr[num_i].score = -1;
389 void zebra_meta_records_destroy (ZebraHandle zh, ZebraMetaRecord *records,
392 assert(zh); /* compiler shut up about unused arg */
402 void resultSetInsertSort (ZebraHandle zh, ZebraSet sset,
403 struct sortKeyInfo *criteria, int num_criteria,
406 struct zset_sort_entry this_entry;
407 struct zset_sort_entry *new_entry = NULL;
408 struct zset_sort_info *sort_info = sset->sort_info;
411 sortIdx_sysno (zh->reg->sortIdx, sysno);
412 for (i = 0; i<num_criteria; i++)
414 sortIdx_type (zh->reg->sortIdx, criteria[i].attrUse);
415 sortIdx_read (zh->reg->sortIdx, this_entry.buf[i]);
417 i = sort_info->num_entries;
421 for (j = 0; j<num_criteria; j++)
423 if (criteria[j].numerical)
425 double diff = atof(this_entry.buf[j]) -
426 atof(sort_info->entries[i]->buf[j]);
435 rel = memcmp (this_entry.buf[j], sort_info->entries[i]->buf[j],
443 if (criteria[j].relation == 'A')
448 else if (criteria[j].relation == 'D')
455 j = sort_info->max_entries;
459 if (sort_info->num_entries == j)
462 j = (sort_info->num_entries)++;
463 new_entry = sort_info->entries[j];
466 sort_info->entries[j] = sort_info->entries[j-1];
469 sort_info->entries[i] = new_entry;
471 for (i = 0; i<num_criteria; i++)
472 memcpy (new_entry->buf[i], this_entry.buf[i], SORT_IDX_ENTRYSIZE);
473 new_entry->sysno = sysno;
474 new_entry->score = -1;
477 void resultSetInsertRank (ZebraHandle zh, struct zset_sort_info *sort_info,
478 zint sysno, int score, int relation)
480 struct zset_sort_entry *new_entry = NULL;
482 assert(zh); /* compiler shut up about unused arg */
484 i = sort_info->num_entries;
489 rel = score - sort_info->entries[i]->score;
496 else if (relation == 'A')
503 j = sort_info->max_entries;
507 if (sort_info->num_entries == j)
510 j = (sort_info->num_entries)++;
512 new_entry = sort_info->entries[j];
515 sort_info->entries[j] = sort_info->entries[j-1];
518 sort_info->entries[i] = new_entry;
520 new_entry->sysno = sysno;
521 new_entry->score = score;
524 void resultSetSort (ZebraHandle zh, NMEM nmem,
525 int num_input_setnames, const char **input_setnames,
526 const char *output_setname,
527 Z_SortKeySpecList *sort_sequence, int *sort_status)
532 if (num_input_setnames == 0)
537 if (num_input_setnames > 1)
542 yaz_log(YLOG_DEBUG, "result set sort input=%s output=%s",
543 *input_setnames, output_setname);
544 sset = resultSetGet (zh, input_setnames[0]);
548 zh->errString = nmem_strdup (nmem, input_setnames[0]);
551 if (!(rset = sset->rset))
554 zh->errString = nmem_strdup (nmem, input_setnames[0]);
557 if (strcmp (output_setname, input_setnames[0]))
559 rset = rset_dup (rset);
560 sset = resultSetAdd (zh, output_setname, 1);
563 resultSetSortSingle (zh, nmem, sset, rset, sort_sequence, sort_status);
566 void resultSetSortSingle (ZebraHandle zh, NMEM nmem,
567 ZebraSet sset, RSET rset,
568 Z_SortKeySpecList *sort_sequence, int *sort_status)
575 struct sortKeyInfo sort_criteria[3];
582 assert(nmem); /* compiler shut up about unused param */
583 sset->sort_info->num_entries = 0;
585 rset_getterms(rset, 0, 0, &n);
586 terms = (TERMID *) nmem_malloc(nmem, sizeof(*terms)*n);
587 rset_getterms(rset, terms, n, &numTerms);
590 num_criteria = sort_sequence->num_specs;
591 if (num_criteria > 3)
593 for (i = 0; i < num_criteria; i++)
595 Z_SortKeySpec *sks = sort_sequence->specs[i];
598 if (*sks->sortRelation == Z_SortKeySpec_ascending)
599 sort_criteria[i].relation = 'A';
600 else if (*sks->sortRelation == Z_SortKeySpec_descending)
601 sort_criteria[i].relation = 'D';
607 if (sks->sortElement->which == Z_SortElement_databaseSpecific)
612 else if (sks->sortElement->which != Z_SortElement_generic)
617 sk = sks->sortElement->u.generic;
620 case Z_SortKey_sortField:
621 yaz_log(YLOG_DEBUG, "Sort: key %d is of type sortField", i+1);
624 case Z_SortKey_elementSpec:
625 yaz_log(YLOG_DEBUG, "Sort: key %d is of type elementSpec", i+1);
628 case Z_SortKey_sortAttributes:
629 yaz_log(YLOG_DEBUG, "Sort: key %d is of type sortAttributes", i+1);
630 sort_criteria[i].attrUse =
