X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=recctrl%2Frectext.c;h=f102e7cde287c35a11161f4a750345824bdc81b6;hb=ec84578232c3de8ad75cc35638bed1bbe14fb7f7;hp=0dba6455a67568022ea0190f718b1a93705e5414;hpb=94ad502634fd97a08c4af0731eb3eddca13a3691;p=idzebra-moved-to-github.git

diff --git a/recctrl/rectext.c b/recctrl/rectext.c
index 0dba645..f102e7c 100644
--- a/recctrl/rectext.c
+++ b/recctrl/rectext.c
@@ -1,10 +1,50 @@
 /*
- * Copyright (C) 1994-1995, Index Data I/S 
+ * Copyright (C) 1994-2001, Index Data 
  * All rights reserved.
  * Sebastian Hammer, Adam Dickmeiss
  *
  * $Log: rectext.c,v $
- * Revision 1.2  1996-10-29 14:02:45  adam
+ * Revision 1.14  2001-01-22 11:41:41  adam
+ * Added support for raw retrieval (element set name "R").
+ *
+ * Revision 1.13  1999/09/07 07:19:21  adam
+ * Work on character mapping. Implemented replace rules.
+ *
+ * Revision 1.12  1999/05/26 07:49:14  adam
+ * C++ compilation.
+ *
+ * Revision 1.11  1999/05/21 12:00:17  adam
+ * Better diagnostics for extraction process.
+ *
+ * Revision 1.10  1999/05/20 12:57:18  adam
+ * Implemented TCL filter. Updated recctrl system.
+ *
+ * Revision 1.9  1998/10/16 08:14:38  adam
+ * Updated record control system.
+ *
+ * Revision 1.8  1998/05/20 10:12:27  adam
+ * Implemented automatic EXPLAIN database maintenance.
+ * Modified Zebra to work with ASN.1 compiled version of YAZ.
+ *
+ * Revision 1.7  1998/03/11 11:19:05  adam
+ * Changed the way sequence numbers are generated.
+ *
+ * Revision 1.6  1998/02/10 12:03:06  adam
+ * Implemented Sort.
+ *
+ * Revision 1.5  1997/10/27 14:33:06  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
+ * operators.
+ *
+ * Revision 1.4  1996/11/04 14:09:16  adam
+ * Minor changes.
+ *
+ * Revision 1.3  1996/11/01 09:00:33  adam
+ * This simple "text" format now supports element specs B and M.
+ *
+ * Revision 1.2  1996/10/29 14:02:45  adam
  * Uses buffered read to speed up things.
  *
  * Revision 1.1  1996/10/11 10:57:28  adam
@@ -42,7 +82,12 @@
 #include <zebrautl.h>
 #include "rectext.h"
 
-static void text_init (void)
+static void *text_init (RecType recType)
+{
+    return 0;
+}
+
+static void text_destroy (void *clientData)
 {
 }
 
@@ -55,10 +100,10 @@ struct buf_info {
 
 struct buf_info *buf_open (struct recExtractCtrl *p)
 {
-    struct buf_info *fi = xmalloc (sizeof(*fi));
+    struct buf_info *fi = (struct buf_info *) xmalloc (sizeof(*fi));
 
     fi->p = p;
-    fi->buf = xmalloc (4096);
+    fi->buf = (char *) xmalloc (4096);
     fi->offset = 1;
     fi->max = 1;
     return fi;
@@ -66,10 +111,10 @@ struct buf_info *buf_open (struct recExtractCtrl *p)
 
 int buf_read (struct buf_info *fi, char *dst)
 {
-    if (fi->max <= 0)
-        return 0;
     if (fi->offset >= fi->max)
     {
+        if (fi->max <= 0)
+            return 0;
         fi->max = (*fi->p->readf)(fi->p->fh, fi->buf, 4096);
         fi->offset = 0;
         if (fi->max <= 0)
@@ -85,47 +130,52 @@ void buf_close (struct buf_info *fi)
     xfree (fi);
 }
 
-static int text_extract (struct recExtractCtrl *p)
+static int text_extract (void *clientData, struct recExtractCtrl *p)
 {
-    char w[256];
+    char w[512];
     RecWord recWord;
-    int r, seqno = 1;
+    int r;
     struct buf_info *fi = buf_open (p);
 
-    (*p->init)(&recWord);
-    recWord.which = Word_String;
+    (*p->init)(p, &recWord);
+    recWord.reg_type = 'w';
     do
     {
         int i = 0;
             
         r = buf_read (fi, w);
-        while (r > 0 && i < 255 && isalnum(w[i]))
+        while (r > 0 && i < 511 && w[i] != '\n' && w[i] != '\r')
         {
             i++;
             r = buf_read (fi, w + i);
         }
         if (i)
         {
-            int j;
-            for (j = 0; j<i; j++)
-                w[j] = tolower(w[j]);
-            w[i] = 0;
-            recWord.seqno = seqno++;
-            recWord.u.string = w;
-            (*p->add)(&recWord);
+            recWord.string = w;
+	    recWord.length = i;
+            (*p->tokenAdd)(&recWord);
         }
     } while (r > 0);
     buf_close (fi);
-    return 0;
+    return RECCTRL_EXTRACT_OK;
 }
 
-static int text_retrieve (struct recRetrieveCtrl *p)
+static int text_retrieve (void *clientData, struct recRetrieveCtrl *p)
 {
     int r, text_ptr = 0;
     static char *text_buf = NULL;
     static int text_size = 0;
     int start_flag = 1;
+    const char *elementSetName = NULL;
+    int no_lines = 0;
+
+    if (p->comp && p->comp->which == Z_RecordComp_simple &&
+        p->comp->u.simple->which == Z_ElementSetNames_generic)
+        elementSetName = p->comp->u.simple->u.generic;
 
+    /* don't make header for the R(aw) element set name */
+    if (elementSetName && !strcmp(elementSetName, "R"))
+        start_flag = 0;
     while (1)
     {
         if (text_ptr + 4096 >= text_size)
@@ -133,7 +183,7 @@ static int text_retrieve (struct recRetrieveCtrl *p)
             char *nb;
 
             text_size = 2*text_size + 8192;
-            nb = xmalloc (text_size);
+            nb = (char *) xmalloc (text_size);
             if (text_buf)
             {
                 memcpy (nb, text_buf, text_ptr);
@@ -157,6 +207,27 @@ static int text_retrieve (struct recRetrieveCtrl *p)
             break;
         text_ptr += r;
     }
+    text_buf[text_ptr] = '\0';
+    if (elementSetName)
+    {
+        if (!strcmp (elementSetName, "B"))
+            no_lines = 4;
+        if (!strcmp (elementSetName, "M"))
+            no_lines = 20;
+    }
+    if (no_lines)
+    {
+        char *p = text_buf;
+        int i = 0;
+
+        while (++i <= no_lines && (p = strchr (p, '\n')))
+            p++;
+        if (p)
+        {
+            p[1] = '\0';
+            text_ptr = p-text_buf;
+        }
+    }
     p->output_format = VAL_SUTRS;
     p->rec_buf = text_buf;
     p->rec_len = text_ptr; 
@@ -166,6 +237,7 @@ static int text_retrieve (struct recRetrieveCtrl *p)
 static struct recType text_type = {
     "text",
     text_init,
+    text_destroy,
     text_extract,
     text_retrieve
 };