1 /* $Id: retrieve.c,v 1.57 2006-11-23 14:01:21 adam Exp $
2 Copyright (C) 1995-2006
5 This file is part of the Zebra server.
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
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
17 You should have received a copy of the GNU General Public License
18 along with this program; if not, write to the Free Software
19 Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
36 #include <yaz/diagbib1.h>
42 static int zebra_create_record_stream(ZebraHandle zh,
44 struct ZebraRecStream *stream){
46 RecordAttr *recordAttr = rec_init_attr(zh->reg->zei, *rec);
48 if ((*rec)->size[recInfo_storeData] > 0){
49 zebra_create_stream_mem(stream, (*rec)->info[recInfo_storeData],
50 (*rec)->size[recInfo_storeData]);
56 if (zh->path_reg && !yaz_is_abspath((*rec)->info[recInfo_filename])){
57 strcpy(full_rep, zh->path_reg);
58 strcat(full_rep, "/");
59 strcat(full_rep, (*rec)->info[recInfo_filename]);
62 strcpy(full_rep, (*rec)->info[recInfo_filename]);
64 if ((fd = open (full_rep, O_BINARY|O_RDONLY)) == -1){
65 yaz_log (YLOG_WARN|YLOG_ERRNO, "Retrieve fail; missing file: %s",
70 zebra_create_stream_fd(stream, fd, recordAttr->recordOffset);
77 static int parse_zebra_elem(const char *elem,
78 const char **index, size_t *index_len,
79 const char **type, size_t *type_len)
90 /* verify that '::' is in the beginning of *elem
91 and something more follows */
93 || !(elem +1) || ':' != *(elem +1)
94 || !(elem +2) || '\0' == *(elem +2))
97 /* pick out info from string after '::' */
99 cp = strchr(elem, ':');
101 if (!cp) /* index, no colon, no type */
104 *index_len = strlen(elem);
106 else if (cp[1] == '\0') /* colon, but no following type */
110 else /* index, colon and type */
113 *index_len = cp - elem;
115 *type_len = strlen(cp+1);
122 int zebra_special_index_fetch(ZebraHandle zh, zint sysno, ODR odr,
124 const char *elemsetname,
125 oid_value input_format,
126 oid_value *output_format,
127 char **rec_bufp, int *rec_lenp)
129 const char *retrieval_index;
130 size_t retrieval_index_len;
131 const char *retrieval_type;
132 size_t retrieval_type_len;
134 zebra_rec_keys_t keys;
136 /* set output variables before processing possible error states */
139 /* only accept XML and SUTRS requests */
140 if (input_format != VAL_TEXT_XML
141 && input_format != VAL_SUTRS){
142 yaz_log(YLOG_WARN, "unsupported format for element set zebra::%s",
144 *output_format = VAL_NONE;
145 return YAZ_BIB1_NO_SYNTAXES_AVAILABLE_FOR_THIS_REQUEST;
148 if (!parse_zebra_elem(elemsetname,
149 &retrieval_index, &retrieval_index_len,
150 &retrieval_type, &retrieval_type_len))
151 return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
153 if (retrieval_type_len != 0 && retrieval_type_len != 1)
155 return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
158 if (retrieval_index_len)
160 char retrieval_index_cstr[256];
162 if (retrieval_index_len < sizeof(retrieval_index_cstr) -1)
164 memcpy(retrieval_index_cstr, retrieval_index, retrieval_index_len);
165 retrieval_index_cstr[retrieval_index_len] = '\0';
167 if (zebraExplain_lookup_attr_str(zh->reg->zei,
168 zinfo_index_category_index,
169 (retrieval_type_len == 0 ? -1 :
171 retrieval_index_cstr) == -1)
172 return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
176 keys = zebra_rec_keys_open();
177 zebra_rec_keys_set_buf(keys, rec->info[recInfo_delKeys],
178 rec->size[recInfo_delKeys], 0);
180 wrbuf = wrbuf_alloc();
181 if (zebra_rec_keys_rewind(keys)){
184 struct it_key key_in;
186 if (input_format == VAL_TEXT_XML){
187 *output_format = VAL_TEXT_XML;
188 /*wrbuf_printf(wrbuf,
189 "<?xml version=\"1.0\" encoding=\"UTF-8\"?>\n");*/
193 "\"http://www.indexdata.com/zebra/\""
194 " sysno=\"" ZINT_FORMAT "\""
195 " set=\"zebra::index%s/\">\n",
198 else if (input_format == VAL_SUTRS)
199 *output_format = VAL_SUTRS;
202 while(zebra_rec_keys_read(keys, &str, &slen, &key_in)){
