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 1.4 2000-02-28 11:20:06 adam
28 * Using autoconf. New definitions: YAZ_BEGIN_CDECL/YAZ_END_CDECL.
30 * Revision 1.3 2000/01/04 17:46:17 ian
31 * Added function to count occurences of a tag spec in a data1 tree.
33 * Revision 1.2 1999/12/21 14:16:19 ian
34 * Changed retrieval module to allow data1 trees with no associated absyn.
35 * Also added a simple interface for extracting values from data1 trees using
36 * a string based tagpath.
38 * Revision 1.1 1999/11/30 13:47:11 adam
39 * Improved installation. Moved header files to include/yaz.
41 * Revision 1.42 1999/10/21 12:06:28 adam
42 * Retrieval module no longer uses ctype.h - functions.
44 * Revision 1.41 1999/07/13 13:23:47 adam
45 * Non-recursive version of data1_read_node. data1_read_nodex reads
46 * stream of bytes (instead of buffer in memory).
48 * Revision 1.40 1998/11/03 10:14:12 adam
49 * Changed definition of data1 node so that it uses less space.
51 * Revision 1.39 1998/10/28 15:10:06 adam
52 * Added --with-yc option to configure. For the data1_node in data1.h:
53 * decreased size of localdata and removed member "line" which wasn't useful.
55 * Revision 1.38 1998/10/15 08:29:15 adam
56 * Tag set type may be specified in reference to it using "tagset"
57 * directive in .abs-files and "include" directive in .tag-files.
59 * Revision 1.37 1998/10/13 16:09:46 adam
60 * Added support for arbitrary OID's for tagsets, schemas and attribute sets.
61 * Added support for multiple attribute set references and tagset references
62 * from an abstract syntax file.
63 * Fixed many bad logs-calls in routines that read the various
64 * specifications regarding data1 (*.abs,*.att,...) and made the messages
65 * consistent whenever possible.
66 * Added extra 'lineno' argument to function readconf_line.
68 * Revision 1.36 1998/05/18 13:06:57 adam
69 * Changed the way attribute sets are handled by the retriaval module.
70 * Extended Explain conversion / schema.
71 * Modified server and client to work with ASN.1 compiled protocol handlers.
73 * Revision 1.35 1998/03/05 08:15:32 adam
74 * Implemented data1_add_insert_taggeddata utility which is more flexible
75 * than data1_insert_taggeddata.
77 * Revision 1.34 1998/02/27 14:08:04 adam
78 * Added const to some char pointer arguments.
79 * Reworked data1_read_node so that it doesn't create a tree with
80 * pointers to original "SGML"-buffer.
82 * Revision 1.33 1997/12/18 10:51:30 adam
83 * Implemented sub-trees feature for schemas - including forward
86 * Revision 1.32 1997/12/09 16:18:16 adam
87 * Work on EXPLAIN schema. First implementation of sub-schema facility
90 * Revision 1.31 1997/11/18 09:51:08 adam
91 * Removed element num_children from data1_node. Minor changes in
94 * Revision 1.30 1997/10/31 12:20:07 adam
95 * Improved memory debugging for xmalloc/nmem.c. References to NMEM
96 * instead of ODR in n ESPEC-1 handling in source d1_espec.c.
97 * Bug fix: missing fclose in data1_read_espec1.
99 * Revision 1.29 1997/10/27 13:54:18 adam
100 * Changed structure field in data1 node to be simple string which
101 * is "unknown" to the retrieval system itself.
103 * Revision 1.28 1997/10/06 09:37:53 adam
104 * Added prototype for data1_get_map_buf.
106 * Revision 1.27 1997/09/24 13:35:44 adam
107 * Added two members to data1_marctab to ease reading of weird MARC records.
109 * Revision 1.26 1997/09/17 12:10:32 adam
112 * Revision 1.25 1997/09/05 09:50:55 adam
113 * Removed global data1_tabpath - uses data1_get_tabpath() instead.
115 * Revision 1.24 1997/09/01 09:30:39 adam
116 * Added include of yaz-util.h.
118 * Revision 1.23 1997/09/01 08:58:04 adam
119 * Removed declaration of data1_matchstr since it's a macro.
121 * Revision 1.22 1997/09/01 08:49:48 adam
122 * New windows NT/95 port using MSV5.0. To export DLL functions the
123 * YAZ_EXPORT modifier was added. Defined in yconfig.h.
