2 * Copyright (c) 1995-1998, Index Data.
4 * Permission to use, copy, modify, distribute, and sell this software and
5 * its documentation, in whole or in part, for any purpose, is hereby granted,
8 * 1. This copyright and permission notice appear in all copies of the
9 * software and its documentation. Notices of copyright or attribution
10 * which appear at the beginning of any file must remain unchanged.
12 * 2. The names of Index Data or the individual authors may not be used to
13 * endorse or promote products derived from this software without specific
14 * prior written permission.
16 * THIS SOFTWARE IS PROVIDED "AS IS" AND WITHOUT WARRANTY OF ANY KIND,
17 * EXPRESS, IMPLIED, OR OTHERWISE, INCLUDING WITHOUT LIMITATION, ANY
18 * WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.
19 * IN NO EVENT SHALL INDEX DATA BE LIABLE FOR ANY SPECIAL, INCIDENTAL,
20 * INDIRECT OR CONSEQUENTIAL DAMAGES OF ANY KIND, OR ANY DAMAGES
21 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER OR
22 * NOT ADVISED OF THE POSSIBILITY OF DAMAGE, AND ON ANY THEORY OF
23 * LIABILITY, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
27 * Revision 1.3 2000-01-04 17:46:17 ian
28 * Added function to count occurences of a tag spec in a data1 tree.
30 * Revision 1.2 1999/12/21 14:16:19 ian
31 * Changed retrieval module to allow data1 trees with no associated absyn.
32 * Also added a simple interface for extracting values from data1 trees using
33 * a string based tagpath.
35 * Revision 1.1 1999/11/30 13:47:11 adam
36 * Improved installation. Moved header files to include/yaz.
38 * Revision 1.42 1999/10/21 12:06:28 adam
39 * Retrieval module no longer uses ctype.h - functions.
41 * Revision 1.41 1999/07/13 13:23:47 adam
42 * Non-recursive version of data1_read_node. data1_read_nodex reads
43 * stream of bytes (instead of buffer in memory).
45 * Revision 1.40 1998/11/03 10:14:12 adam
46 * Changed definition of data1 node so that it uses less space.
48 * Revision 1.39 1998/10/28 15:10:06 adam
49 * Added --with-yc option to configure. For the data1_node in data1.h:
50 * decreased size of localdata and removed member "line" which wasn't useful.
52 * Revision 1.38 1998/10/15 08:29:15 adam
53 * Tag set type may be specified in reference to it using "tagset"
54 * directive in .abs-files and "include" directive in .tag-files.
56 * Revision 1.37 1998/10/13 16:09:46 adam
57 * Added support for arbitrary OID's for tagsets, schemas and attribute sets.
58 * Added support for multiple attribute set references and tagset references
59 * from an abstract syntax file.
60 * Fixed many bad logs-calls in routines that read the various
61 * specifications regarding data1 (*.abs,*.att,...) and made the messages
62 * consistent whenever possible.
63 * Added extra 'lineno' argument to function readconf_line.
65 * Revision 1.36 1998/05/18 13:06:57 adam
66 * Changed the way attribute sets are handled by the retriaval module.
67 * Extended Explain conversion / schema.
68 * Modified server and client to work with ASN.1 compiled protocol handlers.
70 * Revision 1.35 1998/03/05 08:15:32 adam
71 * Implemented data1_add_insert_taggeddata utility which is more flexible
72 * than data1_insert_taggeddata.
74 * Revision 1.34 1998/02/27 14:08:04 adam
75 * Added const to some char pointer arguments.
76 * Reworked data1_read_node so that it doesn't create a tree with
77 * pointers to original "SGML"-buffer.
79 * Revision 1.33 1997/12/18 10:51:30 adam
80 * Implemented sub-trees feature for schemas - including forward
83 * Revision 1.32 1997/12/09 16:18:16 adam
84 * Work on EXPLAIN schema. First implementation of sub-schema facility
87 * Revision 1.31 1997/11/18 09:51:08 adam
88 * Removed element num_children from data1_node. Minor changes in
91 * Revision 1.30 1997/10/31 12:20:07 adam
92 * Improved memory debugging for xmalloc/nmem.c. References to NMEM
93 * instead of ODR in n ESPEC-1 handling in source d1_espec.c.
94 * Bug fix: missing fclose in data1_read_espec1.
96 * Revision 1.29 1997/10/27 13:54:18 adam
97 * Changed structure field in data1 node to be simple string which
98 * is "unknown" to the retrieval system itself.
100 * Revision 1.28 1997/10/06 09:37:53 adam
101 * Added prototype for data1_get_map_buf.
