X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=recctrl%2Frecgrs.c;h=1bae99de456153dd29fedd4b83d0ac2c933d3518;hb=b72718347ce9e412d57b1203441fe7e51746772c;hp=a0cdef7c9a54e146f2b1f9f2cfff24a7301dd4ae;hpb=9823f7a1e997efcb035afcaa26718b8b7f87fc0a;p=idzebra-moved-to-github.git

diff --git a/recctrl/recgrs.c b/recctrl/recgrs.c
index a0cdef7..1bae99d 100644
--- a/recctrl/recgrs.c
+++ b/recctrl/recgrs.c
@@ -1,4 +1,4 @@
-/* $Id: recgrs.c,v 1.75 2003-03-08 14:27:58 pop Exp $
+/* $Id: recgrs.c,v 1.82 2003-08-21 10:29:00 adam Exp $
    Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003
    Index Data Aps
 
@@ -62,7 +62,8 @@ static int read_grs_type (struct grs_handlers *h,
         strcpy (p->type, cp+1);
     for (gh = h->handlers; gh; gh = gh->next)
     {
-        if (!memcmp (type, gh->type->type, cp-type))
+        if (!memcmp (type, gh->type->type, cp-type) && 
+	    gh->type->type[cp-type] == '\0')
 	{
 	    if (!gh->initFlag)
 	    {
@@ -99,6 +100,7 @@ static void *grs_init(RecType recType)
     grs_add_handler (h, recTypeGrs_tcl);
 #endif
     grs_add_handler (h, recTypeGrs_marc);
+    grs_add_handler (h, recTypeGrs_marcxml);
 #if HAVE_EXPAT_H
     grs_add_handler (h, recTypeGrs_xml);
 #endif
@@ -123,87 +125,88 @@ static void grs_destroy(void *clientData)
     xfree (h);
 }
 
