Support for relations on text operands. Open range support in

DFA module (i.e. [-j], [g-]).

diff --git a/CHANGELOG b/CHANGELOG
index 2f2bd84..c783420 100644 (file)
--- a/CHANGELOG
+++ b/CHANGELOG
@@ -1,3 +1,11 @@
+Added support for relational operators on text when using
+RPN queries.
+
+Added support for sort specifications in RPN queries. Type 7
+specifies 'sort' where value 1=ascending, value 2=descending.
+The use attribute specifies the field criteria as usual.
+The term specifies priority (0=first, 1=second).
+
 Changed the way use attributes are specified in the recordId
 specification.
 

diff --git a/dfa/dfa.c b/dfa/dfa.c
index 072c397..e1e1c45 100644 (file)
--- a/dfa/dfa.c
+++ b/dfa/dfa.c
@@ -4,7 +4,11 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: dfa.c,v $
- * Revision 1.21  1998-06-22 11:33:39  adam
+ * Revision 1.22  1998-06-24 12:16:10  adam
+ * Support for relations on text operands. Open range support in
+ * DFA module (i.e. [-j], [g-]).
+ *
+ * Revision 1.21  1998/06/22 11:33:39  adam
  * Added two type casts.
  *
  * Revision 1.20  1998/06/08 14:40:44  adam
@@ -423,18 +427,28 @@ static int read_charset (struct DFA_parse *parse_info)
     {
         if (!esc0 && ch0 == ']')
             break;
-        if (parse_info->cmap)
-        {
-            const char **mapto;
-           char mapfrom[2];
-            const char *mcp = mapfrom;
-            mapfrom[0] = ch0;
-            mapto = (*parse_info->cmap)(parse_info->cmap_data, &mcp, 1);
-            assert (mapto);
-            ch0 = mapto[0][0];
-        }
-        add_BSet (parse_info->charset, parse_info->look_chars, ch0);
-        ch1 = nextchar_set (parse_info, &esc1);
+       if (!esc0 && ch0 == '-')
+       {
+           ch1 = ch0;
+           esc1 = esc0;
+           ch0 = 1;
+           add_BSet (parse_info->charset, parse_info->look_chars, ch0);
+       }
+       else
+       {
+           if (parse_info->cmap)
+           {
+               const char **mapto;
+               char mapfrom[2];
+               const char *mcp = mapfrom;
+               mapfrom[0] = ch0;
+               mapto = (*parse_info->cmap)(parse_info->cmap_data, &mcp, 1);
+               assert (mapto);
+               ch0 = mapto[0][0];
+           }
+           add_BSet (parse_info->charset, parse_info->look_chars, ch0);
+           ch1 = nextchar_set (parse_info, &esc1);
+       }
         if (!esc1 && ch1 == '-')
         {
             int open_range = 0;

diff --git a/dict/lookgrep.c b/dict/lookgrep.c
index 3e094e1..39d2b3f 100644 (file)
--- a/dict/lookgrep.c
+++ b/dict/lookgrep.c
@@ -1,10 +1,14 @@
 /*
- * Copyright (C) 1994-1996, Index Data I/S 
+ * Copyright (C) 1994-1998, Index Data I/S 
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: lookgrep.c,v $
- * Revision 1.20  1997-10-27 14:33:03  adam
+ * Revision 1.21  1998-06-24 12:16:12  adam
+ * Support for relations on text operands. Open range support in
+ * DFA module (i.e. [-j], [g-]).
+ *
+ * Revision 1.20  1997/10/27 14:33:03  adam
  * Moved towards generic character mapping depending on "structure"
  * field in abstract syntax file. Fixed a few memory leaks. Fixed
  * bug with negative integers when doing searches with relational
@@ -419,6 +423,13 @@ int dict_lookup_grep (Dict dict, const char *pattern, int range, void *client,
     struct DFA *dfa = dfa_init();
     int i, d;
 
