Multiple registers (alpha early)

[idzebra-moved-to-github.git] / index / rank1.c

diff --git a/index/rank1.c b/index/rank1.c
index 3369c79..89f8455 100644 (file)
--- a/index/rank1.c
+++ b/index/rank1.c
@@ -1,10 +1,31 @@
 /*
- * Copyright (C) 1998, Index Data I/S
+ * Copyright (C) 1998-1999, Index Data
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: rank1.c,v $
- * Revision 1.1  1998-03-05 08:45:12  adam
+ * 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.
  *
@@ -12,13 +33,13 @@
 
 #include <stdio.h>
 #include <assert.h>
-#ifdef WINDOWS
+#ifdef WIN32
 #include <io.h>
 #else
 #include <unistd.h>
 #endif
 
-#include "zserver.h"
+#include "index.h"
 
 struct rank_class_info {
     int dummy;
@@ -29,6 +50,7 @@ struct rank_term_info {
     int global_occur;
     int global_inv;
     int rank_flag;
+    int rank_weight;
 };
 
 struct rank_set_info {
@@ -50,9 +72,10 @@ 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 (ZebraHandle zh)
+static void *create (struct zebra_register *reg)
 {
-    struct rank_class_info *ci = xmalloc (sizeof(*ci));
+    struct rank_class_info *ci = (struct rank_class_info *)
+       xmalloc (sizeof(*ci));
 
     logf (LOG_DEBUG, "rank-1 create");
     return ci;
@@ -63,9 +86,9 @@ static void *create (ZebraHandle zh)
  *  when the handler is no longer needed - i.e. when the server
  *  dies. The class_handle was previously returned by create.
  */
-static void destroy (ZebraHandle zh, void *class_handle)
+static void destroy (struct zebra_register *reg, void *class_handle)
 {
-    struct rank_class_info *ci = class_handle;
+    struct rank_class_info *ci = (struct rank_class_info *) class_handle;
 
     logf (LOG_DEBUG, "rank-1 destroy");
     xfree (ci);
@@ -77,21 +100,25 @@ static void destroy (ZebraHandle zh, void *class_handle)
  *  each result set. The returned handle is a "set handle" and
  *  will be used in each of the handlers below.
  */
-static void *begin (ZebraHandle zh, void *class_handle, RSET rset)
+static void *begin (struct zebra_register *reg, void *class_handle, RSET rset)
 {
-    struct rank_set_info *si = xmalloc (sizeof(*si));
+    struct rank_set_info *si = (struct rank_set_info *) xmalloc (sizeof(*si));
     int i;
 
     logf (LOG_DEBUG, "rank-1 begin");
     si->no_entries = rset->no_rset_terms;
     si->no_rank_entries = 0;
-    si->entries = xmalloc (sizeof(*si->entries)*si->no_entries);
+    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;
-       if (!strcmp (rset->rset_terms[i]->flags, "rank"))
+       if (!strncmp (rset->rset_terms[i]->flags, "rank,", 5))
        {
+            yaz_log (LOG_LOG, "%s", rset->rset_terms[i]->flags);
            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);
            (si->no_rank_entries)++;
        }
        else
@@ -108,10 +135,11 @@ static void *begin (ZebraHandle zh, void *class_handle, RSET rset)
  * end: Terminates ranking process. Called after a result set
  *  has been ranked.
  */
-static void end (ZebraHandle zh, void *set_handle)
+static void end (struct zebra_register *reg, void *set_handle)
 {
-    struct rank_set_info *si = set_handle;
+    struct rank_set_info *si = (struct rank_set_info *) set_handle;
     logf (LOG_DEBUG, "rank-1 end");
+    xfree (si->entries);
     xfree (si);
 }
 
@@ -122,7 +150,7 @@ static void end (ZebraHandle zh, void *set_handle)
  */
 static void add (void *set_handle, int seqno, int term_index)
 {
-    struct rank_set_info *si = set_handle;
+    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);
     si->last_pos = seqno;
     si->entries[term_index].local_occur++;
@@ -136,19 +164,22 @@ static void add (void *set_handle, int seqno, int term_index)
  */
 static int calc (void *set_handle, int sysno)
 {
-    int i, lu, score = 0;
-    struct rank_set_info *si = set_handle;
+    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;
     for (i = 0; i < si->no_entries; i++)
-       if (si->entries[i].rank_flag && (lu = si->entries[i].local_occur))
-           score += (2+log2_int (lu)) * si->entries[i].global_inv;
-    logf (LOG_DEBUG, " dividend=%d", 60*score);
-    logf (LOG_DEBUG, " divisor=%d", si->no_rank_entries * log2_int (4+si->last_pos));
-    score = (60 * score)/(si->no_rank_entries * log2_int (4+si->last_pos));
+       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 (score > 1000)
+       score = 1000;
     for (i = 0; i < si->no_entries; i++)
        si->entries[i].local_occur = 0;
     return score;