X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=index%2Frank1.c;h=21177046682d0cbc0f99039e8fb4568d3398acaa;hp=89f845560c48eb2bd8f2b4cbc5ffaf12c8bd33d9;hb=2b99f08490d3bd763bb086693044a6db10c0bcf9;hpb=2b1851bd5565e3d21f9cf9a37661a584c063b75f diff --git a/index/rank1.c b/index/rank1.c index 89f8455..2117704 100644 --- a/index/rank1.c +++ b/index/rank1.c @@ -1,35 +1,26 @@ -/* - * Copyright (C) 1998-1999, Index Data - * All rights reserved. - * Sebastian Hammer, Adam Dickmeiss - * - * $Log: rank1.c,v $ - * Revision 1.8 2002-04-04 14:14:13 adam - * Multiple registers (alpha early) - * - * Revision 1.7 2001/11/14 22:06:27 adam - * Rank-weight may be controlled via query. - * - * Revision 1.6 2000/03/15 15:00:30 adam - * First work on threaded version. - * - * Revision 1.5 1999/05/26 07:49:13 adam - * C++ compilation. - * - * Revision 1.4 1999/02/02 14:51:01 adam - * Updated WIN32 code specific sections. Changed header. - * - * Revision 1.3 1998/06/12 12:21:53 adam - * Fixed memory-leak. - * - * Revision 1.2 1998/03/05 13:03:29 adam - * Improved ranking. - * - * Revision 1.1 1998/03/05 08:45:12 adam - * New result set model and modular ranking system. Moved towards - * descent server API. System information stored as "SGML" records. - * - */ +/* $Id: rank1.c,v 1.16 2004-08-06 13:36:23 adam Exp $ + Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003 + 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 @@ -39,6 +30,8 @@ #include #endif +#define DEBUG_RANK 0 + #include "index.h" struct rank_class_info { @@ -47,7 +40,7 @@ struct rank_class_info { struct rank_term_info { int local_occur; - int global_occur; + zint global_occur; int global_inv; int rank_flag; int rank_weight; @@ -72,12 +65,12 @@ static int log2_int (unsigned g) * create: Creates/Initialises this rank handler. This routine is * called exactly once. The routine returns the class_handle. */ -static void *create (struct zebra_register *reg) +static void *create (ZebraHandle zh) { struct rank_class_info *ci = (struct rank_class_info *) xmalloc (sizeof(*ci)); - logf (LOG_DEBUG, "rank-1 create"); + yaz_log (LOG_DEBUG, "rank-1 create"); return ci; } @@ -90,7 +83,7 @@ static void destroy (struct zebra_register *reg, void *class_handle) { struct rank_class_info *ci = (struct rank_class_info *) class_handle; - logf (LOG_DEBUG, "rank-1 destroy"); + yaz_log (LOG_DEBUG, "rank-1 destroy"); xfree (ci); } @@ -105,20 +98,31 @@ static void *begin (struct zebra_register *reg, void *class_handle, RSET rset) struct rank_set_info *si = (struct rank_set_info *) xmalloc (sizeof(*si)); int i; - logf (LOG_DEBUG, "rank-1 begin"); +#if DEBUG_RANK + yaz_log (LOG_LOG, "rank-1 begin"); +#endif si->no_entries = rset->no_rset_terms; si->no_rank_entries = 0; si->entries = (struct rank_term_info *) xmalloc (sizeof(*si->entries)*si->no_entries); for (i = 0; i < si->no_entries; i++) { - int g = rset->rset_terms[i]->nn; + zint g = rset->rset_terms[i]->nn; +#if DEBUG_RANK + yaz_log(LOG_LOG, "i=%d flags=%s", i, rset->rset_terms[i]->flags); +#endif if (!strncmp (rset->rset_terms[i]->flags, "rank,", 5)) { - yaz_log (LOG_LOG, "%s", rset->rset_terms[i]->flags); + const char *cp = strstr(rset->rset_terms[i]->flags+4, ",w="); si->entries[i].rank_flag = 1; - si->entries[i].rank_weight = atoi (rset->rset_terms[i]->flags+5); - yaz_log (LOG_LOG, "i=%d weight=%d", i, si->entries[i].rank_weight); + if (cp) + si->entries[i].rank_weight = atoi (cp+3); + else + si->entries[i].rank_weight = 34; +#if DEBUG_RANK + yaz_log (LOG_LOG, " i=%d weight=%d", i, + si->entries[i].rank_weight); +#endif (si->no_rank_entries)++; } else @@ -126,7 +130,7 @@ static void *begin (struct zebra_register *reg, void *class_handle, RSET rset) si->entries[i].local_occur = 0; si->entries[i].global_occur = g; si->entries[i].global_inv = 32 - log2_int (g); - logf (LOG_DEBUG, "-------- %d ------", 32 - log2_int (g)); + yaz_log (LOG_DEBUG, " global_inv = %d g = " ZINT_FORMAT, (int) (32-log2_int (g)), g); } return si; } @@ -138,7 +142,7 @@ static void *begin (struct zebra_register *reg, void *class_handle, RSET rset) static void end (struct zebra_register *reg, void *set_handle) { struct rank_set_info *si = (struct rank_set_info *) set_handle; - logf (LOG_DEBUG, "rank-1 end"); + yaz_log (LOG_DEBUG, "rank-1 end"); xfree (si->entries); xfree (si); } @@ -151,7 +155,9 @@ static void end (struct zebra_register *reg, void *set_handle) static void add (void *set_handle, int seqno, int term_index) { struct rank_set_info *si = (struct rank_set_info *) set_handle; - logf (LOG_DEBUG, "rank-1 add seqno=%d term_index=%d", seqno, term_index); +#if DEBUG_RANK + yaz_log (LOG_LOG, "rank-1 add seqno=%d term_index=%d", seqno, term_index); +#endif si->last_pos = seqno; si->entries[term_index].local_occur++; } @@ -162,22 +168,32 @@ static void add (void *set_handle, int seqno, int term_index) * score should be between 0 and 1000. If score cannot be obtained * -1 should be returned. */ -static int calc (void *set_handle, int sysno) +static int calc (void *set_handle, zint sysno) { int i, lo, divisor, score = 0; struct rank_set_info *si = (struct rank_set_info *) set_handle; - logf (LOG_DEBUG, "rank-1 calc sysno=%d", sysno); - if (!si->no_rank_entries) return -1; + +#if DEBUG_RANK + yaz_log(LOG_LOG, "calc"); +#endif for (i = 0; i < si->no_entries; i++) + { +#if DEBUG_RANK + yaz_log(LOG_LOG, "i=%d rank_flag=%d lo=%d", + i, si->entries[i].rank_flag, si->entries[i].local_occur); +#endif if (si->entries[i].rank_flag && (lo = si->entries[i].local_occur)) score += (8+log2_int (lo)) * si->entries[i].global_inv * si->entries[i].rank_weight; + } divisor = si->no_rank_entries * (8+log2_int (si->last_pos/si->no_entries)); score = score / divisor; - yaz_log (LOG_LOG, "score=%d", score); +#if DEBUG_RANK + yaz_log (LOG_LOG, "sysno=" ZINT_FORMAT " score=%d", sysno, score); +#endif if (score > 1000) score = 1000; for (i = 0; i < si->no_entries; i++)