631 zebra_maps_sort (zh->reg->zebra_maps,
632 sk->u.sortAttributes,
633 &sort_criteria[i].numerical);
634 yaz_log(YLOG_DEBUG, "use value = %d", sort_criteria[i].attrUse);
635 if (sort_criteria[i].attrUse == -1)
640 if (sortIdx_type (zh->reg->sortIdx, sort_criteria[i].attrUse))
648 rfd = rset_open (rset, RSETF_READ);
649 while (rset_read (rfd, &key, &termid))
650 /* FIXME - pass a TERMID *, and use it for something below !! */
652 zint this_sys = key.mem[0];
654 if (this_sys != psysno)
658 resultSetInsertSort (zh, sset,
659 sort_criteria, num_criteria, psysno);
663 yaz_log(YLOG_DEBUG, ZINT_FORMAT " keys, " ZINT_FORMAT " sysnos, sort",
665 for (i = 0; i < numTerms; i++)
666 yaz_log(YLOG_LOG, "term=\"%s\" type=%s count=" ZINT_FORMAT,
667 terms[i]->name, terms[i]->flags, rset_count(terms[i]->rset));
668 *sort_status = Z_SortResponse_success;
671 RSET resultSetRef (ZebraHandle zh, const char *resultSetId)
675 if ((s = resultSetGet (zh, resultSetId)))
680 void resultSetRank (ZebraHandle zh, ZebraSet zebraSet, RSET rset, NMEM nmem)
690 ZebraRankClass rank_class;
691 struct rank_control *rc;
692 struct zset_sort_info *sort_info;
693 const char *rank_handler_name = res_get_def(zh->res, "rank", "rank-1");
695 zint est=-2; /* -2 not done, -1 can't do, >0 actual estimate*/
699 sort_info = zebraSet->sort_info;
700 sort_info->num_entries = 0;
702 rset_getterms(rset, 0, 0, &n);
703 terms = (TERMID *) nmem_malloc(nmem, sizeof(*terms)*n);
704 rset_getterms(rset, terms, n, &numTerms);
706 rfd = rset_open (rset, RSETF_READ);
708 rank_class = zebraRankLookup (zh, rank_handler_name);
711 yaz_log(YLOG_WARN, "No such rank handler: %s", rank_handler_name);
714 rc = rank_class->control;
716 if (rset_read (rfd, &key, &termid))
718 zint psysno = key.mem[0];
721 (*rc->begin) (zh->reg, rank_class->class_handle, rset, nmem,
724 esthits = atoi(res_get_def(zh->res, "estimatehits","0"));
726 est = -1; /* can not do */
729 zint this_sys = key.mem[0]; /* FIXME - assumes scope==2 */
730 zint seqno = key.mem[key.len-1]; /* FIXME - assumes scope==2 */
732 yaz_log(YLOG_LOG, "got sysno=%lld", this_sys);
733 if (this_sys != psysno)
735 score = (*rc->calc) (handle, psysno);
737 resultSetInsertRank (zh, sort_info, psysno, score, 'A');
741 (*rc->add) (handle, seqno, termid);
743 if ((est==-2) && (zebraSet->hits==esthits))
744 { /* time to estimate the hits */
745 rset_pos(rfd,&cur,&tot);
748 est = (zint)(0.5+zebraSet->hits/ratio);
749 yaz_log(YLOG_LOG, "Estimating hits (%s) "
750 "%0.1f->" ZINT_FORMAT
751 "; %0.1f->" ZINT_FORMAT,
755 i = 0; /* round to 3 significant digits */
762 zebraSet->hits = est;
766 while (rset_read (rfd, &key,&termid) && (est<0) );
767 score = (*rc->calc)(handle, psysno);
768 resultSetInsertRank(zh, sort_info, psysno, score, 'A');
769 (*rc->end) (zh->reg, handle);
773 yaz_log(YLOG_DEBUG, ZINT_FORMAT " keys, " ZINT_FORMAT " sysnos, rank",
774 kno, zebraSet->hits);
775 for (i = 0; i < numTerms; i++)
777 yaz_log(YLOG_LOG, "term=\"%s\" type=%s count=" ZINT_FORMAT,
778 terms[i]->name, terms[i]->flags, rset_count(terms[i]->rset));
782 ZebraRankClass zebraRankLookup (ZebraHandle zh, const char *name)
784 ZebraRankClass p = zh->reg->rank_classes;
785 while (p && strcmp (p->control->name, name))
787 if (p && !p->init_flag)
789 if (p->control->create)
790 p->class_handle = (*p->control->create)(zh);
796 void zebraRankInstall (struct zebra_register *reg, struct rank_control *ctrl)
798 ZebraRankClass p = (ZebraRankClass) xmalloc (sizeof(*p));
799 p->control = (struct rank_control *) xmalloc (sizeof(*p->control));
800 memcpy (p->control, ctrl, sizeof(*p->control));
801 p->control->name = xstrdup (ctrl->name);
803 p->next = reg->rank_classes;
804 reg->rank_classes = p;
807 void zebraRankDestroy (struct zebra_register *reg)
809 ZebraRankClass p = reg->rank_classes;
812 ZebraRankClass p_next = p->next;
813 if (p->init_flag && p->control->destroy)
814 (*p->control->destroy)(reg, p->class_handle);
815 xfree(p->control->name);
820 reg->rank_classes = NULL;