X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=rset%2Frsbool.c;h=380ef6a7ba521750dc97097757de39e39bfc80c8;hb=2e4e9c6def27f1e1463dcb6f205fab6a98054f38;hp=c9c254a5cc77fb0e8de12b6e310997f20dd3d98e;hpb=40ca1d08c1d83b92e0b90951d918b252c61c63b2;p=idzebra-moved-to-github.git diff --git a/rset/rsbool.c b/rset/rsbool.c index c9c254a..380ef6a 100644 --- a/rset/rsbool.c +++ b/rset/rsbool.c @@ -1,105 +1,109 @@ -/* - * Copyright (C) 1994-1995, Index Data I/S - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: rsbool.c,v $ - * Revision 1.6 1995-10-06 14:38:05 adam - * New result set method: r_score. - * Local no (sysno) and score is transferred to retrieveCtrl. - * - * Revision 1.5 1995/09/08 14:52:41 adam - * Work on relevance feedback. - * - * Revision 1.4 1995/09/08 08:54:04 adam - * More efficient and operation. - * - * Revision 1.3 1995/09/07 13:58:43 adam - * New parameter: result-set file descriptor (RSFD) to support multiple - * positions within the same result-set. - * Boolean operators: and, or, not implemented. - * - * Revision 1.2 1995/09/06 16:11:55 adam - * More work on boolean sets. - * - * Revision 1.1 1995/09/06 13:27:15 adam - * New set type: bool. Not finished yet. - * - */ +/* $Id: rsbool.c,v 1.32 2004-08-03 14:54:41 heikki Exp $ + Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004 + Index Data Aps + +This file is part of the Zebra server. + +Zebra is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free +Software Foundation; either version 2, or (at your option) any later +version. + +Zebra is distributed in the hope that it will be useful, but WITHOUT ANY +WARRANTY; without even the implied warranty of MERCHANTABILITY or +FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +for more details. + +You should have received a copy of the GNU General Public License +along with Zebra; see the file LICENSE.zebra. If not, write to the +Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA +02111-1307, USA. +*/ #include +#include +#include #include #include -#include +#include -static rset_control *r_create(const struct rset_control *sel, void *parms); -static RSFD r_open (rset_control *ct, int wflag); +#ifndef RSET_DEBUG +#define RSET_DEBUG 0 +#endif + +static void *r_create(RSET ct, const struct rset_control *sel, void *parms); +static RSFD r_open (RSET ct, int flag); static void r_close (RSFD rfd); -static void r_delete (rset_control *ct); +static void r_delete (RSET ct); static void r_rewind (RSFD rfd); -static int r_count (rset_control *ct); -static int r_read_and (RSFD rfd, void *buf); -static int r_read_or (RSFD rfd, void *buf); -static int r_read_not (RSFD rfd, void *buf); +static int r_forward(RSET ct, RSFD rfd, void *buf, int *term_index, + int (*cmpfunc)(const void *p1, const void *p2), + const void *untilbuf); +/* static void r_pos (RSFD rfd, int *current, int *total); */ +static int r_read_and (RSFD rfd, void *buf, int *term_index); +static int r_read_or (RSFD rfd, void *buf, int *term_index); +static int r_read_not (RSFD rfd, void *buf, int *term_index); static int r_write (RSFD rfd, const void *buf); -static int r_score (RSFD rfd, int *score); -static const rset_control control_and = +static const struct rset_control control_and = { - "AND set type", - 0, + "and", r_create, r_open, r_close, r_delete, r_rewind, - r_count, + r_forward, /* rset_default_forward, */ + rset_default_pos, r_read_and, r_write, - r_score }; -static const rset_control control_or = +static const struct rset_control control_or = { - "OR set type", - 0, + "or", r_create, r_open, r_close, r_delete, r_rewind, - r_count, +#if 1 + r_forward, +#else + rset_default_forward, +#endif + rset_default_pos, r_read_or, r_write, - r_score }; -static const rset_control control_not = +static const struct rset_control control_not = { - "NOT set type", - 0, + "not", r_create, r_open, r_close, r_delete, r_rewind, - r_count, + r_forward, + rset_default_pos, r_read_not, r_write, - r_score }; -const rset_control *rset_kind_and = &control_and; -const rset_control *rset_kind_or = &control_or; -const rset_control *rset_kind_not = &control_not; +const struct rset_control *rset_kind_and = &control_and; +const struct rset_control *rset_kind_or = &control_or; +const struct rset_control *rset_kind_not = &control_not; struct rset_bool_info { int key_size; RSET rset_l; RSET rset_r; + int term_index_s; int (*cmp)(const void *p1, const void *p2); + void (*log_item)(int logmask, const void *p, const char *txt); struct rset_bool_rfd *rfd_list; }; @@ -108,51 +112,69 @@ struct rset_bool_rfd { RSFD rfd_r; int more_l; int more_r; + int term_index_l; + int term_index_r; void *buf_l; void *buf_r; + int tail; struct rset_bool_rfd *next; struct rset_bool_info *info; }; -static rset_control *r_create (const struct rset_control *sel, void *parms) +static void *r_create (RSET ct, const struct rset_control *sel, void *parms) { - rset_control *newct; - rset_bool_parms *bool_parms = parms; + rset_bool_parms *bool_parms = (rset_bool_parms *) parms; struct rset_bool_info *info; - newct = xmalloc(sizeof(*newct)); - memcpy (newct, sel, sizeof(*sel)); - newct->buf = xmalloc (sizeof(struct rset_bool_info)); - info = (struct rset_bool_info*) newct->buf; + info = (struct rset_bool_info *) xmalloc (sizeof(*info)); info->key_size = bool_parms->key_size; info->rset_l = bool_parms->rset_l; info->rset_r = bool_parms->rset_r; + if (rset_is_volatile(info->rset_l) || rset_is_volatile(info->rset_r)) + ct->flags |= RSET_FLAG_VOLATILE; info->cmp = bool_parms->cmp; + info->log_item = bool_parms->log_item; info->rfd_list = NULL; - return newct; + + info->term_index_s = info->rset_l->no_rset_terms; + ct->no_rset_terms = + info->rset_l->no_rset_terms + info->rset_r->no_rset_terms; + ct->rset_terms = (RSET_TERM *) + xmalloc (sizeof (*ct->rset_terms) * ct->no_rset_terms); + + memcpy (ct->rset_terms, info->rset_l->rset_terms, + info->rset_l->no_rset_terms * sizeof(*ct->rset_terms)); + memcpy (ct->rset_terms + info->rset_l->no_rset_terms, + info->rset_r->rset_terms, + info->rset_r->no_rset_terms * sizeof(*ct->rset_terms)); + return info; } -static RSFD r_open (rset_control *ct, int wflag) +static RSFD r_open (RSET ct, int flag) { - struct rset_bool_info *info = ct->buf; + struct rset_bool_info *info = (struct rset_bool_info *) ct->buf; struct rset_bool_rfd *rfd; - if (wflag) + if (flag & RSETF_WRITE) { logf (LOG_FATAL, "bool set type is read-only"); return NULL; } - rfd = xmalloc (sizeof(*rfd)); + rfd = (struct rset_bool_rfd *) xmalloc (sizeof(*rfd)); + logf(LOG_DEBUG,"rsbool (%s) open [%p]", ct->control->desc, rfd); rfd->next = info->rfd_list; info->rfd_list = rfd; + rfd->info = info; rfd->buf_l = xmalloc (info->key_size); rfd->buf_r = xmalloc (info->key_size); - rfd->rfd_l = rset_open (info->rset_l, wflag); - rfd->rfd_r = rset_open (info->rset_r, wflag); - rfd->more_l = rset_read (info->rset_l, rfd->rfd_l, rfd->buf_l); - rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r); - rfd->info = info; + rfd->rfd_l = rset_open (info->rset_l, RSETF_READ); + rfd->rfd_r = rset_open (info->rset_r, RSETF_READ); + rfd->more_l = rset_read (info->rset_l, rfd->rfd_l, rfd->buf_l, + &rfd->term_index_l); + rfd->more_r = rset_read (info->rset_r, rfd->rfd_r, rfd->buf_r, + &rfd->term_index_r); + rfd->tail = 0; return rfd; } @@ -169,79 +191,260 @@ static void r_close (RSFD rfd) rset_close (info->rset_l, (*rfdp)->rfd_l); rset_close (info->rset_r, (*rfdp)->rfd_r); *rfdp = (*rfdp)->next; - free (rfd); + xfree (rfd); return; } logf (LOG_FATAL, "r_close but no rfd match!"); assert (0); } -static void r_delete (rset_control *ct) +static void r_delete (RSET ct) { - struct rset_bool_info *info = ct->buf; + struct rset_bool_info *info = (struct rset_bool_info *) ct->buf; assert (info->rfd_list == NULL); + xfree (ct->rset_terms); + rset_delete (info->rset_l); + rset_delete (info->rset_r); xfree (info); - xfree (ct); } static void r_rewind (RSFD rfd) { struct rset_bool_info *info = ((struct rset_bool_rfd*)rfd)->info; - struct rset_bool_rfd *p = rfd; + struct rset_bool_rfd *p = (struct rset_bool_rfd *) rfd; logf (LOG_DEBUG, "rsbool_rewind"); rset_rewind (info->rset_l, p->rfd_l); rset_rewind (info->rset_r, p->rfd_r); - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, &p->term_index_l); + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, &p->term_index_r); } -static int r_count (rset_control *ct) +static int r_forward (RSET ct, RSFD rfd, void *buf, int *term_index, + int (*cmpfunc)(const void *p1, const void *p2), + const void *untilbuf) { - return 0; + struct rset_bool_info *info = ((struct