103 * Revision 1.27 1997/09/24 13:35:44 adam
104 * Added two members to data1_marctab to ease reading of weird MARC records.
106 * Revision 1.26 1997/09/17 12:10:32 adam
109 * Revision 1.25 1997/09/05 09:50:55 adam
110 * Removed global data1_tabpath - uses data1_get_tabpath() instead.
112 * Revision 1.24 1997/09/01 09:30:39 adam
113 * Added include of yaz-util.h.
115 * Revision 1.23 1997/09/01 08:58:04 adam
116 * Removed declaration of data1_matchstr since it's a macro.
118 * Revision 1.22 1997/09/01 08:49:48 adam
119 * New windows NT/95 port using MSV5.0. To export DLL functions the
120 * YAZ_EXPORT modifier was added. Defined in yconfig.h.
122 * Revision 1.21 1997/05/14 06:53:38 adam
125 * Revision 1.20 1996/10/29 13:34:39 adam
126 * New functions to get/set data1_tabpath.
128 * Revision 1.19 1996/10/11 11:57:16 quinn
131 * Revision 1.18 1996/10/07 15:29:16 quinn
134 * Revision 1.17 1996/07/06 19:58:32 quinn
135 * System headerfiles gathered in yconfig
137 * Revision 1.16 1996/06/10 08:55:34 quinn
138 * Added Summary. Unfinished work
140 * Revision 1.15 1996/06/03 09:46:03 quinn
143 * Revision 1.14 1996/05/09 07:27:11 quinn
144 * Multiple local values supported.
146 * Revision 1.13 1996/02/20 16:32:48 quinn
149 * Revision 1.12 1996/01/18 09:46:34 adam
150 * Changed prototype for reader function parsed to data1_read_record.
152 * Revision 1.11 1995/12/15 16:19:45 quinn
153 * Added formatted_text.
155 * Revision 1.10 1995/12/14 11:09:43 quinn
158 * Revision 1.9 1995/12/13 15:32:47 quinn
161 * Revision 1.8 1995/12/13 13:44:23 quinn
162 * Modified Data1-system to use nmem
164 * Revision 1.7 1995/12/12 16:37:05 quinn
165 * Added destroy element to data1_node.
167 * Revision 1.6 1995/12/11 15:22:12 quinn
168 * Added last_child field to the node.
170 * Revision 1.5 1995/12/05 14:26:40 quinn
171 * Added global lbuf to data1_node.
173 * Revision 1.4 1995/11/13 09:27:29 quinn
174 * Fiddling with the variant stuff.
176 * Revision 1.3 1995/11/01 16:34:52 quinn
177 * Making data1 look for tables in data1_tabpath
179 * Revision 1.2 1995/11/01 13:54:35 quinn
182 * Revision 1.1 1995/11/01 13:07:18 quinn
183 * Data1 module now lives in YAZ.
185 * Revision 1.13 1995/10/25 16:00:49 quinn
186 * USMARC support is now almost operational
188 * Revision 1.12 1995/10/16 14:02:59 quinn
189 * Changes to support element set names and espec1
191 * Revision 1.11 1995/10/13 16:05:10 quinn
192 * Adding Espec1-processing
194 * Revision 1.10 1995/10/11 14:53:46 quinn
197 * Revision 1.9 1995/10/10 16:27:59 quinn
198 * *** empty log message ***
200 * Revision 1.8 1995/10/06 16:44:14 quinn
201 * Work on attribute set mapping, etc.
203 * Revision 1.7 1995/10/06 12:58:36 quinn
206 * Revision 1.6 1995/10/04 09:29:51 quinn
207 * Adjustments to support USGS test data
209 * Revision 1.5 1995/10/03 17:56:44 quinn
212 * Revision 1.4 1995/10/02 14:55:43 quinn
213 * *** empty log message ***
215 * Revision 1.3 1995/09/15 14:41:43 quinn
218 * Revision 1.2 1995/09/14 15:18:14 quinn
221 * Revision 1.1 1995/09/12 11:24:33 quinn
222 * Beginning to add code for structured records.
