1 /* $Id: rsbetween.c,v 1.19 2004-08-23 12:38:53 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 void *r_create_between(RSET ct, const struct rset_control *sel, void *parms);
44 static RSFD r_open_between (RSET ct, int flag);
45 static void r_close_between (RSFD rfd);
46 static void r_delete_between (RSET ct);
47 static void r_rewind_between (RSFD rfd);
48 static int r_forward_between(RSET ct, RSFD rfd, void *buf,
49 int (*cmpfunc)(const void *p1, const void *p2),
50 const void *untilbuf);
51 static int r_read_between (RSFD rfd, void *buf);
52 static int r_write_between (RSFD rfd, const void *buf);
53 static void r_pos_between (RSFD rfd, double *current, double *total);
55 static const struct rset_control control_between =
63 r_forward_between, /* rset_default_forward, */
70 const struct rset_control *rset_kind_between = &control_between;
72 struct rset_between_info {
78 int (*cmp)(const void *p1, const void *p2);
79 char *(*printer)(const void *p1, char *buf);
80 struct rset_between_rfd *rfd_list;
83 struct rset_between_rfd {
97 struct rset_between_rfd *next;
98 struct rset_between_info *info;
103 static void log2 (struct rset_between_rfd *p, char *msg, int cmp_l, int cmp_r)
108 logf(LOG_DEBUG,"btw: %s l=%s(%d/%d) m=%s(%d) r=%s(%d/%d), lev=%d",
110 (*p->info->printer)(p->buf_l, buf_l), p->more_l, cmp_l,
111 (*p->info->printer)(p->buf_m, buf_m), p->more_m,
112 (*p->info->printer)(p->buf_r, buf_r), p->more_r, cmp_r,
117 static void *r_create_between (RSET ct, const struct rset_control *sel,
120 rset_between_parms *between_parms = (rset_between_parms *) parms;
121 struct rset_between_info *info;
123 info = (struct rset_between_info *) xmalloc (sizeof(*info));
124 info->key_size = between_parms->key_size;
125 info->rset_l = between_parms->rset_l;
126 info->rset_m = between_parms->rset_m;
127 info->rset_r = between_parms->rset_r;
128 info->rset_attr = between_parms->rset_attr;
129 info->cmp = between_parms->cmp;
130 info->printer = between_parms->printer;
131 info->rfd_list = NULL;
135 static RSFD r_open_between (RSET ct, int flag)
137 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
138 struct rset_between_rfd *rfd;
140 if (flag & RSETF_WRITE)
142 logf (LOG_FATAL, "between set type is read-only");
145 rfd = (struct rset_between_rfd *) xmalloc (sizeof(*rfd));
146 rfd->next = info->rfd_list;
147 info->rfd_list = rfd;
150 rfd->buf_l = xmalloc (info->key_size);
151 rfd->buf_m = xmalloc (info->key_size);
152 rfd->buf_r = xmalloc (info->key_size);
153 rfd->buf_attr = xmalloc (info->key_size);
155 rfd->rfd_l = rset_open (info->rset_l, RSETF_READ);
156 rfd->rfd_m = rset_open (info->rset_m, RSETF_READ);
157 rfd->rfd_r = rset_open (info->rset_r, RSETF_READ);
159 rfd->more_l = rset_read (info->rset_l, rfd->rfd_l, rfd->buf_l);
160 rfd->more_m = rset_read (info->rset_m, rfd->rfd_m, rfd->buf_m);
161 rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r);
164 rfd->rfd_attr = rset_open (info->rset_attr, RSETF_READ);
165 rfd->more_attr = rset_read (info->rset_attr, rfd->rfd_attr,
173 static void r_close_between (RSFD rfd)
175 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
176 struct rset_between_rfd **rfdp;
178 for (rfdp = &info->rfd_list; *rfdp; rfdp = &(*rfdp)->next)
181 xfree ((*rfdp)->buf_l);
182 xfree ((*rfdp)->buf_m);
183 xfree ((*rfdp)->buf_r);
184 xfree ((*rfdp)->buf_attr);
185 rset_close (info->rset_l, (*rfdp)->rfd_l);
186 rset_close (info->rset_m, (*rfdp)->rfd_m);
187 rset_close (info->rset_r, (*rfdp)->rfd_r);
189 rset_close (info->rset_attr, (*rfdp)->rfd_attr);
191 *rfdp = (*rfdp)->next;
195 logf (LOG_FATAL, "r_close_between but no rfd match!");
199 static void r_delete_between (RSET ct)
201 struct rset_between_info *info = (struct rset_between_info *) ct->buf;
203 assert (info->rfd_list == NULL);
204 rset_delete (info->rset_l);
205 rset_delete (info->rset_m);
206 rset_delete (info->rset_r);
208 rset_delete (info->rset_attr);
212 static void r_rewind_between (RSFD rfd)
214 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
215 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
218 logf (LOG_DEBUG, "rsbetween_rewind");
220 rset_rewind (info->rset_l, p->rfd_l);
221 rset_rewind (info->rset_m, p->rfd_m);
222 rset_rewind (info->rset_r, p->rfd_r);
223 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l);
224 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
225 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r);
228 rset_rewind (info->rset_attr, p->rfd_attr);
229 p->more_attr = rset_read (info->rset_attr, p->rfd_attr, p->buf_attr);
237 static int r_forward_between(RSET ct, RSFD rfd, void *buf,
