X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=index%2Finvstat.c;h=ca44bda61b1bb6247bffb998be591356517a7e44;hb=175aced6e1fba324771b361822e8a27a38e567bb;hp=f4b548c59f451b65a66548313c1c99db28ee7e09;hpb=3e72747f4b28856e5f684dbe8984f8fb28680e62;p=idzebra-moved-to-github.git

diff --git a/index/invstat.c b/index/invstat.c
index f4b548c..ca44bda 100644
--- a/index/invstat.c
+++ b/index/invstat.c
@@ -20,6 +20,9 @@ struct inv_stat_info {
     int isam_bounds[20];
     int isam_occurrences[20];
     char tmp[128];
+    int isamb_levels[10][5];
+    int isamb_sizes[10];
+    int isamb_blocks[10];
 };
 
 #define SINGLETON_TYPE 8 /* the type to use for singletons that */ 
@@ -116,11 +119,21 @@ static int inv_stat_handle (char *name, const char *info, int pos,
     {
         ISAMB_PP pp;
         struct it_key key;
+        int cat = isam_p & 3;
+        int level;
+        int size;
+        int blocks;
+        
+        pp = isamb_pp_open_x(stat_info->zh->reg->isamb, isam_p, &level);
 
-        pp = isamb_pp_open(stat_info->zh->reg->isamb, isam_p);
         while (isamb_pp_read(pp, &key))
             occur++;
-        isamb_pp_close (pp);
+        isamb_pp_close_x (pp, &size, &blocks);
+        stat_info->isamb_blocks[cat] += blocks;
+        stat_info->isamb_sizes[cat] += size;
+        if (level > 4)
+            level = 4;
+        stat_info->isamb_levels[cat][level] ++;
     }
 
     while (occur > stat_info->isam_bounds[i] && stat_info->isam_bounds[i])
@@ -175,6 +188,15 @@ void zebra_register_statistics (ZebraHandle zh)
     for (i = 0; i<20; i++)
         stat_info.isam_occurrences[i] = 0;
 
+    for (i = 0; i<10; i++)
+    {
+        int j;
+        for (j = 0; j<5; j++)
+            stat_info.isamb_levels[i][j] = 0;
+        stat_info.isamb_sizes[i] = 0;
+        stat_info.isamb_blocks[i] = 0;
+    }
+
     dict_scan (zh->reg->dict, term_dict, &before, &after, &stat_info,
                inv_stat_handle);
 
@@ -230,6 +252,28 @@ void zebra_register_statistics (ZebraHandle zh)
     if ( (zh->reg->isamd) && (zh->reg->isamd->method->debug>0))
         fprintf (stderr, "\n%d words using %d bytes\n",
              stat_info.no_dict_entries, stat_info.no_dict_bytes);
+
+    if (zh->reg->isamb)
+    {
+        for (i = 0; i<4; i++)
+        {
+            int j;
+            int bsize = isamb_block_info(zh->reg->isamb, i);
+            if (bsize < 0)
+                break;
+            fprintf (stderr, "Category   %d\n", i);
+            fprintf (stderr, "Block size %d\n", bsize);
+            fprintf (stderr, "Blocks:    %d\n", stat_info.isamb_blocks[i]);
+            fprintf (stderr, "Size:      %d\n", stat_info.isamb_sizes[i]);
+            fprintf (stderr, "Total      %d\n", stat_info.isamb_blocks[i]*
+                     bsize);
+            for (j = 0; j<5; j++)
+                if (stat_info.isamb_levels[i][j])
+                    fprintf (stderr, "Level%d     %d\n", j,
+                             stat_info.isamb_levels[i][j]);
+        }
+    }
+
     fprintf (stderr, "    Occurrences     Words\n");
     prev = 1;
     for (i = 0; stat_info.isam_bounds[i]; i++)
@@ -249,7 +293,13 @@ void zebra_register_statistics (ZebraHandle zh)
 /*
  *
  * $Log: invstat.c,v $
- * Revision 1.25  2002-04-26 08:44:47  adam
+ * Revision 1.27  2002-04-30 08:28:37  adam
+ * isamb fixes for pp_read. Statistics
+ *
+ * Revision 1.26  2002/04/29 18:03:46  adam
+ * More isamb statistics
+ *
+ * Revision 1.25  2002/04/26 08:44:47  adam
  * Index statistics working again
  *
  * Revision 1.24  2002/04/05 08:46:26  adam