232 #include <yaz/nmem.h>
234 #include <yaz/proto.h>
236 #include <yaz/d1_attset.h>
237 #include <yaz/d1_map.h>
238 #include <yaz/yaz-util.h>
239 #include <yaz/wrbuf.h>
241 #define d1_isspace(c) strchr(" \r\n\t\f", c)
242 #define d1_isdigit(c) ((c) <= '9' && (c) >= '0')
247 #define data1_matchstr(s1, s2) yaz_matchstr(s1, s2)
249 #define DATA1_MAX_SYMBOL 31
251 typedef struct data1_name
254 struct data1_name *next;
257 typedef struct data1_absyn_cache_info *data1_absyn_cache;
258 typedef struct data1_attset_cache_info *data1_attset_cache;
260 typedef enum data1_datatype
268 DATA1K_generalizedtime,
275 typedef struct data1_marctab
280 char record_status[2];
281 char implementation_codes[5];
282 int indicator_length;
283 int identifier_length;
284 char user_systems[4];
286 int length_data_entry;
288 int length_implementation;
291 int force_indicator_length;
292 int force_identifier_length;
293 struct data1_marctab *next;
296 typedef struct data1_esetname
300 struct data1_esetname *next;
304 * Variant set definitions.
307 typedef struct data1_vartype
310 struct data1_varclass *zclass;
312 data1_datatype datatype;
313 struct data1_vartype *next;
316 typedef struct data1_varclass
319 struct data1_varset *set;
321 data1_vartype *types;
322 struct data1_varclass *next;
325 typedef struct data1_varset
329 data1_varclass *classes;
338 typedef struct data1_tag
341 #define DATA1T_numeric 1
342 #define DATA1T_string 2
351 struct data1_tagset *tagset;
352 struct data1_tag *next;
355 typedef struct data1_tagset data1_tagset;
359 int type; /* type of tagset in current context */
360 char *name; /* symbolic name */
362 data1_tag *tags; /* tags defined by this set */
363 data1_tagset *children; /* children */
364 data1_tagset *next; /* sibling */
367 typedef struct data1_termlist
371 struct data1_termlist *next;
375 * abstract syntax specification
378 typedef struct data1_element
382 data1_termlist *termlists;
384 struct data1_element *children;
385 struct data1_element *next;
388 typedef struct data1_sub_elements {
390 struct data1_sub_elements *next;
391 data1_element *elements;
392 } data1_sub_elements;
394 typedef struct data1_absyn
398 data1_tagset *tagset;
399 data1_attset *attset;
400 data1_varset *varset;
401 data1_esetname *esetnames;
402 data1_maptab *maptabs;
404 data1_sub_elements *sub_elements;
405 data1_element *main_elements;
409 * record data node (tag/data/variant)
412 typedef struct data1_node
414 /* the root of a record (containing global data) */
415 #define DATA1N_root 1
418 /* some data under a leaf tag or variant */
419 #define DATA1N_data 3
420 /* variant specification (a triple, actually) */
421 #define DATA1N_variant 4
429 data1_absyn *absyn; /* abstract syntax for this type */
435 data1_element *element;
436 int no_data_requested;
438 unsigned node_selected : 1;
439 unsigned make_variantlist : 1;
444 char *data; /* filename or data */
447 #define DATA1I_inctxt 1
448 /* binary data inclusion */
449 #define DATA1I_incbin 2
451 #define DATA1I_text 3
454 /* object identifier */
457 unsigned formatted_text : 1; /* newlines are significant */
467 void (*destroy)(struct data1_node *n);
468 #define DATA1_LOCALDATA 12
469 char lbuf[DATA1_LOCALDATA]; /* small buffer for local data */
470 struct data1_node *next;
471 struct data1_node *child;
472 struct data1_node *last_child;
473 struct data1_node *parent;
474 struct data1_node *root;
477 YAZ_EXPORT data1_handle data1_create (void);
478 YAZ_EXPORT void data1_destroy(data1_handle dh);
479 YAZ_EXPORT data1_node *get_parent_tag(data1_handle dh, data1_node *n);
480 YAZ_EXPORT data1_node *data1_read_node(data1_handle dh, const char **buf,
482 YAZ_EXPORT data1_node *data1_read_nodex (data1_handle dh, NMEM m,
483 int (*get_byte)(void *fh), void *fh,
485 YAZ_EXPORT data1_node *data1_read_record(data1_handle dh,
486 int (*rf)(void *, char *, size_t),
