Added nfa.h to makefile.am. Work continues on the nfa

diff --git a/include/yaz/Makefile.am b/include/yaz/Makefile.am
index 7f56a0b..e515f18 100644 (file)
--- a/include/yaz/Makefile.am
+++ b/include/yaz/Makefile.am
@@ -1,7 +1,8 @@
-## $Id: Makefile.am,v 1.28 2006-05-02 20:47:45 adam Exp $
+## $Id: Makefile.am,v 1.29 2006-05-03 11:09:59 heikki Exp $

 
 pkginclude_HEADERS= backend.h ccl.h cql.h comstack.h \
 
 pkginclude_HEADERS= backend.h ccl.h cql.h comstack.h \

- diagbib1.h diagsrw.h sortspec.h log.h logrpn.h marcdisp.h nmem.h odr.h \
+ diagbib1.h diagsrw.h sortspec.h log.h logrpn.h marcdisp.h nfa.h \
+ nmem.h odr.h \

  oid.h options.h otherinfo.h pquery.h prt-ext.h querytowrbuf.h \
  readconf.h record_conv.h statserv.h \
  tcpip.h test.h unix.h tpath.h wrbuf.h xmalloc.h \
  oid.h options.h otherinfo.h pquery.h prt-ext.h querytowrbuf.h \
  readconf.h record_conv.h statserv.h \
  tcpip.h test.h unix.h tpath.h wrbuf.h xmalloc.h \

diff --git a/include/yaz/nfa.h b/include/yaz/nfa.h
index e89a405..a61b5c1 100644 (file)
--- a/include/yaz/nfa.h
+++ b/include/yaz/nfa.h
@@ -1,6 +1,6 @@
 /*  Copyright (C) 2006, Index Data ApS
  *  See the file LICENSE for details.
 /*  Copyright (C) 2006, Index Data ApS
  *  See the file LICENSE for details.

- *  $Id: nfa.h,v 1.1 2006-05-03 09:04:33 heikki Exp $
+ *  $Id: nfa.h,v 1.2 2006-05-03 11:09:59 heikki Exp $

  */
 
 /**
  */
 
 /**

@@ -91,16 +91,15 @@ void *yaz_nfa_get_result(
  *  Each state can be the beginning and/or ending of a backref
  *  sequence. This call sets those flags in the states. After matching,
  *  we can get hold of the backrefs that matched, and use them in our
  *  Each state can be the beginning and/or ending of a backref
  *  sequence. This call sets those flags in the states. After matching,
  *  we can get hold of the backrefs that matched, and use them in our

- *  translations.
+ *  translations. The backrefs start at 1, not zero!

  *
  *  \param n   the nfa
  *  \param s   the state to add to
  *  \param backref_number is the number of the back reference. 0 for clearing
  *  \param is_start is 1 for start of the backref, 0 for end
  *
  *  \param n   the nfa
  *  \param s   the state to add to
  *  \param backref_number is the number of the back reference. 0 for clearing
  *  \param is_start is 1 for start of the backref, 0 for end

- *  \return
- *     0 for OK
- *     1 if the backref is already set
- *     
+ *  \retval   0 for OK
+ *  \retval   1 if the backref is already set
+ *  \retval   2 for ending a backref that has not been started

  *     
  */
 
  *     
  */
 

diff --git a/src/nfa.c b/src/nfa.c
index 5e8d56e..24fb6aa 100644 (file)
--- a/src/nfa.c
+++ b/src/nfa.c
@@ -1,7 +1,7 @@
 /*  Copyright (C) 2006, Index Data ApS
  *  See the file LICENSE for details.
  * 
 /*  Copyright (C) 2006, Index Data ApS
  *  See the file LICENSE for details.
  * 

- *  $Id: nfa.c,v 1.1 2006-05-03 09:04:33 heikki Exp $ 
+ *  $Id: nfa.c,v 1.2 2006-05-03 11:10:00 heikki Exp $ 

  *
  *  This is a simple NFA-based system for character set normalizing
  *  in yaz and zebra. Unlike standard NFA's, this operates on ranges of
  *
  *  This is a simple NFA-based system for character set normalizing
  *  in yaz and zebra. Unlike standard NFA's, this operates on ranges of

@@ -81,6 +81,10 @@ yaz_nfa *yaz_nfa_init() {
     n->nmem=my_nmem;
     n->nstates=0;
     n->laststate=0;
     n->nmem=my_nmem;
     n->nstates=0;
     n->laststate=0;

+    n->firststate=0;
+    n->nbackrefs=0;
+    n->curr_backrefs=0;
+    n->best_backrefs=0;

     return n;
 }
 
     return n;
 }
 

@@ -132,11 +136,19 @@ int yaz_nfa_set_backref(yaz_nfa *n, yaz_nfa_state *s,
         if (s->backref_start)
             return 1;
         s->backref_start=backref_number;
         if (s->backref_start)
             return 1;
         s->backref_start=backref_number;

