1 /* $Id: rsbetween.c,v 1.22 2004-08-26 11:11:59 heikki Exp $
2 Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002
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
24 /* rsbetween is (mostly) used for xml searches. It returns the hits of the
25 * "middle" rset, that are in between the "left" and "right" rsets. For
26 * example "Shakespeare" in between "<title>" and </title>. The thing is
27 * complicated by the inclusion of attributes (from their own rset). If attrs
28 * specified, they must match the "left" rset (start tag). "Hamlet" between
29 * "<title lang=eng>" and "</title>". (This assumes that the attributes are
30 * indexed to the same seqno as the tags).
39 #include <rsbetween.h>
41 #define RSBETWEEN_DEBUG 0
43 static RSFD r_open_between (RSET ct, int flag);
44 static void r_close_between (RSFD rfd);
45 static void r_delete_between (RSET ct);
46 static void r_rewind_between (RSFD rfd);
47 static int r_forward_between(RSET ct, RSFD rfd, void *buf,
48 int (*cmpfunc)(const void *p1, const void *p2),
49 const void *untilbuf);
50 static int r_read_between (RSFD rfd, void *buf);
51 static int r_write_between (RSFD rfd, const void *buf);
52 static void r_pos_between (RSFD rfd, double *current, double *total);
54 static const struct rset_control control =
61 r_forward_between, /* rset_default_forward, */
68 const struct rset_control *rset_kind_between = &control;
70 struct rset_between_info {
76 int (*cmp)(const void *p1, const void *p2);
77 char *(*printer)(const void *p1, char *buf);
78 struct rset_between_rfd *rfd_list;
79 struct rset_between_rfd *free_list;
82 struct rset_between_rfd {
96 struct rset_between_rfd *next;
97 struct rset_between_info *info;
102 static void log2 (struct rset_between_rfd *p, char *msg, int cmp_l, int cmp_r)
107 logf(LOG_DEBUG,"btw: %s l=%s(%d/%d) m=%s(%d) r=%s(%d/%d), lev=%d",
109 (*p->info->printer)(p->buf_l, buf_l), p->more_l, cmp_l,
110 (*p->info->printer)(p->buf_m, buf_m), p->more_m,
111 (*p->info->printer)(p->buf_r, buf_r), p->more_r, cmp_r,
116 RSET rsbetween_create( NMEM nmem, int key_size,
117 int (*cmp)(const void *p1, const void *p2),
118 RSET rset_l, RSET rset_m, RSET rset_r, RSET rset_attr,
119 char *(*printer)(const void *p1, char *buf) )
121 RSET rnew=rset_create_base(&control, nmem);
122 struct rset_between_info *info;
123 info = (struct rset_between_info *) nmem_malloc(rnew->nmem,sizeof(*info));
124 info->key_size = key_size;
125 info->rset_l = rset_l;
126 info->rset_m = rset_m;
127 info->rset_r = rset_r;
128 info->rset_attr = rset_attr;
130 info->printer = printer;
131 info->rfd_list = NULL;
132 info->free_list = NULL;
139 static void r_delete_between (RSET ct)
141 struct rset_between_info *info = (struct rset_between_info *) ct->priv;
143 assert (info->rfd_list == NULL);
144 rset_delete (info->rset_l);
145 rset_delete (info->rset_m);
146 rset_delete (info->rset_r);
148 rset_delete (info->rset_attr);
152 static RSFD r_open_between (RSET ct, int flag)
154 struct rset_between_info *info = (struct rset_between_info *) ct->priv;
155 struct rset_between_rfd *rfd;
157 if (flag & RSETF_WRITE)
159 logf (LOG_FATAL, "between set type is read-only");
164 info->free_list=rfd->next;
166 rfd = (struct rset_between_rfd *) nmem_malloc(ct->nmem, (sizeof(*rfd)));
167 rfd->buf_l = nmem_malloc(ct->nmem, (info->key_size));
168 rfd->buf_m = nmem_malloc(ct->nmem, (info->key_size));
169 rfd->buf_r = nmem_malloc(ct->nmem, (info->key_size));
170 rfd->buf_attr = nmem_malloc(ct->nmem, (info->key_size));
172 rfd->next = info->rfd_list;
173 info->rfd_list = rfd;
176 rfd->rfd_l = rset_open (info->rset_l, RSETF_READ);
177 rfd->rfd_m = rset_open (info->rset_m, RSETF_READ);
178 rfd->rfd_r = rset_open (info->rset_r, RSETF_READ);
180 rfd->more_l = rset_read (info->rset_l, rfd->rfd_l, rfd->buf_l);
181 rfd->more_m = rset_read (info->rset_m, rfd->rfd_m, rfd->buf_m);
182 rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r);
185 rfd->rfd_attr = rset_open (info->rset_attr, RSETF_READ);
186 rfd->more_attr = rset_read (info->rset_attr, rfd->rfd_attr,
194 static void r_close_between (RSFD rfd)
196 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
197 struct rset_between_rfd **rfdp;
199 for (rfdp = &info->rfd_list; *rfdp; rfdp = &(*rfdp)->next)
202 struct rset_between_rfd *rfd_tmp=*rfdp;
203 rset_close (info->rset_l, (*rfdp)->rfd_l);
204 rset_close (info->rset_m, (*rfdp)->rfd_m);
205 rset_close (info->rset_r, (*rfdp)->rfd_r);
207 rset_close (info->rset_attr, (*rfdp)->rfd_attr);
208 *rfdp = (*rfdp)->next;
209 rfd_tmp->next=info->free_list;
210 info->free_list=rfd_tmp;
213 logf (LOG_FATAL, "r_close_between but no rfd match!");
217 static void r_rewind_between (RSFD rfd)
219 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
220 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
223 logf (LOG_DEBUG, "rsbetween_rewind");
225 rset_rewind (info->rset_l, p->rfd_l);
226 rset_rewind (info->rset_m, p->rfd_m);
227 rset_rewind (info->rset_r, p->rfd_r);