488 YAZ_EXPORT data1_absyn *data1_read_absyn(data1_handle dh, const char *file);
489 YAZ_EXPORT data1_tag *data1_gettagbynum(data1_handle dh,
491 int type, int value);
492 YAZ_EXPORT data1_tagset *data1_empty_tagset (data1_handle dh);
493 YAZ_EXPORT data1_tagset *data1_read_tagset(data1_handle dh,
496 YAZ_EXPORT data1_element *data1_getelementbytagname(data1_handle dh,
498 data1_element *parent,
499 const char *tagname);
500 YAZ_EXPORT Z_GenericRecord *data1_nodetogr(data1_handle dh, data1_node *n,
503 YAZ_EXPORT data1_tag *data1_gettagbyname(data1_handle dh, data1_tagset *s,
505 YAZ_EXPORT void data1_free_tree(data1_handle dh, data1_node *t);
506 YAZ_EXPORT char *data1_nodetobuf(data1_handle dh, data1_node *n,
507 int select, int *len);
508 YAZ_EXPORT data1_node *data1_insert_taggeddata(data1_handle dh,
511 const char *tagname, NMEM m);
512 YAZ_EXPORT data1_node *data1_add_taggeddata(data1_handle dh, data1_node *root,
514 const char *tagname, NMEM m);
515 YAZ_EXPORT data1_datatype data1_maptype(data1_handle dh, char *t);
516 YAZ_EXPORT data1_varset *data1_read_varset(data1_handle dh, const char *file);
517 YAZ_EXPORT data1_vartype *data1_getvartypebyct(data1_handle dh,
519 char *zclass, char *type);
520 YAZ_EXPORT Z_Espec1 *data1_read_espec1(data1_handle dh, const char *file);
521 YAZ_EXPORT int data1_doespec1(data1_handle dh, data1_node *n, Z_Espec1 *e);
522 YAZ_EXPORT data1_esetname *data1_getesetbyname(data1_handle dh,
525 YAZ_EXPORT data1_element *data1_getelementbyname(data1_handle dh,
528 YAZ_EXPORT data1_node *data1_mk_node(data1_handle dh, NMEM m);
529 YAZ_EXPORT data1_absyn *data1_get_absyn(data1_handle dh, const char *name);
530 YAZ_EXPORT data1_attset *data1_get_attset (data1_handle dh, const char *name);
531 YAZ_EXPORT data1_maptab *data1_read_maptab(data1_handle dh, const char *file);
532 YAZ_EXPORT data1_node *data1_map_record(data1_handle dh, data1_node *n,
533 data1_maptab *map, NMEM m);
534 YAZ_EXPORT data1_marctab *data1_read_marctab (data1_handle dh,
536 YAZ_EXPORT char *data1_nodetomarc(data1_handle dh, data1_marctab *p,
537 data1_node *n, int selected, int *len);
538 YAZ_EXPORT char *data1_nodetoidsgml(data1_handle dh, data1_node *n,
539 int select, int *len);
540 YAZ_EXPORT Z_ExplainRecord *data1_nodetoexplain(data1_handle dh,
541 data1_node *n, int select,
543 YAZ_EXPORT Z_BriefBib *data1_nodetosummary(data1_handle dh,
544 data1_node *n, int select,
546 YAZ_EXPORT char *data1_nodetosoif(data1_handle dh, data1_node *n, int select,
548 YAZ_EXPORT void data1_set_tabpath(data1_handle dh, const char *path);
549 YAZ_EXPORT const char *data1_get_tabpath(data1_handle dh);
551 YAZ_EXPORT WRBUF data1_get_wrbuf (data1_handle dp);
552 YAZ_EXPORT char **data1_get_read_buf (data1_handle dp, int **lenp);
553 YAZ_EXPORT char **data1_get_map_buf (data1_handle dp, int **lenp);
554 YAZ_EXPORT data1_absyn_cache *data1_absyn_cache_get (data1_handle dh);
555 YAZ_EXPORT data1_attset_cache *data1_attset_cache_get (data1_handle dh);
556 YAZ_EXPORT NMEM data1_nmem_get (data1_handle dh);
557 YAZ_EXPORT void data1_pr_tree (data1_handle dh, data1_node *n, FILE *out);
558 YAZ_EXPORT char *data1_insert_string (data1_handle dh, data1_node *res,
559 NMEM m, const char *str);
560 YAZ_EXPORT data1_node *data1_read_sgml (data1_handle dh, NMEM m,
562 YAZ_EXPORT void data1_absyn_trav (data1_handle dh, void *handle,
563 void (*fh)(data1_handle dh,
564 void *h, data1_absyn *a));
566 YAZ_EXPORT data1_attset *data1_attset_search_id (data1_handle dh, int id);
568 YAZ_EXPORT data1_node
569 *data1_add_insert_taggeddata(data1_handle dh, data1_node *root,
570 data1_node *at, const char *tagname, NMEM m,
571 int first_flag, int local_allowed);
573 YAZ_EXPORT char *data1_getNodeValue(data1_node* node, char* pTagPath);
574 YAZ_EXPORT data1_node *data1_LookupNode(data1_node* node, char* pTagPath);
575 YAZ_EXPORT int data1_CountOccurences(data1_node* node, char* pTagPath);