-int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p) {
-  int res = 1;
-  char *attname;
-  data1_xattr *attr;
-
-  if (!p) {
-    return (1);
-  } else {
-    if (p->which == XPATH_PREDICATE_RELATION) {
-      if (p->u.relation.name[0]) {
-	if (*p->u.relation.name != '@') {
-	  logf(LOG_WARN, 
-	       "  Only attributes (@) are supported in xelm xpath predicates");
-	  logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
-	  return (1);
-	}
-	attname = p->u.relation.name + 1;
-	res = 0;
-	/* looking for the attribute with a specified name */
-	for (attr = n->u.tag.attributes; attr; attr = attr->next) {
-	  logf(LOG_DEBUG,"  - attribute %s <-> %s", attname, attr->name );
-
-	  if (!strcmp(attr->name, attname)) {
-	    if (p->u.relation.op[0]) {
-	      if (*p->u.relation.op != '=') {
-		logf(LOG_WARN, 
-		     "Only '=' relation is supported (%s)",p->u.relation.op);
-		logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
-		res = 1; break;
-	      } else {
-		logf(LOG_DEBUG,"    - value %s <-> %s", 
-		     p->u.relation.value, attr->value );
-		if (!strcmp(attr->value, p->u.relation.value)) {
-		  res = 1; break;
-		} 
-	      }
-	    } else {
-	      /* attribute exists, no value specified */
-	      res = 1; break;
-	    }
-	  }
-	}
-	return (res);
-      } else {
-	return (1);
-      }
-    } 
-    else if (p->which == XPATH_PREDICATE_BOOLEAN) {
-      if (!strcmp(p->u.boolean.op,"and")) {
-	return (d1_check_xpath_predicate(n, p->u.boolean.left) 
-		&& d1_check_xpath_predicate(n, p->u.boolean.right)); 
-      }
-      else if (!strcmp(p->u.boolean.op,"or")) {
-	return (d1_check_xpath_predicate(n, p->u.boolean.left) 
-		|| d1_check_xpath_predicate(n, p->u.boolean.right)); 
-      } else {
-	logf(LOG_WARN, "Unknown boolean relation %s, ignored",p->u.boolean.op);
-	return (1);
-      }
+int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p)
+{
+    int res = 1;
+    char *attname;
+    data1_xattr *attr;
+    
+    if (!p) {
+        return 1;
+    } else {
+        if (p->which == XPATH_PREDICATE_RELATION) {
+            if (p->u.relation.name[0]) {
+                if (*p->u.relation.name != '@') {
+                    logf(LOG_WARN, 
+                         "  Only attributes (@) are supported in xelm xpath predicates");
+                    logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
+                    return (1);
+                }
+                attname = p->u.relation.name + 1;
+                res = 0;
+                /* looking for the attribute with a specified name */
+                for (attr = n->u.tag.attributes; attr; attr = attr->next) {
+                    logf(LOG_DEBUG,"  - attribute %s <-> %s", attname, attr->name );
+                    
+                    if (!strcmp(attr->name, attname)) {
+                        if (p->u.relation.op[0]) {
+                            if (*p->u.relation.op != '=') {
+                                logf(LOG_WARN, 
+                                     "Only '=' relation is supported (%s)",p->u.relation.op);
+                                logf(LOG_WARN, "predicate %s ignored", p->u.relation.name);
+                                res = 1; break;
+                            } else {
+                                logf(LOG_DEBUG,"    - value %s <-> %s", 
+                                     p->u.relation.value, attr->value );
+                                if (!strcmp(attr->value, p->u.relation.value)) {
+                                    res = 1; break;
+                                } 
+                            }
+                        } else {
+                            /* attribute exists, no value specified */
+                            res = 1; break;
+                        }
+                    }
+                }
+                return res;
+            } else {
+                return 1;
+            }
+        } 
+        else if (p->which == XPATH_PREDICATE_BOOLEAN) {
+            if (!strcmp(p->u.boolean.op,"and")) {
+                return d1_check_xpath_predicate(n, p->u.boolean.left) 
+                    && d1_check_xpath_predicate(n, p->u.boolean.right); 
+            }
+            else if (!strcmp(p->u.boolean.op,"or")) {
+                return (d1_check_xpath_predicate(n, p->u.boolean.left) 
+                        || d1_check_xpath_predicate(n, p->u.boolean.right)); 
+            } else {
+                logf(LOG_WARN, "Unknown boolean relation %s, ignored",p->u.boolean.op);
+                return 1;
+            }
+        }
     }
-  }
-  return 0;
+    return 0;
 }
 
 
 /* *ostrich*
    
-   New function, looking for xpath "element" definitions in abs, by
-   tagpath, using a kind of ugly regxp search.The DFA was built while
-   parsing abs, so here we just go trough them and try to match
-   against the given tagpath. The first matching entry is returned.
+New function, looking for xpath "element" definitions in abs, by
+tagpath, using a kind of ugly regxp search.The DFA was built while
+parsing abs, so here we just go trough them and try to match
+against the given tagpath. The first matching entry is returned.
 
-   pop, 2002-12-13
+pop, 2002-12-13
 
-   Added support for enhanced xelm. Now [] predicates are considered
-   as well, when selecting indexing rules... (why the hell it's called
-   termlist???)
+Added support for enhanced xelm. Now [] predicates are considered
+as well, when selecting indexing rules... (why the hell it's called
+termlist???)
 
-   pop, 2003-01-17
+pop, 2003-01-17
 
- */
+*/
 
 data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
 {
@@ -218,6 +221,7 @@ data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
     int ok = 0;
     
