X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=util%2Fzebramap.c;h=e05bdcefdde6c78f46ae9ad826209eedc3664001;hb=56795fa9571dcb939b249bcd654ae5094fee835f;hp=4e24daa67ca596f132dac88dde5ffe8c6c009c15;hpb=01ddc55fd5a59535e229c09440cfdadccadf3555;p=idzebra-moved-to-github.git

diff --git a/util/zebramap.c b/util/zebramap.c
index 4e24daa..e05bdce 100644
--- a/util/zebramap.c
+++ b/util/zebramap.c
@@ -4,7 +4,19 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: zebramap.c,v $
- * Revision 1.16  1999-09-07 07:19:21  adam
+ * Revision 1.20  2000-03-02 14:35:19  adam
+ * Added structure year and date.
+ *
+ * Revision 1.19  1999/11/30 13:48:04  adam
+ * Improved installation. Updated for inclusion of YAZ header files.
+ *
+ * Revision 1.18  1999/10/15 08:27:46  adam
+ * Fixed replace handler. 8-bit fix.
+ *
+ * Revision 1.17  1999/09/08 12:13:21  adam
+ * Fixed minor bug "replace"-mappings. Removed some logging messages.
+ *
+ * Revision 1.16  1999/09/07 07:19:21  adam
  * Work on character mapping. Implemented replace rules.
  *
  * Revision 1.15  1999/05/26 07:49:14  adam
@@ -63,15 +75,17 @@
 #include <assert.h>
 #include <ctype.h>
 
-#include <yaz-util.h>
+#include <yaz/yaz-util.h>
 #include <charmap.h>
 #include <zebramap.h>
 
 #define ZEBRA_MAP_TYPE_SORT  1
 #define ZEBRA_MAP_TYPE_INDEX 2
 
+#define ZEBRA_REPLACE_ANY  300
+
 struct zm_token {
-    char *token_from;
+    int *token_from;
     char *token_to;
     int token_min;
     struct zm_token *next;
@@ -166,6 +180,7 @@ static void zebra_map_read (ZebraMaps zms, const char *name)
 	    (*zm)->maptab = NULL;
 	    (*zm)->completeness = 0;
 	    (*zm)->positioned = 0;
+	    (*zm)->replace_tokens = 0;
 	}
 	else if (zm && !yaz_matchstr (argv[0], "charmap") && argc == 2)
 	{
@@ -187,27 +202,43 @@ static void zebra_map_read (ZebraMaps zms, const char *name)
         else if (zm && !yaz_matchstr (argv[0], "replace") && argc >= 2)
         {
 	    struct zm_token *token = nmem_malloc (zms->nmem, sizeof(*token));
-	    char *cp, *dp;
 	    token->next = (*zm)->replace_tokens;
 	    (*zm)->replace_tokens = token;
-	    dp = token->token_from = nmem_strdup (zms->nmem, cp = argv[1]);
-	    while (*cp)
-	    {
-		if (*cp == '$')
-		{
-		    *dp++ = ' ';
-		    cp++;
-		}
-		else
-		    *dp++ = zebra_prim(&cp);
+#if 0
+	    logf (LOG_LOG, "replace %s", argv[1]);
+#endif
+	    token->token_from = 0;
+            if (argc >= 2)
+            {
+	        char *cp = argv[1];
+	        int *dp = token->token_from = (int *)
+                    nmem_malloc (zms->nmem, (1+strlen(cp))*sizeof(int));
+	        while (*cp)
+		    if (*cp == '$')
+		    {
+		        *dp++ = ' ';
+		        cp++;
+		    }
+		    else if (*cp == '.')
+                    {
+                        *dp++ = ZEBRA_REPLACE_ANY;
+                        cp++;
+                    }
+                    else
+		    {
+		        *dp++ = zebra_prim(&cp);
+#if 0
+			logf (LOG_LOG, "  char %2X %c", dp[-1], dp[-1]);
+#endif
+		    }
+	        *dp = '\0';
 	    }
-	    *dp = '\0';
-
 	    if (argc >= 3)
 	    {
-		dp = token->token_to = nmem_strdup (zms->nmem, cp = argv[2]);
+                char *cp = argv[2];
+		char *dp = token->token_to =
+                    nmem_malloc (zms->nmem, strlen(cp)+1);
 		while (*cp)
-		{
 		    if (*cp == '$')
 		    {
 			*dp++ = ' ';
@@ -215,7 +246,6 @@ static void zebra_map_read (ZebraMaps zms, const char *name)
 		    }
 		    else
 			*dp++ = zebra_prim(&cp);
-		}
 		*dp = '\0';
 	    }
 	    else
@@ -279,7 +309,7 @@ chrmaptab zebra_charmap_get (ZebraMaps zms, unsigned reg_id)
 	logf (LOG_WARN, "Unknown register type: %c", reg_id);
 
 	zm->reg_id = reg_id;
-	zm->maptab_name = NULL;
+	zm->maptab_name = nmem_strdup (zms->nmem, "@");
 	zm->maptab = NULL;
 	zm->type = ZEBRA_MAP_TYPE_INDEX;
 	zm->completeness = 0;
@@ -545,6 +575,14 @@ int zebra_maps_attr (ZebraMaps zms, Z_AttributesPlusTerm *zapt,
         *reg_id = '0';
         *search_type = "phrase";
         break;
+    case 4:  /* year */
+        *reg_id = 'y';
+        *search_type = "phrase";
+        break;
+    case 5:  /* date */
+        *reg_id = 'd';
+        *search_type = "phrase";
+        break;
     default:
 	return -1;
     }
@@ -561,11 +599,10 @@ WRBUF zebra_replace(ZebraMaps zms, unsigned reg_id, const char *ex_list,
 
     wrbuf_rewind(zms->wrbuf_1);
     wrbuf_write(zms->wrbuf_1, input_str, input_len);
-    if (!zm->replace_tokens)
+    if (!zm || !zm->replace_tokens)
 	return zms->wrbuf_1;
-   
-#if 0 
-    logf (LOG_LOG, "zebra_replace");
+  
+#if 0
     logf (LOG_LOG, "in:%.*s:", wrbuf_len(zms->wrbuf_1),
 	  wrbuf_buf(zms->wrbuf_1));
 #endif
@@ -615,8 +652,8 @@ int zebra_replace_sub(ZebraMaps zms, unsigned reg_id, const char *ex_list,
 		if (i+j < 0 || j+i >= input_len)
 		    c = ' ';
 		else
-		    c = tolower(input_str[j+i]);
-		if (token->token_from[j] == '.')
+		    c = input_str[j+i] & 255;
+		if (token->token_from[j] == ZEBRA_REPLACE_ANY)
 		{
 		    if (c == ' ')
 			break;
@@ -625,7 +662,9 @@ int zebra_replace_sub(ZebraMaps zms, unsigned reg_id, const char *ex_list,
 		else
 		{
 		    if (c != token->token_from[j])
+		    {
 			break;
+		    }
 		    if (!replace_done)
 		    {
 			const char *cp = token->token_to;