X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=dict%2Fdicttest.c;h=b74b795daa4f5bc7df4b869af5de38a7d72aebe3;hb=cfecd7d576e8de7cfa19751b8ec7eebf265b2346;hp=a8647c1b27ff76d24aa179dcc3e05df55ae1dd82;hpb=1be0464201292e60fc8dcd3b922d9c4c4ac52d41;p=idzebra-moved-to-github.git

diff --git a/dict/dicttest.c b/dict/dicttest.c
index a8647c1..b74b795 100644
--- a/dict/dicttest.c
+++ b/dict/dicttest.c
@@ -4,7 +4,27 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: dicttest.c,v $
- * Revision 1.4  1994-09-01 17:49:37  adam
+ * Revision 1.9  1994-09-22 14:43:56  adam
+ * First functional version of lookup with error correction. A 'range'
+ * specified the maximum number of insertions+deletions+substitutions.
+ *
+ * Revision 1.8  1994/09/22  10:43:44  adam
+ * Two versions of depend. Type 1 is the tail-type compatible with
+ * all make programs. Type 2 is the GNU make with include facility.
+ * Type 2 is default. depend rule chooses current rule.
+ *
+ * Revision 1.7  1994/09/19  16:34:26  adam
+ * Depend rule change. Minor changes in dicttest.c
+ *
+ * Revision 1.6  1994/09/16  15:39:12  adam
+ * Initial code of lookup - not tested yet.
+ *
+ * Revision 1.5  1994/09/06  13:05:14  adam
+ * Further development of insertion. Some special cases are
+ * not properly handled yet! assert(0) are put here. The
+ * binary search in each page definitely reduce usr CPU.
+ *
+ * Revision 1.4  1994/09/01  17:49:37  adam
  * Removed stupid line. Work on insertion in dictionary. Not finished yet.
  *
  * Revision 1.3  1994/09/01  17:44:06  adam
@@ -21,32 +41,45 @@
 #include <stdlib.h>
 #include <string.h>
 #include <stdio.h>
+#include <ctype.h>
 
 #include <dict.h>
 
 char *prog;
 Dict dict;
 
+static int lookup_handle (Dict_char *name)
+{
+    printf ("%s\n", name);
+    return 0;
+}
+
 int main (int argc, char **argv)
 {
     const char *name = NULL;
     const char *inputfile = NULL;
     const char *base = NULL;
+    int range = -1;
     int rw = 0;
-    int infosize = 2;
+    int infosize = 4;
     int cache = 10;
     int ret;
+    int no_of_iterations = 0;
+    int no_of_new = 0, no_of_same = 0, no_of_change = 0;
+    int no_of_hits = 0, no_of_misses = 0;
+    int unique = 0;
     char *arg;
     
     prog = argv[0];
     if (argc < 2)
     {
-        fprintf (stderr, "usage:\n"
-                 "  %s [-s n] [-v n] [-i f] [-w] [-c n] base file\n",
+        fprintf (stderr, "usage:\n "
+                 " %s [-r n] [-u] [-s n] [-v n] [-i f] [-w] [-c n]"
+                 " base file\n",
                  prog);
         exit (1);
     }
-    while ((ret = options ("s:v:i:wc:", argv, argc, &arg)) != -2)
+    while ((ret = options ("r:us:v:i:wc:", argv, argc, &arg)) != -2)
     {
         if (ret == 0)
         {
@@ -60,6 +93,14 @@ int main (int argc, char **argv)
                 exit (1);
             }
         }
+        else if (ret == 'r')
+        {
+            range = atoi (arg);
+        }
+        else if (ret == 'u')
+        {
+            unique = 1;
+        }
         else if (ret == 'c')
         {
             cache = atoi(arg);
@@ -69,10 +110,7 @@ int main (int argc, char **argv)
         else if (ret == 'w')
             rw = 1;
         else if (ret == 'i')
-        {
             inputfile = arg;
-            rw = 1;
-        }
         else if (ret == 's')
         {
             infosize = atoi(arg);
@@ -107,8 +145,7 @@ int main (int argc, char **argv)
     if (inputfile)
     {
         FILE *ipf;
-        char ipf_buf[256];
-        char word[256];
+        char ipf_buf[1024];
         int line = 1;
         char infobytes[120];
         memset (infobytes, 0, 120);
@@ -119,17 +156,75 @@ int main (int argc, char **argv)
             exit (1);
         }
         
-        while (fgets (ipf_buf, 255, ipf))
+        while (fgets (ipf_buf, 1023, ipf))
         {
-            if (sscanf (ipf_buf, "%s", word) == 1)
+            char *ipf_ptr = ipf_buf;
+            sprintf (infobytes, "%d", line);
+            for (;*ipf_ptr && *ipf_ptr != '\n';ipf_ptr++)
             {
-                sprintf (infobytes, "%d", line);
-                dict_insert (dict, word, infosize, infobytes);
+                if (isalpha(*ipf_ptr) || *ipf_ptr == '_')
+                {
+                    int i = 1;
+                    while (ipf_ptr[i] && (isalnum(ipf_ptr[i]) ||
+                                          ipf_ptr[i] == '_'))
+                        i++;
+                    if (ipf_ptr[i])
+                        ipf_ptr[i++] = '\0';
+                    if (rw)
+                    {
+                        switch(dict_insert (dict, ipf_ptr,
+                                            infosize, infobytes))
+                        {
+                        case 0:
+                            no_of_new++;
+                            break;
+                        case 1:
+                            no_of_change++;
+                        if (unique)
+                            log (LOG_LOG, "%s change\n", ipf_ptr);
+                            break;
+                        case 2:
+                            if (unique)
+                                log (LOG_LOG, "%s duplicate\n", ipf_ptr);
+                            no_of_same++;
+                            break;
+                        }
+                    }
+                    else if(range < 0)
+                    {
+                        char *cp;
+
+                        cp = dict_lookup (dict, ipf_ptr);
+                        if (cp && *cp)
+                            no_of_hits++;
+                        else
+                            no_of_misses++;
+                    }
+                    else
+                    {
+                        dict_lookup_ec (dict, ipf_ptr, range, lookup_handle);
+                    }
+                    ++no_of_iterations;
+                    ipf_ptr += (i-1);
+                }
             }
             ++line;
         }
         fclose (ipf);
     }
+    if (rw)
+    {
+        log (LOG_LOG, "Insertions.... %d", no_of_iterations);
+        log (LOG_LOG, "No of new..... %d", no_of_new);
+        log (LOG_LOG, "No of change.. %d", no_of_change);
+        log (LOG_LOG, "No of same.... %d", no_of_same);
+    }
+    else
+    {
+        log (LOG_LOG, "Lookups....... %d", no_of_iterations);
+        log (LOG_LOG, "No of hits.... %d", no_of_hits);
+        log (LOG_LOG, "No of misses.. %d", no_of_misses);
+    }
     dict_close (dict);
     res_close (common_resource);
     return 0;