     sprintf (pexpr, "%s\n", tagpath);
+    logf(LOG_DEBUG,"Checking tagpath %s",tagpath);
     while (xpe) 
     {
         struct DFA_state **dfaar = xpe->dfa->states;
@@ -227,9 +231,11 @@ data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
         int i;
         unsigned char c;
         int start_line = 1;
-        
+
+	logf(LOG_DEBUG," - xpath %s",xpe->xpath_expr);
+
         c = *pexpr++; t = s->trans; i = s->tran_no;
-        if (c >= t->ch[0] && c <= t->ch[1]) {
+	if ((c >= t->ch[0] && c <= t->ch[1]) || (!t->ch[0])) {
             p = pexpr;
             do {
                 if ((s = dfaar[t->to])->rule_no && 
@@ -243,43 +249,43 @@ data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
                 }
                 p++;
             } while (i >= 0);
-        }
+	}
         pexpr--;
         if (ok) {
 #ifdef ENHANCED_XELM 
-	  /* we have to check the perdicates up to the root node */
-	  xp = xpe->xpath;
-
-	  /* find the first tag up in the node structure */
-	  nn = n; while (nn && nn->which != DATA1N_tag) {
-	    nn = nn->parent;
-	  }
-
-	  /* go from inside out in the node structure, while going
-	     backwards trough xpath location steps ... */
-	  for (i=xpe->xpath_len - 1; i>0; i--) {
-	    
-	    logf(LOG_DEBUG,"Checking step %d: %s on tag %s",
+            /* we have to check the perdicates up to the root node */
+            xp = xpe->xpath;
+            
+            /* find the first tag up in the node structure */
+            nn = n; while (nn && nn->which != DATA1N_tag) {
+                nn = nn->parent;
+            }
+            
+            /* go from inside out in the node structure, while going
+               backwards trough xpath location steps ... */
+            for (i=xpe->xpath_len - 1; i>0; i--) {
+                
+                logf(LOG_DEBUG,"Checking step %d: %s on tag %s",
 		     i,xp[i].part,nn->u.tag.tag);
-
-	    if (!d1_check_xpath_predicate(nn, xp[i].predicate)) {
-	      logf(LOG_DEBUG,"  Predicates didn't match");
-	      ok = 0;
-	      break;
-	    }
-
-	    if (nn->which == DATA1N_tag) {
-	      nn = nn->parent;
-	    }
-	  }
+                
+                if (!d1_check_xpath_predicate(nn, xp[i].predicate)) {
+                    logf(LOG_DEBUG,"  Predicates didn't match");
+                    ok = 0;
+                    break;
+                }
+                
+                if (nn->which == DATA1N_tag) {
+                    nn = nn->parent;
+                }
+            }
 #endif
-	  if (ok) {
-	    break;
-	  }
+            if (ok) {
+                break;
+            }
 	}
         xpe = xpe->next;
     } 
-
+    
     xfree(pexpr);
     
     if (ok) {
@@ -308,6 +314,34 @@ data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
   modified by pop, 2002-12-13
 */
 
+/* add xpath index for an attribute */
+static void index_xpath_attr (char *tag_path, char *name, char *value,
+			      char *structure, struct recExtractCtrl *p,
+			      RecWord *wrd)
+{
+    wrd->attrSet = VAL_IDXPATH;
+    wrd->attrUse = 1;
+    wrd->reg_type = '0';
+    wrd->string = tag_path;
+    wrd->length = strlen(tag_path);
+    (*p->tokenAdd)(wrd);
+    
+    if (value) {
+        wrd->attrUse = 1015;
+        wrd->reg_type = 'w';
+        wrd->string = value;
+        wrd->length = strlen(value);
+        (*p->tokenAdd)(wrd);
+    }
+    
+    wrd->attrUse = 2;
+    wrd->reg_type = '0';
+    wrd->string = tag_path;
+    wrd->length = strlen(tag_path);
+    (*p->tokenAdd)(wrd);
+}
+
+
 static void index_xpath (data1_node *n, struct recExtractCtrl *p,
                          int level, RecWord *wrd, int use)
 {
@@ -315,6 +349,12 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p,
     char tag_path_full[1024];
     size_t flen = 0;
     data1_node *nn;
+    int termlist_only = 1;
+
+    if ((!n->root->u.root.absyn) ||
+	(n->root->u.root.absyn->enable_xpath_indexing)) {
+      termlist_only = 0;
+    }
 
     switch (n->which)
     {
@@ -323,16 +363,18 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p,
         wrd->length = n->u.data.len;
         if (p->flagShowRecords)
         {
-            printf("%*s data=", (level + 1) * 4, "");
+            printf("%*s XData:\"", (level + 1) * 4, "");
             for (i = 0; i<wrd->length && i < 8; i++)
                 fputc (wrd->string[i], stdout);
-            printf("\n");
+            printf("\"\n");
         }  
         else  {
             data1_termlist *tl;
             int xpdone = 0;
             flen = 0;
             