125 * Revision 1.21 1997/05/14 06:53:38 adam
128 * Revision 1.20 1996/10/29 13:34:39 adam
129 * New functions to get/set data1_tabpath.
131 * Revision 1.19 1996/10/11 11:57:16 quinn
134 * Revision 1.18 1996/10/07 15:29:16 quinn
137 * Revision 1.17 1996/07/06 19:58:32 quinn
138 * System headerfiles gathered in yconfig
140 * Revision 1.16 1996/06/10 08:55:34 quinn
141 * Added Summary. Unfinished work
143 * Revision 1.15 1996/06/03 09:46:03 quinn
146 * Revision 1.14 1996/05/09 07:27:11 quinn
147 * Multiple local values supported.
149 * Revision 1.13 1996/02/20 16:32:48 quinn
152 * Revision 1.12 1996/01/18 09:46:34 adam
153 * Changed prototype for reader function parsed to data1_read_record.
155 * Revision 1.11 1995/12/15 16:19:45 quinn
156 * Added formatted_text.
158 * Revision 1.10 1995/12/14 11:09:43 quinn
161 * Revision 1.9 1995/12/13 15:32:47 quinn
164 * Revision 1.8 1995/12/13 13:44:23 quinn
165 * Modified Data1-system to use nmem
167 * Revision 1.7 1995/12/12 16:37:05 quinn
168 * Added destroy element to data1_node.
170 * Revision 1.6 1995/12/11 15:22:12 quinn
171 * Added last_child field to the node.
173 * Revision 1.5 1995/12/05 14:26:40 quinn
174 * Added global lbuf to data1_node.
176 * Revision 1.4 1995/11/13 09:27:29 quinn
177 * Fiddling with the variant stuff.
179 * Revision 1.3 1995/11/01 16:34:52 quinn
180 * Making data1 look for tables in data1_tabpath
182 * Revision 1.2 1995/11/01 13:54:35 quinn
185 * Revision 1.1 1995/11/01 13:07:18 quinn
186 * Data1 module now lives in YAZ.
188 * Revision 1.13 1995/10/25 16:00:49 quinn
189 * USMARC support is now almost operational
191 * Revision 1.12 1995/10/16 14:02:59 quinn
192 * Changes to support element set names and espec1
194 * Revision 1.11 1995/10/13 16:05:10 quinn
195 * Adding Espec1-processing
197 * Revision 1.10 1995/10/11 14:53:46 quinn
200 * Revision 1.9 1995/10/10 16:27:59 quinn
201 * *** empty log message ***
203 * Revision 1.8 1995/10/06 16:44:14 quinn
204 * Work on attribute set mapping, etc.
206 * Revision 1.7 1995/10/06 12:58:36 quinn
209 * Revision 1.6 1995/10/04 09:29:51 quinn
210 * Adjustments to support USGS test data
212 * Revision 1.5 1995/10/03 17:56:44 quinn
215 * Revision 1.4 1995/10/02 14:55:43 quinn
216 * *** empty log message ***
218 * Revision 1.3 1995/09/15 14:41:43 quinn
221 * Revision 1.2 1995/09/14 15:18:14 quinn
224 * Revision 1.1 1995/09/12 11:24:33 quinn
225 * Beginning to add code for structured records.
235 #include <yaz/nmem.h>
237 #include <yaz/proto.h>
239 #include <yaz/d1_attset.h>
240 #include <yaz/d1_map.h>
241 #include <yaz/yaz-util.h>
242 #include <yaz/wrbuf.h>
244 #define d1_isspace(c) strchr(" \r\n\t\f", c)
245 #define d1_isdigit(c) ((c) <= '9' && (c) >= '0')
249 #define data1_matchstr(s1, s2) yaz_matchstr(s1, s2)
251 #define DATA1_MAX_SYMBOL 31
253 typedef struct data1_name
256 struct data1_name *next;
259 typedef struct data1_absyn_cache_info *data1_absyn_cache;
260 typedef struct data1_attset_cache_info *data1_attset_cache;
262 typedef enum data1_datatype
270 DATA1K_generalizedtime,
277 typedef struct data1_marctab
282 char record_status[2];
283 char implementation_codes[5];
284 int indicator_length;
285 int identifier_length;
286 char user_systems[4];
288 int length_data_entry;
290 int length_implementation;
293 int force_indicator_length;
294 int force_identifier_length;
295 struct data1_marctab *next;
298 typedef struct data1_esetname
302 struct data1_esetname *next;
306 * Variant set definitions.