+#if 0
+    debug_dfa_trav = 1;
+    debug_dfa_tran = 1;
+    debug_dfa_followpos = 1;
+    dfa_verbose = 1;
+#endif
+
     logf (LOG_DEBUG, "dict_lookup_grep range=%d", range);
     for (i = 0; pattern[i]; i++)
     {

diff --git a/index/zebraapi.c b/index/zebraapi.c
index 94a31f9..4e5d2fd 100644 (file)
--- a/index/zebraapi.c
+++ b/index/zebraapi.c
@@ -4,7 +4,11 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: zebraapi.c,v $
- * Revision 1.6  1998-06-22 11:36:47  adam
+ * Revision 1.7  1998-06-24 12:16:13  adam
+ * Support for relations on text operands. Open range support in
+ * DFA module (i.e. [-j], [g-]).
+ *
+ * Revision 1.6  1998/06/22 11:36:47  adam
  * Added authentication check facility to zebra.
  *
  * Revision 1.5  1998/06/13 00:14:08  adam
@@ -279,8 +283,8 @@ void zebra_scan (ZebraHandle zh, ODR stream, Z_AttributesPlusTerm *zapt,
 }
 
 void zebra_sort (ZebraHandle zh, ODR stream,
-                int num_input_setnames, char **input_setnames,
-                char *output_setname, Z_SortKeySpecList *sort_sequence,
+                int num_input_setnames, const char **input_setnames,
+                const char *output_setname, Z_SortKeySpecList *sort_sequence,
                 int *sort_status)
 {
     zh->errCode = 0;

diff --git a/index/zrpn.c b/index/zrpn.c
index a5617dc..5a1b42a 100644 (file)
--- a/index/zrpn.c
+++ b/index/zrpn.c
@@ -4,7 +4,11 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: zrpn.c,v $
- * Revision 1.80  1998-06-23 15:33:34  adam
+ * Revision 1.81  1998-06-24 12:16:14  adam
+ * Support for relations on text operands. Open range support in
+ * DFA module (i.e. [-j], [g-]).
+ *
+ * Revision 1.80  1998/06/23 15:33:34  adam
  * Added feature to specify sort criteria in query (type 7 specifies
  * sort flags).
  *
@@ -720,20 +724,33 @@ static void gen_regular_rel (char *dst, int val, int islt)
     strcat (dst, "))");
 }
 
+void string_rel_add_char (char **term_p, const char *src, int *indx)
+{
+    if (src[*indx] == '\\')
+       *(*term_p)++ = src[(*indx)++];
+    *(*term_p)++ = src[(*indx)++];
+}
+
+/*
+ *   >  abc     ([b-].*|a[c-].*|ab[d-].*|abc.+)
+ *              ([^-a].*|a[^-b].*ab[^-c].*|abc.+)
+ *   >= abc     ([b-].*|a[c-].*|ab[c-].*)
+ *              ([^-a].*|a[^-b].*|ab[c-].*)
+ *   <  abc     ([-0].*|a[-a].*|ab[-b].*)
+ *              ([^a-].*|a[^b-].*|ab[^c-].*)
+ *   <= abc     ([-0].*|a[-a].*|ab[-b].*|abc)
+ *              ([^a-].*|a[^b-].*|ab[^c-].*|abc)
+ */
 static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
-                                const char **term_sub,
-                                char *term_dict,
-                                oid_value attributeSet,
-                                struct grep_info *grep_info,
-                                int *max_pos,
-                                int reg_type,
-                                char *term_dst)
+                           const char **term_sub, char *term_dict,
+                           oid_value attributeSet,
+                           int reg_type, int space_split, char *term_dst)
 {
     AttrType relation;
     int relation_value;
-    int term_value;
-    int r;
+    int i;
     char *term_tmp = term_dict + strlen(term_dict);
+    char term_component[256];
 