228 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l);
229 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
230 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r);
233 rset_rewind (info->rset_attr, p->rfd_attr);
234 p->more_attr = rset_read (info->rset_attr, p->rfd_attr, p->buf_attr);
242 static int r_forward_between(RSET ct, RSFD rfd, void *buf,
243 int (*cmpfunc)(const void *p1, const void *p2),
244 const void *untilbuf)
246 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
247 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
250 log2( p, "fwd: before forward", 0,0);
252 /* It is enough to forward the m pointer here, the read will */
253 /* naturally forward the l, m, and attr pointers */
255 p->more_m=rset_forward(info->rset_m,p->rfd_m, p->buf_m,
258 log2( p, "fwd: after forward M", 0,0);
260 rc = r_read_between(rfd, buf);
262 log2( p, "fwd: after forward", 0,0);
270 static int r_read_between (RSFD rfd, void *buf)
272 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
273 struct rset_between_info *info = p->info;
281 log2( p, "start of loop", cmp_l, cmp_r);
284 /* forward L until past m, count levels, note rec boundaries */
286 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
290 cmp_l=2; /* past this record */
293 log2( p, "after first L", cmp_l, cmp_r);
296 while (cmp_l < 0) /* l before m */
299 p->level=0; /* earlier record */
302 p->level++; /* relevant start tag */
304 if (!info->rset_attr)
312 cmp_attr = (*info->cmp)(p->buf_attr, p->buf_l);
318 else if (cmp_attr > 0)
320 else if (cmp_attr==-1)
321 p->more_attr = rset_read (info->rset_attr,
322 p->rfd_attr, p->buf_attr);
323 /* if we had a forward that went all the way to
324 * the seqno, we could use that. But fwd only goes
326 else if (cmp_attr==-2)
328 p->more_attr = rset_forward(
329 info->rset_attr, p->rfd_attr,
330 p->buf_attr, info->cmp, p->buf_l);
332 logf(LOG_DEBUG, "btw: after frowarding attr m=%d",p->more_attr);
335 } /* while more_attr */
344 p->more_l=rset_forward( info->rset_l, p->rfd_l,
345 p->buf_l, info->cmp, p->buf_m);
347 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
351 log2( p, "after forwarding L", cmp_l, cmp_r);
356 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l);
359 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l);
363 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
368 log2( p, "end of L loop", cmp_l, cmp_r);
373 /* forward R until past m, count levels */
375 log2( p, "Before moving R", cmp_l, cmp_r);
378 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
382 log2( p, "after first R", cmp_l, cmp_r);
384 while (cmp_r < 0) /* r before m */
386 /* -2, earlier record, don't count level */
388 p->level--; /* relevant end tag */
394 p->more_r=rset_forward( info->rset_r, p->rfd_r,
395 p->buf_r, info->cmp, p->buf_m);
398 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r);
401 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
404 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r);
405 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
411 log2( p, "End of R loop", cmp_l, cmp_r);
415 if ( ( p->level <= 0 ) && ! p->more_l)
416 return 0; /* no more start tags, nothing more to find */
418 if ( attr_match && p->level > 0) /* within a tag pair (or deeper) */
420 memcpy (buf, p->buf_m, info->key_size);
422 log2( p, "Returning a hit (and forwarding m)", cmp_l, cmp_r);
424 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
430 else if ( ! p->more_l ) /* not in data, no more starts */
433 log2( p, "no more starts, exiting without a hit", cmp_l, cmp_r);
435 return 0; /* ergo, nothing can be found. stop scanning */
441 p->more_m=rset_forward( info->rset_m, p->rfd_m,
442 p->buf_m, info->cmp, p->buf_l);
445 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
450 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
453 log2( p, "End of M loop", cmp_l, cmp_r);
458 log2( p, "Exiting, nothing more in m", cmp_l, cmp_r);
460 return 0; /* no more data possible */
466 static int r_write_between (RSFD rfd, const void *buf)
468 logf (LOG_FATAL, "between set type is read-only");
473 static void r_pos_between (RSFD rfd, double *current, double *total)
475 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
476 struct rset_between_info *info = p->info;
482 rset_pos(info->rset_l, p->rfd_l, &lcur, <ot);
483 rset_pos(info->rset_m, p->rfd_m, &mcur, &mtot);
484 rset_pos(info->rset_r, p->rfd_r, &rcur, &rtot);
485 if ( (ltot<0) && (mtot<0) && (rtot<0) ) { /*no position */
486 *current=mcur; /* return same as you got */
487 *total=mtot; /* probably -1 for not available */
489 if ( ltot<0) { ltot=0; lcur=0;} /* if only one useful, use it */
490 if ( mtot<0) { mtot=0; mcur=0;}
491 if ( rtot<0) { rtot=0; rcur=0;}
492 if ( ltot+mtot+rtot < 1 ) { /* empty rset */
497 r=1.0*(lcur+mcur+rcur)/(ltot+mtot+rtot); /* weighed average of l and r */
501 yaz_log(LOG_DEBUG,"betw_pos: (%s/%s) %0.1f/%0.1f= %0.4f ",
502 info->rset_l->control->desc, info->rset_r->control->desc,
503 *current, *total, r);