2 * Copyright (c) 1995-2000, 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 1.5 2000-11-29 14:22:47 adam
28 * Implemented XML/SGML attributes for data1 so that d1_read reads them
29 * and d1_write generates proper attributes for XML/SGML records. Added
30 * register locking for threaded version.
32 * Revision 1.4 2000/02/28 11:20:06 adam
33 * Using autoconf. New definitions: YAZ_BEGIN_CDECL/YAZ_END_CDECL.
35 * Revision 1.3 2000/01/04 17:46:17 ian
36 * Added function to count occurences of a tag spec in a data1 tree.
38 * Revision 1.2 1999/12/21 14:16:19 ian
39 * Changed retrieval module to allow data1 trees with no associated absyn.
40 * Also added a simple interface for extracting values from data1 trees using
41 * a string based tagpath.
43 * Revision 1.1 1999/11/30 13:47:11 adam
44 * Improved installation. Moved header files to include/yaz.
46 * Revision 1.42 1999/10/21 12:06:28 adam
47 * Retrieval module no longer uses ctype.h - functions.
49 * Revision 1.41 1999/07/13 13:23:47 adam
50 * Non-recursive version of data1_read_node. data1_read_nodex reads
51 * stream of bytes (instead of buffer in memory).
53 * Revision 1.40 1998/11/03 10:14:12 adam
54 * Changed definition of data1 node so that it uses less space.
56 * Revision 1.39 1998/10/28 15:10:06 adam
57 * Added --with-yc option to configure. For the data1_node in data1.h:
58 * decreased size of localdata and removed member "line" which wasn't useful.
60 * Revision 1.38 1998/10/15 08:29:15 adam
61 * Tag set type may be specified in reference to it using "tagset"
62 * directive in .abs-files and "include" directive in .tag-files.
64 * Revision 1.37 1998/10/13 16:09:46 adam
65 * Added support for arbitrary OID's for tagsets, schemas and attribute sets.
66 * Added support for multiple attribute set references and tagset references
67 * from an abstract syntax file.
68 * Fixed many bad logs-calls in routines that read the various
69 * specifications regarding data1 (*.abs,*.att,...) and made the messages
70 * consistent whenever possible.
71 * Added extra 'lineno' argument to function readconf_line.
73 * Revision 1.36 1998/05/18 13:06:57 adam
74 * Changed the way attribute sets are handled by the retriaval module.
75 * Extended Explain conversion / schema.
76 * Modified server and client to work with ASN.1 compiled protocol handlers.
78 * Revision 1.35 1998/03/05 08:15:32 adam
79 * Implemented data1_add_insert_taggeddata utility which is more flexible
80 * than data1_insert_taggeddata.
82 * Revision 1.34 1998/02/27 14:08:04 adam
83 * Added const to some char pointer arguments.
84 * Reworked data1_read_node so that it doesn't create a tree with
85 * pointers to original "SGML"-buffer.
87 * Revision 1.33 1997/12/18 10:51:30 adam
88 * Implemented sub-trees feature for schemas - including forward
91 * Revision 1.32 1997/12/09 16:18:16 adam
92 * Work on EXPLAIN schema. First implementation of sub-schema facility
95 * Revision 1.31 1997/11/18 09:51:08 adam
96 * Removed element num_children from data1_node. Minor changes in
99 * Revision 1.30 1997/10/31 12:20:07 adam
100 * Improved memory debugging for xmalloc/nmem.c. References to NMEM
101 * instead of ODR in n ESPEC-1 handling in source d1_espec.c.
102 * Bug fix: missing fclose in data1_read_espec1.
104 * Revision 1.29 1997/10/27 13:54:18 adam
105 * Changed structure field in data1 node to be simple string which
106 * is "unknown" to the retrieval system itself.
108 * Revision 1.28 1997/10/06 09:37:53 adam
109 * Added prototype for data1_get_map_buf.
111 * Revision 1.27 1997/09/24 13:35:44 adam
112 * Added two members to data1_marctab to ease reading of weird MARC records.
114 * Revision 1.26 1997/09/17 12:10:32 adam
117 * Revision 1.25 1997/09/05 09:50:55 adam
118 * Removed global data1_tabpath - uses data1_get_tabpath() instead.
120 * Revision 1.24 1997/09/01 09:30:39 adam
121 * Added include of yaz-util.h.
123 * Revision 1.23 1997/09/01 08:58:04 adam
124 * Removed declaration of data1_matchstr since it's a macro.