rset_bool_rfd*)rfd)->info; + struct rset_bool_rfd *p = (struct rset_bool_rfd *) rfd; + int rc; + +#if RSET_DEBUG + logf (LOG_DEBUG, "rsbool_forward (L) [%p] '%s' (ct=%p rfd=%p m=%d,%d)", + rfd, ct->control->desc, ct, rfd, p->more_l, p->more_r); +#endif + if ( p->more_l && ((cmpfunc)(untilbuf,p->buf_l)==2) ) + p->more_l = rset_forward(info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l, info->cmp, untilbuf); +#if RSET_DEBUG + logf (LOG_DEBUG, "rsbool_forward (R) [%p] '%s' (ct=%p rfd=%p m=%d,%d)", + rfd, ct->control->desc, ct, rfd, p->more_l, p->more_r); +#endif + if ( p->more_r && ((cmpfunc)(untilbuf,p->buf_r)==2)) + p->more_r = rset_forward(info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r, info->cmp, untilbuf); +#if RSET_DEBUG + logf (LOG_DEBUG, "rsbool_forward [%p] calling read, m=%d,%d t=%d", + rfd, p->more_l, p->more_r, p->tail); +#endif + + p->tail=0; + rc = rset_read(ct,rfd,buf,term_index); +#if RSET_DEBUG + logf (LOG_DEBUG, "rsbool_forward returning [%p] %d m=%d,%d", + rfd, rc, p->more_l, p->more_r); +#endif + return rc; } -static int r_read_and (RSFD rfd, void *buf) + +/* + 1,1 1,3 + 1,9 2,1 + 1,11 3,1 + 2,9 + + 1,1 1,1 + 1,3 1,3 + 1,9 + 1,11 + 2,1 2,1 + 2,9 + 3,1 +*/ + +static int r_read_and (RSFD rfd, void *buf, int *term_index) { - struct rset_bool_rfd *p = rfd; + struct rset_bool_rfd *p = (struct rset_bool_rfd *) rfd; struct rset_bool_info *info = p->info; - while (p->more_l && p->more_r) + while (p->more_l || p->more_r) { int cmp; - cmp = (*info->cmp)(p->buf_l, p->buf_r); + if (p->more_l && p->more_r) + cmp = (*info->cmp)(p->buf_l, p->buf_r); + else if (p->more_l) + cmp = -2; + else + cmp = 2; +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] looping: m=%d/%d c=%d t=%d", + rfd, p->more_l, p->more_r, cmp, p->tail); + (*info->log_item)(LOG_DEBUG, p->buf_l, "left "); + (*info->log_item)(LOG_DEBUG, p->buf_r, "right "); +#endif if (!cmp) { memcpy (buf, p->buf_l, info->key_size); - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); - return 1; + *term_index = p->term_index_l; + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); + p->tail = 1; } else if (cmp == 1) { memcpy (buf, p->buf_r, info->key_size); - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); + *term_index = p->term_index_r + info->term_index_s; + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r); + p->tail = 1; +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] returning R m=%d/%d c=%d", + rfd, p->more_l, p->more_r, cmp); + key_logdump(LOG_DEBUG,buf); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif return 1; } else if (cmp == -1) { memcpy (buf, p->buf_l, info->key_size); - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); + *term_index = p->term_index_l; + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); + p->tail = 1; +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] returning L m=%d/%d c=%d", + rfd, p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif return 1; } - else if (cmp > 1) - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); - else - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); + else if (cmp > 1) /* cmp == 2 */ + { +#define OLDCODE 0 +#if OLDCODE + memcpy (buf, p->buf_r, info->key_size); + *term_index = p->term_index_r + info->term_index_s; + + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r); + if (p->tail) + { + if (!p->more_r || (*info->cmp)(p->buf_r, buf) > 1) + p->tail = 0; +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and returning C m=%d/%d c=%d", + p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif + return 1; + } +#else + + if (p->tail) + { + memcpy (buf, p->buf_r, info->key_size); + *term_index = p->term_index_r + info->term_index_s; + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r); + if (!p->more_r || (*info->cmp)(p->buf_r, buf) > 1) + p->tail = 0; +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] returning R tail m=%d/%d c=%d", + rfd, p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif + return 1; + } + else + { +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] about to forward R m=%d/%d c=%d", + rfd, p->more_l, p->more_r, cmp); +#endif + if (p->more_r && p->more_l) + p->more_r = rset_forward( + info->rset_r, p->rfd_r, + p->buf_r, &p->term_index_r, + (info->cmp), p->buf_l); + else + return 0; /* no