+	    //	    logf (LOG_LOG,"Index value, %s",wrd->string);
+
             /* we have to fetch the whole path to the data tag */
             for (nn = n; nn; nn = nn->parent) {
                 if (nn->which == DATA1N_tag) {
@@ -366,14 +408,14 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p,
                 }
             }
             /* xpath indexing is done, if there was no termlist given, 
-               or no ! attribute... */
-            if (!xpdone) {
+               or no ! in the termlist, and default indexing is enabled... */
+            if ((!xpdone) && (!termlist_only)) {
                 wrd->attrSet = VAL_IDXPATH;
                 wrd->attrUse = use;
                 wrd->reg_type = 'w';
                 (*p->tokenAdd)(wrd);
             }
-        }
+	}
         break;
     case DATA1N_tag:
         flen = 0;
@@ -410,95 +452,139 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p,
         else
         {
             data1_xattr *xp;
-            (*p->tokenAdd)(wrd);   /* index element pag (AKA tag path) */
-            if (use == 1)
+            data1_termlist *tl;
+	    int do_xpindex;
+            
+	    tag_path_full[flen] = 0;
+            
+            /* Add tag start/end xpath index, only when there is a ! in the apropriate xelm
+               directive, or default xpath indexing is enabled */
+	    if (!(do_xpindex = 1 - termlist_only)) {
+                if ((tl = xpath_termlist_by_tagpath(tag_path_full, n))) {
+                    for (; tl; tl = tl->next) { if (!tl->att) {do_xpindex = 1;} }
+                }
+	    }
+	    if (do_xpindex) {
+                (*p->tokenAdd)(wrd);   /* index element pag (AKA tag path) */
+	    }
+            
+            if (use == 1) /* only for the starting tag... */
             {
-                for (xp = n->u.tag.attributes; xp; xp = xp->next)
-                {
+                
+#define MAX_ATTR_COUNT 50
+                
+                data1_termlist *tll[MAX_ATTR_COUNT];
+                
+                int i = 0;
+                
+                /* get termlists for attributes, and find out, if we have to do xpath indexing */
+                for (xp = n->u.tag.attributes; xp; xp = xp->next) {
+                    i++;
+                }
+                
+                i = 0;
+                for (xp = n->u.tag.attributes; xp; xp = xp->next) {
                     char comb[512];
-                    /* attribute  (no value) */
-                    wrd->reg_type = '0';
-                    wrd->attrUse = 3;
-                    wrd->string = xp->name;
-                    wrd->length = strlen(xp->name);
+                    int do_xpindex = 1 - termlist_only;
+                    data1_termlist *tl;
+                    char attr_tag_path_full[1024]; 
+                    int int_len = flen;
                     
-                    wrd->seqno--;
-                    (*p->tokenAdd)(wrd);
-
-                    if (xp->value &&
-                        strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2)
+                    /* this could be cached as well */
+                    sprintf (attr_tag_path_full, "@%s/%.*s",
+                             xp->name, int_len, tag_path_full);
+                    
+                    tll[i] = xpath_termlist_by_tagpath(attr_tag_path_full,n);
+                    
+                    /* if there is a ! in the xelm termlist, or default indexing is on, 
+                       proceed with xpath idx */
+                    if ((tl = tll[i]))
                     {
-                        /* attribute value exact */
-                        strcpy (comb, xp->name);
-                        strcat (comb, "=");
-                        strcat (comb, xp->value);
+                        for (; tl; tl = tl->next)
+                        { 
+                            if (!tl->att)
+                                do_xpindex = 1;
+                        }
+                    }
+                    
+                    if (do_xpindex) {
                         