-        if (n->nbackrefs<backref_number)
+        if (n->nbackrefs<backref_number) {

             n->nbackrefs=backref_number;
             n->nbackrefs=backref_number;

+            n->curr_backrefs=0;
+            n->best_backrefs=0;
+            /* clear them just in case we have already matched on */
+            /* with this nfa, and created a too small backref table */
+            /* we will reallocate when matching. */
+        }

     } else {
         if (s->backref_end)
             return 1;
     } else {
         if (s->backref_end)
             return 1;

+        if (n->nbackrefs<backref_number) 
+            return 2; /* can't end a backref that has not been started */

         s->backref_end=backref_number;
     }
     return 0; /* ok */
         s->backref_end=backref_number;
     }
     return 0; /* ok */

@@ -245,6 +257,7 @@ yaz_nfa_state *yaz_nfa_add_sequence(yaz_nfa *n,
  * * * * * * */
 
 struct matcher {
  * * * * * * */
 
 struct matcher {

+    yaz_nfa *n; 

     yaz_nfa_char *longest;
     int bestnode;
     void *result;
     yaz_nfa_char *longest;
     int bestnode;
     void *result;

@@ -253,13 +266,19 @@ struct matcher {
 };
 
 /* Finds the longest match. In case of ties, chooses the one with the 
 };
 
 /* Finds the longest match. In case of ties, chooses the one with the 

- * lowest numbered state */
-static void match_state(yaz_nfa_state *s, 
+ * lowest numbered state. Keep track of the back references. Recursively
+ * traverses the NFA. Keeps track of the best hit it has found. */
+
+static void match_state(
+              yaz_nfa_state *s, 

               yaz_nfa_char *inchar,
               size_t incharsleft,
               struct matcher *m ) {
     yaz_nfa_transition *t=s->lasttrans;
               yaz_nfa_char *inchar,
               size_t incharsleft,
               struct matcher *m ) {
     yaz_nfa_transition *t=s->lasttrans;

-    
+    if (s->backref_start) 
+        m->n->curr_backrefs[s->backref_start].start=inchar;
+    if (s->backref_end) 
+        m->n->curr_backrefs[s->backref_end].end=inchar;

     if (t) {
         if (incharsleft) {
             do {
     if (t) {
         if (incharsleft) {
             do {

@@ -284,11 +303,19 @@ static void match_state(yaz_nfa_state *s,
         if ( (m->longest < inchar) ||  /* longer result */
              ((m->longest == inchar)&&(m->bestnode<s->num)) ){
               /* or as long, but with lower node number. Still better */
         if ( (m->longest < inchar) ||  /* longer result */
              ((m->longest == inchar)&&(m->bestnode<s->num)) ){
               /* or as long, but with lower node number. Still better */

+           int i;

            m->longest=inchar;
            m->bestnode=s->num;
            m->result=s->result;
            m->longest=inchar;
            m->bestnode=s->num;
            m->result=s->result;

+           if (m->n->curr_backrefs) 
+               for (i=0; i<m->n->nbackrefs;i++) 
+                   m->n->best_backrefs[i]=m->n->curr_backrefs[i];

         }
     }
         }
     }

+    if (s->backref_start) 
+        m->n->curr_backrefs[s->backref_start].start=0;
+    if (s->backref_end) 
+        m->n->curr_backrefs[s->backref_end].end=0;

 } /* match_state */
 
 int yaz_nfa_match(yaz_nfa *n, 
 } /* match_state */
 
 int yaz_nfa_match(yaz_nfa *n, 

@@ -298,14 +325,15 @@ int yaz_nfa_match(yaz_nfa *n,
     struct matcher m;
     if (!n->firststate)
         return YAZ_NFA_NOMATCH; 
     struct matcher m;
     if (!n->firststate)
         return YAZ_NFA_NOMATCH; 

+    m.n=n;

     m.longest=*inbuff;
     m.bestnode=n->nstates;
     m.result=0;
     m.errorcode=0;
     m.longest=*inbuff;
     m.bestnode=n->nstates;
     m.result=0;
     m.errorcode=0;

-    if (!n->curr_backrefs) {
-        int sz=sizeof( struct yaz_nfa_backref_info);
-        n->curr_backrefs=nmem_malloc( n->nmem, n->nbackrefs* sz);
-        n->best_backrefs=nmem_malloc( n->nmem, n->nbackrefs* sz);
+    if ((!n->curr_backrefs) && (n->nbackrefs)){
+        int sz=sizeof( struct yaz_nfa_backref_info) * (n->nbackrefs+1);
+        n->curr_backrefs=nmem_malloc( n->nmem, sz);
+        n->best_backrefs=nmem_malloc( n->nmem, sz);

     }
     
 
     }