     attr_init (&relation, zapt, 2);
     relation_value = attr_find (&relation, NULL);
@@ -742,52 +759,141 @@ static int string_relation (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
     switch (relation_value)
     {
     case 1:
-        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
-                      term_dst))
+        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component,
+                      space_split, term_dst))
             return 0;
-        term_value = atoi (term_tmp);
         logf (LOG_DEBUG, "Relation <");
-        gen_regular_rel (term_tmp, term_value-1, 1);
+       
+       *term_tmp++ = '(';
+       for (i = 0; term_component[i]; )
+       {
+           int j = 0;
+
+           if (i)
+               *term_tmp++ = '|';
+           while (j < i)
+               string_rel_add_char (&term_tmp, term_component, &j);
+
+           *term_tmp++ = '[';
+
+           *term_tmp++ = '^';
+           string_rel_add_char (&term_tmp, term_component, &i);
+           *term_tmp++ = '-';
+
+           *term_tmp++ = ']';
+           *term_tmp++ = '.';
+           *term_tmp++ = '*';
+       }
+       *term_tmp++ = ')';
+       *term_tmp = '\0';
         break;
     case 2:
-        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
-                      term_dst))
+        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component,
+                      space_split, term_dst))
             return 0;
-        term_value = atoi (term_tmp);
         logf (LOG_DEBUG, "Relation <=");
-        gen_regular_rel (term_tmp, term_value, 1);
-        break;
-    case 4:
-        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
-                      term_dst))
-            return 0;
-        term_value = atoi (term_tmp);
-        logf (LOG_DEBUG, "Relation >=");
-        gen_regular_rel (term_tmp, term_value, 0);
+
+       *term_tmp++ = '(';
+       for (i = 0; term_component[i]; )
+       {
+           int j = 0;
+
+           while (j < i)
+               string_rel_add_char (&term_tmp, term_component, &j);
+           *term_tmp++ = '[';
+
+           *term_tmp++ = '^';
+           string_rel_add_char (&term_tmp, term_component, &i);
+           *term_tmp++ = '-';
+
+           *term_tmp++ = ']';
+           *term_tmp++ = '.';
+           *term_tmp++ = '*';
+
+           *term_tmp++ = '|';
+       }
+       for (i = 0; term_component[i]; )
+           string_rel_add_char (&term_tmp, term_component, &i);
+       *term_tmp++ = ')';
+       *term_tmp = '\0';
         break;
     case 5:
-        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp, 1,
-                      term_dst))
+        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component,
+                      space_split, term_dst))
             return 0;
-        term_value = atoi (term_tmp);
         logf (LOG_DEBUG, "Relation >");
-        gen_regular_rel (term_tmp, term_value+1, 0);
+
+       *term_tmp++ = '(';
+       for (i = 0; term_component[i];)
+       {
+           int j = 0;
+
+           while (j < i)
+               string_rel_add_char (&term_tmp, term_component, &j);
+           *term_tmp++ = '[';
+           
+           *term_tmp++ = '^';
+           *term_tmp++ = '-';
+           string_rel_add_char (&term_tmp, term_component, &i);
+
+           *term_tmp++ = ']';
+           *term_tmp++ = '.';
+           *term_tmp++ = '*';
+
+           *term_tmp++ = '|';
+       }
+       for (i = 0; term_component[i];)
+           string_rel_add_char (&term_tmp, term_component, &i);
+       *term_tmp++ = '.';
+       *term_tmp++ = '+';
+       *term_tmp++ = ')';
+       *term_tmp = '\0';
+        break;
+    case 4:
+        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component,
+                      space_split, term_dst))
+            return 0;
+        logf (LOG_DEBUG, "Relation >=");
+
+       *term_tmp++ = '(';
+       for (i = 0; term_component[i];)
+       {
+           int j = 0;
+
+           if (i)
+               *term_tmp++ = '|';
+           while (j < i)
+               string_rel_add_char (&term_tmp, term_component, &j);
+           *term_tmp++ = '[';
+
+           if (term_component[i+1])
+           {
+               *term_tmp++ = '^';
+               *term_tmp++ = '-';
+               string_rel_add_char (&term_tmp, term_component, &i);
+           }
+           else
+           {
+               string_rel_add_char (&term_tmp, term_component, &i);
+               *term_tmp++ = '-';
+           }
+           *term_tmp++ = ']';
+           *term_tmp++ = '.';
+           *term_tmp++ = '*';
+       }
+       *term_tmp++ = ')';
+       *term_tmp = '\0';
         break;
     case 3:
     default:
         logf (LOG_DEBUG, "Relation =");
-       *term_tmp = '(';
-        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_tmp+1, 1,
-                      term_dst))
+        if (!term_100 (zh->zebra_maps, reg_type, term_sub, term_component,
+                      space_split, term_dst))
             return 0;
+       strcat (term_tmp, "(");
+       strcat (term_tmp, term_component);
        strcat (term_tmp, ")");
     }
-    logf (LOG_DEBUG, "dict_lookup_grep: %s", term_tmp);
-    r = dict_lookup_grep (zh->dict, term_dict, 0, grep_info, max_pos,
-                          0, grep_handle);
-    if (r)
-        logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r);
-    logf (LOG_DEBUG, "%d positions", grep_info->isam_p_indx);
     return 1;
 }
 