238 int (*cmpfunc)(const void *p1, const void *p2),
239 const void *untilbuf)
241 struct rset_between_info *info = ((struct rset_between_rfd*)rfd)->info;
242 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
245 log2( p, "fwd: before forward", 0,0);
247 /* It is enough to forward the m pointer here, the read will */
248 /* naturally forward the l, m, and attr pointers */
250 p->more_m=rset_forward(info->rset_m,p->rfd_m, p->buf_m,
253 log2( p, "fwd: after forward M", 0,0);
255 rc = r_read_between(rfd, buf);
257 log2( p, "fwd: after forward", 0,0);
265 static int r_read_between (RSFD rfd, void *buf)
267 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
268 struct rset_between_info *info = p->info;
276 log2( p, "start of loop", cmp_l, cmp_r);
279 /* forward L until past m, count levels, note rec boundaries */
281 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
285 cmp_l=2; /* past this record */
288 log2( p, "after first L", cmp_l, cmp_r);
291 while (cmp_l < 0) /* l before m */
294 p->level=0; /* earlier record */
297 p->level++; /* relevant start tag */
299 if (!info->rset_attr)
307 cmp_attr = (*info->cmp)(p->buf_attr, p->buf_l);
313 else if (cmp_attr > 0)
315 else if (cmp_attr==-1)
316 p->more_attr = rset_read (info->rset_attr,
317 p->rfd_attr, p->buf_attr);
318 /* if we had a forward that went all the way to
319 * the seqno, we could use that. But fwd only goes
321 else if (cmp_attr==-2)
323 p->more_attr = rset_forward(
324 info->rset_attr, p->rfd_attr,
325 p->buf_attr, info->cmp, p->buf_l);
327 logf(LOG_DEBUG, "btw: after frowarding attr m=%d",p->more_attr);
330 } /* while more_attr */
339 p->more_l=rset_forward( info->rset_l, p->rfd_l,
340 p->buf_l, info->cmp, p->buf_m);
342 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
346 log2( p, "after forwarding L", cmp_l, cmp_r);
351 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l);
354 p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l);
358 cmp_l= (*info->cmp)(p->buf_l, p->buf_m);
363 log2( p, "end of L loop", cmp_l, cmp_r);
368 /* forward R until past m, count levels */
370 log2( p, "Before moving R", cmp_l, cmp_r);
373 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
377 log2( p, "after first R", cmp_l, cmp_r);
379 while (cmp_r < 0) /* r before m */
381 /* -2, earlier record, don't count level */
383 p->level--; /* relevant end tag */
389 p->more_r=rset_forward( info->rset_r, p->rfd_r,
390 p->buf_r, info->cmp, p->buf_m);
393 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r);
396 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
399 p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r);
400 cmp_r= (*info->cmp)(p->buf_r, p->buf_m);
406 log2( p, "End of R loop", cmp_l, cmp_r);
410 if ( ( p->level <= 0 ) && ! p->more_l)
411 return 0; /* no more start tags, nothing more to find */
413 if ( attr_match && p->level > 0) /* within a tag pair (or deeper) */
415 memcpy (buf, p->buf_m, info->key_size);
417 log2( p, "Returning a hit (and forwarding m)", cmp_l, cmp_r);
419 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
425 else if ( ! p->more_l ) /* not in data, no more starts */
428 log2( p, "no more starts, exiting without a hit", cmp_l, cmp_r);
430 return 0; /* ergo, nothing can be found. stop scanning */
436 p->more_m=rset_forward( info->rset_m, p->rfd_m,
437 p->buf_m, info->cmp, p->buf_l);
440 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
445 p->more_m = rset_read (info->rset_m, p->rfd_m, p->buf_m);
448 log2( p, "End of M loop", cmp_l, cmp_r);
453 log2( p, "Exiting, nothing more in m", cmp_l, cmp_r);
455 return 0; /* no more data possible */
461 static int r_write_between (RSFD rfd, const void *buf)
463 logf (LOG_FATAL, "between set type is read-only");
468 static void r_pos_between (RSFD rfd, double *current, double *total)
470 struct rset_between_rfd *p = (struct rset_between_rfd *) rfd;
471 struct rset_between_info *info = p->info;
477 rset_pos(info->rset_l, p->rfd_l, &lcur, <ot);
478 rset_pos(info->rset_m, p->rfd_m, &mcur, &mtot);
479 rset_pos(info->rset_r, p->rfd_r, &rcur, &rtot);
480 if ( (ltot<0) && (mtot<0) && (rtot<0) ) { /*no position */
481 *current=mcur; /* return same as you got */
482 *total=mtot; /* probably -1 for not available */
484 if ( ltot<0) { ltot=0; lcur=0;} /* if only one useful, use it */
485 if ( mtot<0) { mtot=0; mcur=0;}
486 if ( rtot<0) { rtot=0; rcur=0;}
487 if ( ltot+mtot+rtot < 1 ) { /* empty rset */
492 r=1.0*(lcur+mcur+rcur)/(ltot+mtot+rtot); /* weighed average of l and r */
496 yaz_log(LOG_DEBUG,"betw_pos: (%s/%s) %0.1f/%0.1f= %0.4f ",
497 info->rset_l->control->desc, info->rset_r->control->desc,
498 *current, *total, r);