309 typedef struct data1_vartype
312 struct data1_varclass *zclass;
314 data1_datatype datatype;
315 struct data1_vartype *next;
318 typedef struct data1_varclass
321 struct data1_varset *set;
323 data1_vartype *types;
324 struct data1_varclass *next;
327 typedef struct data1_varset
331 data1_varclass *classes;
340 typedef struct data1_tag
343 #define DATA1T_numeric 1
344 #define DATA1T_string 2
353 struct data1_tagset *tagset;
354 struct data1_tag *next;
357 typedef struct data1_tagset data1_tagset;
361 int type; /* type of tagset in current context */
362 char *name; /* symbolic name */
364 data1_tag *tags; /* tags defined by this set */
365 data1_tagset *children; /* children */
366 data1_tagset *next; /* sibling */
369 typedef struct data1_termlist
373 struct data1_termlist *next;
377 * abstract syntax specification
380 typedef struct data1_element
384 data1_termlist *termlists;
386 struct data1_element *children;
387 struct data1_element *next;
390 typedef struct data1_sub_elements {
392 struct data1_sub_elements *next;
393 data1_element *elements;
394 } data1_sub_elements;
396 typedef struct data1_absyn
400 data1_tagset *tagset;
401 data1_attset *attset;
402 data1_varset *varset;
403 data1_esetname *esetnames;
404 data1_maptab *maptabs;
406 data1_sub_elements *sub_elements;
407 data1_element *main_elements;
411 * record data node (tag/data/variant)
414 typedef struct data1_node
416 /* the root of a record (containing global data) */
417 #define DATA1N_root 1
420 /* some data under a leaf tag or variant */
421 #define DATA1N_data 3
422 /* variant specification (a triple, actually) */
423 #define DATA1N_variant 4
431 data1_absyn *absyn; /* abstract syntax for this type */
437 data1_element *element;
438 int no_data_requested;
440 unsigned node_selected : 1;
441 unsigned make_variantlist : 1;
446 char *data; /* filename or data */
449 #define DATA1I_inctxt 1
450 /* binary data inclusion */
451 #define DATA1I_incbin 2
453 #define DATA1I_text 3
456 /* object identifier */
459 unsigned formatted_text : 1; /* newlines are significant */
469 void (*destroy)(struct data1_node *n);
470 #define DATA1_LOCALDATA 12
471 char lbuf[DATA1_LOCALDATA]; /* small buffer for local data */
472 struct data1_node *next;
473 struct data1_node *child;
474 struct data1_node *last_child;
475 struct data1_node *parent;
476 struct data1_node *root;
479 YAZ_EXPORT data1_handle data1_create (void);
480 YAZ_EXPORT void data1_destroy(data1_handle dh);
481 YAZ_EXPORT data1_node *get_parent_tag(data1_handle dh, data1_node *n);
482 YAZ_EXPORT data1_node *data1_read_node(data1_handle dh, const char **buf,
484 YAZ_EXPORT data1_node *data1_read_nodex (data1_handle dh, NMEM m,
485 int (*get_byte)(void *fh), void *fh,
487 YAZ_EXPORT data1_node *data1_read_record(data1_handle dh,
488 int (*rf)(void *, char *, size_t),
490 YAZ_EXPORT data1_absyn *data1_read_absyn(data1_handle dh, const char *file);
491 YAZ_EXPORT data1_tag *data1_gettagbynum(data1_handle dh,
493 int type, int value);
494 YAZ_EXPORT data1_tagset *data1_empty_tagset (data1_handle dh);
495 YAZ_EXPORT data1_tagset *data1_read_tagset(data1_handle dh,
498 YAZ_EXPORT data1_element *data1_getelementbytagname(data1_handle dh,
500 data1_element *parent,
501 const char *tagname);
502 YAZ_EXPORT Z_GenericRecord *data1_nodetogr(data1_handle dh, data1_node *n,
