recctrl/danbibr.c

   1 /* $Id: danbibr.c,v 1.3 2004-05-26 13:26:17 adam Exp $
   2    Copyright (C) 2004
   3    Index Data Aps
   4
   5 This file is part of the Zebra server.
   6
   7 Zebra is free software; you can redistribute it and/or modify it under
   8 the terms of the GNU General Public License as published by the Free
   9 Software Foundation; either version 2, or (at your option) any later
  10 version.
  11
  12 Zebra is distributed in the hope that it will be useful, but WITHOUT ANY
  13 WARRANTY; without even the implied warranty of MERCHANTABILITY or
  14 FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  15 for more details.
  16
  17 You should have received a copy of the GNU General Public License
  18 along with Zebra; see the file LICENSE.zebra.  If not, write to the
  19 Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA
  20 02111-1307, USA.
  21 */
  22
  23 #include <assert.h>
  24 #include <stdio.h>
  25 #include <stdlib.h>
  26
  27 #include <yaz/log.h>
  28
  29 #include "grsread.h"
  30
  31 #include <yaz/xmalloc.h>
  32 #include <yaz/log.h>
  33 #include <data1.h>
  34
  35 #define READ_CHUNK 200
  36
  37 struct danbibr_info {
  38     WRBUF rec_buf;
  39     char read_buf[READ_CHUNK+1];  /* space for \0 */
  40 };
  41
  42 static void *grs_init_danbib(void)
  43 {
  44     struct danbibr_info *p = (struct danbibr_info *) xmalloc (sizeof(*p));
  45
  46     p->rec_buf = wrbuf_alloc();
  47     wrbuf_puts(p->rec_buf, "");
  48     return p;
  49 }
  50
  51 static int read_rec(struct grs_read_info *p)
  52 {
  53     struct danbibr_info *info = p->clientData;
  54
  55     wrbuf_rewind(info->rec_buf);
  56     while(1)
  57     {
  58         char *cp_split = 0;
  59         int r = (*p->readf)(p->fh, info->read_buf, READ_CHUNK);
  60         if (r <= 0)
  61         {
  62             if (wrbuf_len(info->rec_buf) > 0)
  63                 return 1;
  64             else
  65                 return 0;
  66         }
  67         info->read_buf[r] = '\0';
  68         wrbuf_puts(info->rec_buf, info->read_buf);
  69
  70         cp_split = strstr(wrbuf_buf(info->rec_buf), "\n$");
  71         if (cp_split)
  72         {
  73             cp_split++; /* now at $ */
  74             if (p->endf)
  75                 (*p->endf)(p->fh, p->offset +
  76                            (cp_split - wrbuf_buf(info->rec_buf)));
  77
  78             cp_split[0] = '\0';
  79             return 1;
  80         }
  81     }
  82 }
  83
  84 static data1_node *mk_tree(struct grs_read_info *p, const char *rec_buf)
  85 {
  86     data1_node *root = data1_mk_root(p->dh, p->mem, "danbib");
  87     const char *cp = rec_buf;
  88
  89     root = data1_mk_tag(p->dh, p->mem, "danbib", 0, root);
  90
  91     if (1)  /* <text> all </text> */
  92     {
  93         data1_node *text_node = data1_mk_tag(p->dh, p->mem, "text", 0, root);
  94         data1_mk_text_n(p->dh, p->mem, rec_buf, strlen(rec_buf), text_node);
  95     }
  96     while (*cp)
  97     {
  98         const char *start_tag = cp;
  99         const char *start_text;
 100         if (*cp == '\n')
 101         {
 102             cp++;
 103             continue;
 104         }
 105         else if (*cp == ' ')  /* bad continuation */
 106         {
 107             while (*cp && *cp != '\n')
 108                 cp++;
 109         }
 110         else if (*cp == '$')  /* header */
 111         {
 112             int no = 1;
 113             cp++;
 114             start_text = cp;
 115             for(start_text = cp; *cp && *cp != '\n'; cp++)
 116                 if (*cp == ':')
 117                 {
 118                     if (start_text != cp)
 119                     {
 120                         char elemstr[20];
 121                         data1_node *hnode;
 122                         sprintf(elemstr, "head%d", no);
 123
 124                         hnode = data1_mk_tag(p->dh, p->mem, elemstr, 0, root);
 125                         data1_mk_text_n(p->dh, p->mem, start_text,
 126                                         cp - start_text, hnode);
 127                         start_text = cp+1;
 128                     }
 129                     no++;
 130                 }
 131         }
 132         else /* other */
 133         {
 134             while (*cp != ' ' && *cp && *cp != '\n')
 135                 cp++;
 136             if (*cp == ' ')
 137             {
 138                 data1_node *tag_node =
 139                     data1_mk_tag_n(p->dh, p->mem,
 140                                    start_tag, cp - start_tag, 0, root);
 141                 cp++;
 142                 start_text = cp;
 143                 while (*cp != '\n' && *cp)
 144                 {
 145                     if (*cp == '*' && cp[1]) /* subfield */
 146                     {
 147                         data1_node *sub_tag_node;
 148                         if (start_text != cp)
 149                             data1_mk_text_n(p->dh, p->mem, start_text,
 150                                             cp-start_text, tag_node);
 151                         cp++;
 152                         sub_tag_node =
 153                             data1_mk_tag_n(p->dh, p->mem, cp, 1, 0, tag_node);
 154                         cp++;
 155                         start_text = cp;
 156                         while (*cp)
 157                         {
 158                             if (*cp == '\n' && cp[1] == ' ')
 159                             {
 160                                 cp++;
 161                                 if (start_text != cp)
 162                                     data1_mk_text_n(p->dh, p->mem, start_text,
 163                                                     cp-start_text, sub_tag_node);
 164                                 while (*cp == ' ')
 165                                     cp++;
 166                                 start_text = cp;
 167                             }
 168                             else if (*cp == '\n')
 169                                 break;
 170                             else if (*cp == '*')
 171                                 break;
 172                             else
 173                                 cp++;
 174                         }
 175                         if (start_text != cp)
 176                             data1_mk_text_n(p->dh, p->mem, start_text,
 177                                             cp-start_text, sub_tag_node);
 178                         start_text = cp;
 179                     }
 180                     else
 181                         cp++;
 182                 }
 183                 if (start_text != cp)
 184                     data1_mk_text_n(p->dh, p->mem, start_text,
 185                                     cp-start_text, tag_node);
 186             }
 187         }
 188     }
 189     return root;
 190 }
 191
 192 static data1_node *grs_read_danbib (struct grs_read_info *p)
 193 {
 194     struct danbibr_info *info = p->clientData;
 195
 196     if (read_rec(p))
 197         return mk_tree(p, wrbuf_buf(info->rec_buf));
 198     return 0;
 199 }
 200
 201 static void grs_destroy_danbib(void *clientData)
 202 {
 203     struct danbibr_info *p = (struct danbibr_info *) clientData;
 204
 205     wrbuf_free(p->rec_buf, 1);
 206     xfree (p);
 207 }
 208
 209 static struct recTypeGrs danbib_type = {
 210     "danbib",
 211     grs_init_danbib,
 212     grs_destroy_danbib,
 213     grs_read_danbib
 214 };
 215
 216 RecTypeGrs recTypeGrs_danbib = &danbib_type;
 217