-                        wrd->attrUse = 3;
+                        /* attribute  (no value) */
                         wrd->reg_type = '0';
-                        wrd->string = comb;
-                        wrd->length = strlen(comb);
-                        wrd->seqno--;
+                        wrd->attrUse = 3;
+                        wrd->string = xp->name;
+                        wrd->length = strlen(xp->name);
                         
+                        wrd->seqno--;
                         (*p->tokenAdd)(wrd);
-                    }
-                }                
-                for (xp = n->u.tag.attributes; xp; xp = xp->next)
-                {
+                        
+                        if (xp->value &&
+                            strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2) {
+                            
+                            /* attribute value exact */
+                            strcpy (comb, xp->name);
+                            strcat (comb, "=");
+                            strcat (comb, xp->value);
+                            
+                            wrd->attrUse = 3;
+                            wrd->reg_type = '0';
+                            wrd->string = comb;
+                            wrd->length = strlen(comb);
+                            wrd->seqno--;
+                            
+                            (*p->tokenAdd)(wrd);
+                        }
+                    }                
+                    i++;
+                }
+                
+                i = 0;
+                for (xp = n->u.tag.attributes; xp; xp = xp->next) {
+                    data1_termlist *tl;
                     char attr_tag_path_full[1024];
                     int int_len = flen;
+                    int xpdone = 0;
                     
                     sprintf (attr_tag_path_full, "@%s/%.*s",
                              xp->name, int_len, tag_path_full);
-                    wrd->reg_type = '0';
-                    wrd->attrUse = 1;
-                    wrd->string = attr_tag_path_full;
-                    wrd->length = strlen(attr_tag_path_full);
-                    (*p->tokenAdd)(wrd);
                     
-		    if (xp->value)
- 		    {
-                        /* the same jokes, as with the data nodes ... */
-                        data1_termlist *tl;
-                        int xpdone = 0;
-                        
-                        wrd->string = xp->value;
-                        wrd->length = strlen(xp->value);
-                        wrd->reg_type = 'w';
-                        
-                        if ((tl = xpath_termlist_by_tagpath(attr_tag_path_full,
-                                                           n))) {
-                            for (; tl; tl = tl->next) {
-                                wrd->reg_type = *tl->structure;
-                                if (!tl->att) {
-                                    wrd->attrSet = VAL_IDXPATH;
-                                    wrd->attrUse = 1015;
-                                    (*p->tokenAdd)(wrd);
-                                    xpdone = 1;
-                                } else {
-                                    wrd->attrSet = (int) (tl->att->parent->reference);
+                    if ((tl = tll[i]))
+                    {
+                        /* If there is a termlist given (=xelm directive) */
+                        for (; tl; tl = tl->next)
+                        {
+                            if (!tl->att) {
+                                /* add xpath index for the attribute */
+                                index_xpath_attr (attr_tag_path_full, xp->name,
+                                                  xp->value, tl->structure,
+                                                  p, wrd);
+                                xpdone = 1;
+                            } else {
+                                /* add attribute based index for the attribute */
+                                if (xp->value) {
+                                    wrd->attrSet = (int) 
+                                        (tl->att->parent->reference);
                                     wrd->attrUse = tl->att->locals->local;
+                                    wrd->reg_type = *tl->structure;
+                                    wrd->string = xp->value;
+                                    wrd->length = strlen(xp->value);
                                     (*p->tokenAdd)(wrd);
                                 }
                             }
-                            
-                        } 
-                        if (!xpdone) {
-                            wrd->attrSet = VAL_IDXPATH;
-                            wrd->attrUse = 1015;
-                            wrd->reg_type = 'w';
-                            (*p->tokenAdd)(wrd);
                         }
                     }
-                    
-		    wrd->attrSet = VAL_IDXPATH;
-                    wrd->reg_type = '0';
-                    wrd->attrUse = 2;
-                    wrd->string = attr_tag_path_full;
-                    wrd->length = strlen(attr_tag_path_full);
-                    (*p->tokenAdd)(wrd);
+                    /* if there was no termlist for the given path, 
+                       or the termlist didn't have a ! element, index 
+                       the attribute as "w" */
+                    if ((!xpdone) && (!termlist_only))
+                    {
+                        index_xpath_attr (attr_tag_path_full, xp->name,
+                                          xp->value,  "w", p, wrd);
+                    }
+                    i++;
                 }
             }
-        }
+	}
     }
 }
 