@@ -798,7 +904,7 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
                        int num_bases, char **basenames,
                        char *term_dst)
 {
-    char term_dict[2*IT_MAX_WORD+2];
+    char term_dict[2*IT_MAX_WORD+4000];
     int j, r, base_no;
     AttrType truncation;
     int truncation_value;
@@ -882,20 +988,14 @@ static int string_term (ZebraHandle zh, Z_AttributesPlusTerm *zapt,
        case -1:         /* not specified */
        case 100:        /* do not truncate */
            if (!string_relation (zh, zapt, &termp, term_dict,
-                                 attributeSet, grep_info, &max_pos,
-                                 reg_type, term_dst))
+                                 attributeSet,
+                                 reg_type, space_split, term_dst))
                return 0;
-#if 0
-           term_dict[j++] = '(';   
-           if (!term_100 (zh->zebra_maps, reg_type,
-                          &termp, term_dict + j, space_split, term_dst))
-               return 0;
-           strcat (term_dict, ")");
-           r = dict_lookup_grep (zh->dict, term_dict, 0, grep_info,
-                                 &max_pos, 0, grep_handle);
+           logf (LOG_DEBUG, "dict_lookup_grep: %s", term_dict+prefix_len);
+           r = dict_lookup_grep (zh->dict, term_dict, 0, grep_info, &max_pos,
+                                 0, grep_handle);
            if (r)
-               logf (LOG_WARN, "dict_lookup_grep err, trunc=none:%d", r);
-#endif
+               logf (LOG_WARN, "dict_lookup_grep fail, rel=gt: %d", r);
            break;
        case 1:          /* right truncation */
            term_dict[j++] = '(';

diff --git a/index/zserver.c b/index/zserver.c
index 4718912..34e7406 100644 (file)
--- a/index/zserver.c
+++ b/index/zserver.c
@@ -4,7 +4,11 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: zserver.c,v $
- * Revision 1.60  1998-06-22 11:36:49  adam
+ * Revision 1.61  1998-06-24 12:16:15  adam
+ * Support for relations on text operands. Open range support in
+ * DFA module (i.e. [-j], [g-]).
+ *
+ * Revision 1.60  1998/06/22 11:36:49  adam
  * Added authentication check facility to zebra.
  *
  * Revision 1.59  1998/06/12 12:22:13  adam
@@ -386,7 +390,8 @@ int bend_sort (void *handle, bend_sort_rr *rr)
 {
     ZebraHandle zh = handle;
 
-    zebra_sort (zh, rr->stream, rr->num_input_setnames, rr->input_setnames,
+    zebra_sort (zh, rr->stream,
+                rr->num_input_setnames, (const char **) rr->input_setnames,
                rr->output_setname, rr->sort_sequence, &rr->sort_status);
     rr->errcode = zh->errCode;
     rr->errstring = zh->errString;

diff --git a/index/zserver.h b/index/zserver.h
index 7337b77..51b59bb 100644 (file)
--- a/index/zserver.h
+++ b/index/zserver.h
@@ -4,7 +4,11 @@
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: zserver.h,v $
- * Revision 1.35  1998-06-23 15:33:35  adam
+ * Revision 1.36  1998-06-24 12:16:16  adam
+ * Support for relations on text operands. Open range support in
+ * DFA module (i.e. [-j], [g-]).
+ *
+ * Revision 1.35  1998/06/23 15:33:35  adam
  * Added feature to specify sort criteria in query (type 7 specifies
  * sort flags).
  *
@@ -230,8 +234,8 @@ void resultSetSort (ZebraHandle zh, ODR stream,
                    Z_SortKeySpecList *sort_sequence, int *sort_status);
 
 void zebra_sort (ZebraHandle zh, ODR stream,
-                int num_input_setnames, char **input_setnames,
-                char *output_setname, Z_SortKeySpecList *sort_sequence,
+                int num_input_setnames, const char **input_setnames,
+                const char *output_setname, Z_SortKeySpecList *sort_sequence,
                 int *sort_status);
 
 void zlog_rpn (Z_RPNQuery *rpn);