126 * Revision 1.22 1997/09/01 08:49:48 adam
127 * New windows NT/95 port using MSV5.0. To export DLL functions the
128 * YAZ_EXPORT modifier was added. Defined in yconfig.h.
130 * Revision 1.21 1997/05/14 06:53:38 adam
133 * Revision 1.20 1996/10/29 13:34:39 adam
134 * New functions to get/set data1_tabpath.
136 * Revision 1.19 1996/10/11 11:57:16 quinn
139 * Revision 1.18 1996/10/07 15:29:16 quinn
142 * Revision 1.17 1996/07/06 19:58:32 quinn
143 * System headerfiles gathered in yconfig
145 * Revision 1.16 1996/06/10 08:55:34 quinn
146 * Added Summary. Unfinished work
148 * Revision 1.15 1996/06/03 09:46:03 quinn
151 * Revision 1.14 1996/05/09 07:27:11 quinn
152 * Multiple local values supported.
154 * Revision 1.13 1996/02/20 16:32:48 quinn
157 * Revision 1.12 1996/01/18 09:46:34 adam
158 * Changed prototype for reader function parsed to data1_read_record.
160 * Revision 1.11 1995/12/15 16:19:45 quinn
161 * Added formatted_text.
163 * Revision 1.10 1995/12/14 11:09:43 quinn
166 * Revision 1.9 1995/12/13 15:32:47 quinn
169 * Revision 1.8 1995/12/13 13:44:23 quinn
170 * Modified Data1-system to use nmem
172 * Revision 1.7 1995/12/12 16:37:05 quinn
173 * Added destroy element to data1_node.
175 * Revision 1.6 1995/12/11 15:22:12 quinn
176 * Added last_child field to the node.
178 * Revision 1.5 1995/12/05 14:26:40 quinn
179 * Added global lbuf to data1_node.
181 * Revision 1.4 1995/11/13 09:27:29 quinn
182 * Fiddling with the variant stuff.
184 * Revision 1.3 1995/11/01 16:34:52 quinn
185 * Making data1 look for tables in data1_tabpath
187 * Revision 1.2 1995/11/01 13:54:35 quinn
190 * Revision 1.1 1995/11/01 13:07:18 quinn
191 * Data1 module now lives in YAZ.
193 * Revision 1.13 1995/10/25 16:00:49 quinn
194 * USMARC support is now almost operational
196 * Revision 1.12 1995/10/16 14:02:59 quinn
197 * Changes to support element set names and espec1
199 * Revision 1.11 1995/10/13 16:05:10 quinn
200 * Adding Espec1-processing
202 * Revision 1.10 1995/10/11 14:53:46 quinn
205 * Revision 1.9 1995/10/10 16:27:59 quinn
206 * *** empty log message ***
208 * Revision 1.8 1995/10/06 16:44:14 quinn
209 * Work on attribute set mapping, etc.
211 * Revision 1.7 1995/10/06 12:58:36 quinn
214 * Revision 1.6 1995/10/04 09:29:51 quinn
215 * Adjustments to support USGS test data
217 * Revision 1.5 1995/10/03 17:56:44 quinn
220 * Revision 1.4 1995/10/02 14:55:43 quinn
221 * *** empty log message ***
223 * Revision 1.3 1995/09/15 14:41:43 quinn
226 * Revision 1.2 1995/09/14 15:18:14 quinn
229 * Revision 1.1 1995/09/12 11:24:33 quinn
230 * Beginning to add code for structured records.
240 #include <yaz/nmem.h>
242 #include <yaz/proto.h>
244 #include <yaz/d1_attset.h>
245 #include <yaz/d1_map.h>
246 #include <yaz/yaz-util.h>
247 #include <yaz/wrbuf.h>
249 #define d1_isspace(c) strchr(" \r\n\t\f", c)
250 #define d1_isdigit(c) ((c) <= '9' && (c) >= '0')
252 #define DATA1_USING_XATTR 1
256 #define data1_matchstr(s1, s2) yaz_matchstr(s1, s2)
258 #define DATA1_MAX_SYMBOL 31
260 typedef struct data1_name
263 struct data1_name *next;
266 typedef struct data1_absyn_cache_info *data1_absyn_cache;
267 typedef struct data1_attset_cache_info *data1_attset_cache;
269 typedef enum data1_datatype
277 DATA1K_generalizedtime,
284 typedef struct data1_marctab
289 char record_status[2];
290 char implementation_codes[5];
291 int indicator_length;
292 int identifier_length;
293 char user_systems[4];
295 int length_data_entry;
297 int length_implementation;
300 int force_indicator_length;
301 int force_identifier_length;
302 struct data1_marctab *next;
305 typedef struct data1_esetname
309 struct data1_esetname *next;
313 * Variant set definitions.