@@ -562,11 +648,11 @@ static void index_termlist (data1_node *par, data1_node *n,
 		       tlist->att->parent->name,
 		       tlist->att->name, tlist->att->value,
 		       tlist->source);
-		printf (" data=\"");
-		for (i = 0; i<wrd->length && i < 8; i++)
+		printf (" XData:\"");
+		for (i = 0; i<wrd->length && i < 40; i++)
 		    fputc (wrd->string[i], stdout);
 		fputc ('"', stdout);
-		if (wrd->length > 8)
+		if (wrd->length > 40)
 		    printf (" ...");
 		fputc ('\n', stdout);
 	    }
@@ -627,13 +713,9 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level,
 	{
             index_termlist (n, n, p, level, wrd);
             /* index start tag */
-            assert (n->root->u.root.absyn);
-            
-            if (!n->root->u.root.absyn)
-                index_xpath (n, p, level, wrd, 1);
-            else if (n->root->u.root.absyn->enable_xpath_indexing)
-                index_xpath (n, p, level, wrd, 1);
-	}
+	    if (n->root->u.root.absyn)
+      	        index_xpath (n, p, level, wrd, 1);
+ 	}
 
 	if (n->child)
 	    if (dumpkeys(n->child, p, level + 1, wrd) < 0)
@@ -649,8 +731,8 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level,
 		printf("%*s", level * 4, "");
 		printf("Data: ");
 		if (n->u.data.len > 256)
-		    printf("'%.240s ... %.6s'\n", n->u.data.data,
-			   n->u.data.data + n->u.data.len-6);
+		    printf("'%.170s ... %.70s'\n", n->u.data.data,
+			   n->u.data.data + n->u.data.len-70);
 		else if (n->u.data.len > 0)
 		    printf("'%.*s'\n", n->u.data.len, n->u.data.data);
 		else
@@ -659,19 +741,14 @@ static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level,
 
 	    if (par)
 		index_termlist (par, n, p, level, wrd);
-            if (!n->root->u.root.absyn)
-                index_xpath (n, p, level, wrd, 1016);
-            else if (n->root->u.root.absyn->enable_xpath_indexing)
-                index_xpath (n, p, level, wrd, 1016);
+
+	    index_xpath (n, p, level, wrd, 1016);
  	}
 
 	if (n->which == DATA1N_tag)
 	{
             /* index end tag */
-            if (!n->root->u.root.absyn)
-                index_xpath (n, p, level, wrd, 2);
-            else if (n->root->u.root.absyn->enable_xpath_indexing)
-                index_xpath (n, p, level, wrd, 2);
+	    index_xpath (n, p, level, wrd, 2);
 	}
 
 	if (p->flagShowRecords && n->which == DATA1N_root)
@@ -736,6 +813,7 @@ static int grs_extract_sub(struct grs_handlers *h, struct recExtractCtrl *p,
         if ((oid_ent_to_oid (&oe, oidtmp)))
             (*p->schemaAdd)(p, oidtmp);
     }
+    data1_concat_text(p->dh, mem, n);
 
     /* ensure our data1 tree is UTF-8 */
     data1_iconv (p->dh, mem, n, "UTF-8", data1_get_encoding(p->dh, n));
@@ -925,6 +1003,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p)
         nmem_destroy (mem);
 	return 0;
     }
+    data1_concat_text(p->dh, mem, node);
+
     /* ensure our data1 tree is UTF-8 */
     data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node));