point in reading further */ + } +#endif + } + else /* cmp == -2 */ + { +#if OLDCODE + memcpy (buf, p->buf_l, info->key_size); + *term_index = p->term_index_l; + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); + if (p->tail) + { + if (!p->more_l || (*info->cmp)(p->buf_l, buf) > 1) + p->tail = 0; +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] returning R tail m=%d/%d c=%d", + rfd, p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif + return 1; + } +#else + if (p->tail) + { + memcpy (buf, p->buf_l, info->key_size); + *term_index = p->term_index_l; + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); + if (!p->more_l || (*info->cmp)(p->buf_l, buf) > 1) + p->tail = 0; +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] returning L tail m=%d/%d c=%d", + rfd, p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif + return 1; + } + else + { +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] about to forward L m=%d/%d c=%d", + rfd, p->more_l, p->more_r, cmp); +#endif + if (p->more_r && p->more_l) + p->more_l = rset_forward( + /* p->more_l = rset_default_forward( */ + info->rset_l, p->rfd_l, + p->buf_l, &p->term_index_l, + (info->cmp), p->buf_r); + else + return 0; /* no point in reading further */ + } +#endif + } } +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_and [%p] reached its end",rfd); +#endif return 0; } -static int r_read_or (RSFD rfd, void *buf) +static int r_read_or (RSFD rfd, void *buf, int *term_index) { - struct rset_bool_rfd *p = rfd; + struct rset_bool_rfd *p = (struct rset_bool_rfd *) rfd; struct rset_bool_info *info = p->info; while (p->more_l || p->more_r) @@ -257,29 +460,51 @@ static int r_read_or (RSFD rfd, void *buf) if (!cmp) { memcpy (buf, p->buf_l, info->key_size); - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); + *term_index = p->term_index_l; + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r); +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_or returning A m=%d/%d c=%d", + p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif return 1; } else if (cmp > 0) { memcpy (buf, p->buf_r, info->key_size); - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); + *term_index = p->term_index_r + info->term_index_s; + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r); +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_or returning B m=%d/%d c=%d", + p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif return 1; } else { memcpy (buf, p->buf_l, info->key_size); - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); + *term_index = p->term_index_l; + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); +#if RSET_DEBUG + logf (LOG_DEBUG, "r_read_or returning C m=%d/%d c=%d", + p->more_l, p->more_r, cmp); + (*info->log_item)(LOG_DEBUG, buf, ""); +#endif return 1; } } return 0; } -static int r_read_not (RSFD rfd, void *buf) +static int r_read_not (RSFD rfd, void *buf, int *term_index) { - struct rset_bool_rfd *p = rfd; + struct rset_bool_rfd *p = (struct rset_bool_rfd *) rfd; struct rset_bool_info *info = p->info; while (p->more_l || p->more_r) @@ -295,24 +520,38 @@ static int r_read_not (RSFD rfd, void *buf) if (cmp < -1) { memcpy (buf, p->buf_l, info->key_size); - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); + *term_index = p->term_index_l; + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); return 1; } else if (cmp > 1) - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); + { +#if 0 + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r); +#else + p->more_r = rset_forward( + info->rset_r, p->rfd_r, + p->buf_r, &p->term_index_r, + (info->cmp), p->buf_l); +#endif + } else { memcpy (buf, p->buf_l, info->key_size); do - { - p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l); + { + p->more_l = rset_read (info->rset_l, p->rfd_l, p->buf_l, + &p->term_index_l); if (!p->more_l) break; cmp = (*info->cmp)(p->buf_l, buf); } while (cmp >= -1 && cmp <= 1); do { - p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r); + p->more_r = rset_read (info->rset_r, p->rfd_r, p->buf_r, + &p->term_index_r); if (!p->more_r) break; cmp = (*info->cmp)(p->buf_r, buf); @@ -329,9 +568,3 @@ static int r_write (RSFD rfd, const void *buf) return -1; } -static int r_score (RSFD rfd, int *score) -{ - *score = -1; - return -1; -} -