505 YAZ_EXPORT data1_tag *data1_gettagbyname(data1_handle dh, data1_tagset *s,
507 YAZ_EXPORT void data1_free_tree(data1_handle dh, data1_node *t);
508 YAZ_EXPORT char *data1_nodetobuf(data1_handle dh, data1_node *n,
509 int select, int *len);
510 YAZ_EXPORT data1_node *data1_insert_taggeddata(data1_handle dh,
513 const char *tagname, NMEM m);
514 YAZ_EXPORT data1_node *data1_add_taggeddata(data1_handle dh, data1_node *root,
516 const char *tagname, NMEM m);
517 YAZ_EXPORT data1_datatype data1_maptype(data1_handle dh, char *t);
518 YAZ_EXPORT data1_varset *data1_read_varset(data1_handle dh, const char *file);
519 YAZ_EXPORT data1_vartype *data1_getvartypebyct(data1_handle dh,
521 char *zclass, char *type);
522 YAZ_EXPORT Z_Espec1 *data1_read_espec1(data1_handle dh, const char *file);
523 YAZ_EXPORT int data1_doespec1(data1_handle dh, data1_node *n, Z_Espec1 *e);
524 YAZ_EXPORT data1_esetname *data1_getesetbyname(data1_handle dh,
527 YAZ_EXPORT data1_element *data1_getelementbyname(data1_handle dh,
530 YAZ_EXPORT data1_node *data1_mk_node(data1_handle dh, NMEM m);
531 YAZ_EXPORT data1_absyn *data1_get_absyn(data1_handle dh, const char *name);
532 YAZ_EXPORT data1_attset *data1_get_attset (data1_handle dh, const char *name);
533 YAZ_EXPORT data1_maptab *data1_read_maptab(data1_handle dh, const char *file);
534 YAZ_EXPORT data1_node *data1_map_record(data1_handle dh, data1_node *n,
535 data1_maptab *map, NMEM m);
536 YAZ_EXPORT data1_marctab *data1_read_marctab (data1_handle dh,
538 YAZ_EXPORT char *data1_nodetomarc(data1_handle dh, data1_marctab *p,
539 data1_node *n, int selected, int *len);
540 YAZ_EXPORT char *data1_nodetoidsgml(data1_handle dh, data1_node *n,
541 int select, int *len);
542 YAZ_EXPORT Z_ExplainRecord *data1_nodetoexplain(data1_handle dh,
543 data1_node *n, int select,
545 YAZ_EXPORT Z_BriefBib *data1_nodetosummary(data1_handle dh,
546 data1_node *n, int select,
548 YAZ_EXPORT char *data1_nodetosoif(data1_handle dh, data1_node *n, int select,
550 YAZ_EXPORT void data1_set_tabpath(data1_handle dh, const char *path);
551 YAZ_EXPORT const char *data1_get_tabpath(data1_handle dh);
553 YAZ_EXPORT WRBUF data1_get_wrbuf (data1_handle dp);
554 YAZ_EXPORT char **data1_get_read_buf (data1_handle dp, int **lenp);
555 YAZ_EXPORT char **data1_get_map_buf (data1_handle dp, int **lenp);
556 YAZ_EXPORT data1_absyn_cache *data1_absyn_cache_get (data1_handle dh);
557 YAZ_EXPORT data1_attset_cache *data1_attset_cache_get (data1_handle dh);
558 YAZ_EXPORT NMEM data1_nmem_get (data1_handle dh);
559 YAZ_EXPORT void data1_pr_tree (data1_handle dh, data1_node *n, FILE *out);
560 YAZ_EXPORT char *data1_insert_string (data1_handle dh, data1_node *res,
561 NMEM m, const char *str);
562 YAZ_EXPORT data1_node *data1_read_sgml (data1_handle dh, NMEM m,
564 YAZ_EXPORT void data1_absyn_trav (data1_handle dh, void *handle,
565 void (*fh)(data1_handle dh,
566 void *h, data1_absyn *a));
568 YAZ_EXPORT data1_attset *data1_attset_search_id (data1_handle dh, int id);
570 YAZ_EXPORT data1_node
571 *data1_add_insert_taggeddata(data1_handle dh, data1_node *root,
572 data1_node *at, const char *tagname, NMEM m,
573 int first_flag, int local_allowed);
575 YAZ_EXPORT char *data1_getNodeValue(data1_node* node, char* pTagPath);
576 YAZ_EXPORT data1_node *data1_LookupNode(data1_node* node, char* pTagPath);
577 YAZ_EXPORT int data1_CountOccurences(data1_node* node, char* pTagPath);