316 typedef struct data1_vartype
319 struct data1_varclass *zclass;
321 data1_datatype datatype;
322 struct data1_vartype *next;
325 typedef struct data1_varclass
328 struct data1_varset *set;
330 data1_vartype *types;
331 struct data1_varclass *next;
334 typedef struct data1_varset
338 data1_varclass *classes;
347 typedef struct data1_tag
350 #define DATA1T_numeric 1
351 #define DATA1T_string 2
360 struct data1_tagset *tagset;
361 struct data1_tag *next;
364 typedef struct data1_tagset data1_tagset;
368 int type; /* type of tagset in current context */
369 char *name; /* symbolic name */
371 data1_tag *tags; /* tags defined by this set */
372 data1_tagset *children; /* children */
373 data1_tagset *next; /* sibling */
376 typedef struct data1_termlist
380 struct data1_termlist *next;
384 * abstract syntax specification
387 typedef struct data1_element
391 data1_termlist *termlists;
393 struct data1_element *children;
394 struct data1_element *next;
397 typedef struct data1_sub_elements {
399 struct data1_sub_elements *next;
400 data1_element *elements;
401 } data1_sub_elements;
403 #if DATA1_USING_XATTR
404 typedef struct data1_xattr {
407 struct data1_xattr *next;
411 typedef struct data1_absyn
415 data1_tagset *tagset;
416 data1_attset *attset;
417 data1_varset *varset;
418 data1_esetname *esetnames;
419 data1_maptab *maptabs;
421 data1_sub_elements *sub_elements;
422 data1_element *main_elements;
426 * record data node (tag/data/variant)
429 typedef struct data1_node
431 /* the root of a record (containing global data) */
432 #define DATA1N_root 1
435 /* some data under a leaf tag or variant */
436 #define DATA1N_data 3
437 /* variant specification (a triple, actually) */
438 #define DATA1N_variant 4
446 data1_absyn *absyn; /* abstract syntax for this type */
452 data1_element *element;
453 int no_data_requested;
455 unsigned node_selected : 1;
456 unsigned make_variantlist : 1;
457 #if DATA1_USING_XATTR
458 data1_xattr *attributes;
464 char *data; /* filename or data */
467 #define DATA1I_inctxt 1
468 /* binary data inclusion */
469 #define DATA1I_incbin 2
471 #define DATA1I_text 3
474 /* object identifier */
477 unsigned formatted_text : 1; /* newlines are significant */
487 void (*destroy)(struct data1_node *n);
488 #define DATA1_LOCALDATA 12
489 char lbuf[DATA1_LOCALDATA]; /* small buffer for local data */
490 struct data1_node *next;
491 struct data1_node *child;
492 struct data1_node *last_child;
493 struct data1_node *parent;
494 struct data1_node *root;
497 YAZ_EXPORT data1_handle data1_create (void);
498 YAZ_EXPORT void data1_destroy(data1_handle dh);
499 YAZ_EXPORT data1_node *get_parent_tag(data1_handle dh, data1_node *n);
500 YAZ_EXPORT data1_node *data1_read_node(data1_handle dh, const char **buf,
502 YAZ_EXPORT data1_node *data1_read_nodex (data1_handle dh, NMEM m,
503 int (*get_byte)(void *fh), void *fh,
505 YAZ_EXPORT data1_node *data1_read_record(data1_handle dh,
506 int (*rf)(void *, char *, size_t),
508 YAZ_EXPORT data1_absyn *data1_read_absyn(data1_handle dh, const char *file);
509 YAZ_EXPORT data1_tag *data1_gettagbynum(data1_handle dh,
511 int type, int value);
512 YAZ_EXPORT data1_tagset *data1_empty_tagset (data1_handle dh);
513 YAZ_EXPORT data1_tagset *data1_read_tagset(data1_handle dh,
516 YAZ_EXPORT data1_element *data1_getelementbytagname(data1_handle dh,
518 data1_element *parent,
519 const char *tagname);
520 YAZ_EXPORT Z_GenericRecord *data1_nodetogr(data1_handle dh, data1_node *n,
523 YAZ_EXPORT data1_tag *data1_gettagbyname(data1_handle dh, data1_tagset *s,
525 YAZ_EXPORT void data1_free_tree(data1_handle dh, data1_node *t);
526 YAZ_EXPORT char *data1_nodetobuf(data1_handle dh, data1_node *n,
527 int select, int *len);
528 YAZ_EXPORT data1_node *data1_insert_taggeddata(data1_handle dh,
531 const char *tagname, NMEM m);
532 YAZ_EXPORT data1_node *data1_add_taggeddata(data1_handle dh, data1_node *root,
534 const char *tagname, NMEM m);
535 YAZ_EXPORT data1_datatype data1_maptype(data1_handle dh, char *t);
536 YAZ_EXPORT data1_varset *data1_read_varset(data1_handle dh, const char *file);
537 YAZ_EXPORT data1_vartype *data1_getvartypebyct(data1_handle dh,
539 char *zclass, char *type);
540 YAZ_EXPORT Z_Espec1 *data1_read_espec1(data1_handle dh, const char *file);
541 YAZ_EXPORT int data1_doespec1(data1_handle dh, data1_node *n, Z_Espec1 *e);
542 YAZ_EXPORT data1_esetname *data1_getesetbyname(data1_handle dh,
545 YAZ_EXPORT data1_element *data1_getelementbyname(data1_handle dh,
548 YAZ_EXPORT data1_node *data1_mk_node(data1_handle dh, NMEM m);
549 YAZ_EXPORT data1_node *data1_mk_node_type (data1_handle dh, NMEM m, int type);
550 YAZ_EXPORT data1_absyn *data1_get_absyn(data1_handle dh, const char *name);
551 YAZ_EXPORT data1_attset *data1_get_attset (data1_handle dh, const char *name);
552 YAZ_EXPORT data1_maptab *data1_read_maptab(data1_handle dh, const char *file);
553 YAZ_EXPORT data1_node *data1_map_record(data1_handle dh, data1_node *n,
554 data1_maptab *map, NMEM m);
555 YAZ_EXPORT data1_marctab *data1_read_marctab (data1_handle dh,
557 YAZ_EXPORT char *data1_nodetomarc(data1_handle dh, data1_marctab *p,
558 data1_node *n, int selected, int *len);
559 YAZ_EXPORT char *data1_nodetoidsgml(data1_handle dh, data1_node *n,
560 int select, int *len);
561 YAZ_EXPORT Z_ExplainRecord *data1_nodetoexplain(data1_handle dh,
562 data1_node *n, int select,
564 YAZ_EXPORT Z_BriefBib *data1_nodetosummary(data1_handle dh,
565 data1_node *n, int select,
567 YAZ_EXPORT char *data1_nodetosoif(data1_handle dh, data1_node *n, int select,
569 YAZ_EXPORT void data1_set_tabpath(data1_handle dh, const char *path);
570 YAZ_EXPORT const char *data1_get_tabpath(data1_handle dh);
572 YAZ_EXPORT WRBUF data1_get_wrbuf (data1_handle dp);
573 YAZ_EXPORT char **data1_get_read_buf (data1_handle dp, int **lenp);
574 YAZ_EXPORT char **data1_get_map_buf (data1_handle dp, int **lenp);
575 YAZ_EXPORT data1_absyn_cache *data1_absyn_cache_get (data1_handle dh);
576 YAZ_EXPORT data1_attset_cache *data1_attset_cache_get (data1_handle dh);
577 YAZ_EXPORT NMEM data1_nmem_get (data1_handle dh);
578 YAZ_EXPORT void data1_pr_tree (data1_handle dh, data1_node *n, FILE *out);
579 YAZ_EXPORT char *data1_insert_string (data1_handle dh, data1_node *res,
580 NMEM m, const char *str);
581 YAZ_EXPORT data1_node *data1_read_sgml (data1_handle dh, NMEM m,
583 YAZ_EXPORT void data1_absyn_trav (data1_handle dh, void *handle,
584 void (*fh)(data1_handle dh,
585 void *h, data1_absyn *a));
587 YAZ_EXPORT data1_attset *data1_attset_search_id (data1_handle dh, int id);
589 YAZ_EXPORT data1_node
590 *data1_add_insert_taggeddata(data1_handle dh, data1_node *root,
591 data1_node *at, const char *tagname, NMEM m,
592 int first_flag, int local_allowed);
594 YAZ_EXPORT char *data1_getNodeValue(data1_node* node, char* pTagPath);
595 YAZ_EXPORT data1_node *data1_LookupNode(data1_node* node, char* pTagPath);
596 YAZ_EXPORT int data1_CountOccurences(data1_node* node, char* pTagPath);