204 int ord = CAST_ZINT_TO_INT(key_in.mem[0]);
207 const char *string_index = 0;
208 size_t string_index_len;
209 char dst_buf[IT_MAX_WORD];
211 zebraExplain_lookup_ord(zh->reg->zei, ord, &index_type, &db,
213 string_index_len = strlen(string_index);
215 /* process only if index is not defined,
216 or if defined and matching */
217 if (retrieval_index == 0
218 || (string_index_len == retrieval_index_len
219 && !memcmp(string_index, retrieval_index,
222 /* process only if type is not defined, or is matching */
223 if (retrieval_type == 0
224 || (retrieval_type_len == 1
225 && retrieval_type[0] == index_type)){
228 zebra_term_untrans(zh, index_type, dst_buf, str);
229 if (strlen(dst_buf)){
231 if (input_format == VAL_TEXT_XML){
232 wrbuf_printf(wrbuf, " <index name=\"%s\"",
235 wrbuf_printf(wrbuf, " type=\"%c\"", index_type);
237 wrbuf_printf(wrbuf, " seq=\"" ZINT_FORMAT "\">",
238 key_in.mem[key_in.len -1]);
240 wrbuf_xmlputs(wrbuf, dst_buf);
241 wrbuf_printf(wrbuf, "</index>\n");
243 else if (input_format == VAL_SUTRS){
244 wrbuf_printf(wrbuf, "%s ", string_index);
246 wrbuf_printf(wrbuf, "%c", index_type);
248 for (i = 1; i < key_in.len; i++)
249 wrbuf_printf(wrbuf, " " ZINT_FORMAT,
252 /* zebra_term_untrans(zh, index_type, dst_buf, str); */
253 wrbuf_printf(wrbuf, " %s", dst_buf);
255 wrbuf_printf(wrbuf, "\n");
262 if (input_format == VAL_TEXT_XML)
263 wrbuf_printf(wrbuf, "</record>\n");
265 *rec_lenp = wrbuf_len(wrbuf);
266 *rec_bufp = odr_malloc(odr, *rec_lenp);
267 memcpy(*rec_bufp, wrbuf_buf(wrbuf), *rec_lenp);
268 wrbuf_free(wrbuf, 1);
269 zebra_rec_keys_close(keys);
275 int zebra_special_fetch(ZebraHandle zh, zint sysno, int score, ODR odr,
276 const char *elemsetname,
277 oid_value input_format,
278 oid_value *output_format,
279 char **rec_bufp, int *rec_lenp)
283 /* set output variables before processing possible error states */
286 /* only accept XML and SUTRS requests */
287 if (input_format != VAL_TEXT_XML
288 && input_format != VAL_SUTRS){
289 yaz_log(YLOG_WARN, "unsupported format for element set zebra::%s",
291 return YAZ_BIB1_NO_SYNTAXES_AVAILABLE_FOR_THIS_REQUEST;
294 /* processing zebra::meta::sysno elemset without fetching binary data */
295 if (elemsetname && 0 == strcmp(elemsetname, "meta::sysno"))
298 if (input_format == VAL_SUTRS){
299 sprintf(rec_str, ZINT_FORMAT, sysno);
300 *output_format = VAL_SUTRS;
302 else if (input_format == VAL_TEXT_XML){
303 sprintf(rec_str, "<record xmlns="
304 "\"http://www.indexdata.com/zebra/\""
305 " sysno=\"" ZINT_FORMAT "\""
306 " set=\"zebra::%s\"/>\n",
308 *output_format = VAL_TEXT_XML;
310 *rec_lenp = strlen(rec_str);
312 *rec_bufp = odr_strdup(odr, rec_str);
315 return YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
319 /* fetching binary record up for all other display elementsets */
320 rec = rec_get(zh->reg->records, sysno);
322 yaz_log(YLOG_WARN, "rec_get fail on sysno=" ZINT_FORMAT, sysno);
323 return YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
326 /* processing special elementsetnames zebra::meta:: */
327 if (elemsetname && 0 == strcmp(elemsetname, "meta")){
330 RecordAttr *recordAttr = rec_init_attr(zh->reg->zei, rec);
332 if (input_format == VAL_TEXT_XML){
333 *output_format = VAL_TEXT_XML;
337 "\"http://www.indexdata.com/zebra/\""
338 " sysno=\"" ZINT_FORMAT "\""
343 " rank=\"" ZINT_FORMAT "\""
345 " set=\"zebra::%s/\">\n",
347 rec->size[recInfo_databaseName], rec->info[recInfo_databaseName],
348 rec->size[recInfo_filename], rec->info[recInfo_filename],
349 rec->size[recInfo_fileType], rec->info[recInfo_fileType],
351 recordAttr->staticrank,
352 recordAttr->recordSize,
355 else if (input_format == VAL_SUTRS){
356 *output_format = VAL_SUTRS;
358 "sysno " ZINT_FORMAT "\n"
363 "rank " ZINT_FORMAT "\n"
367 rec->size[recInfo_databaseName], rec->info[recInfo_databaseName],
368 rec->size[recInfo_filename], rec->info[recInfo_filename],
369 rec->size[recInfo_fileType], rec->info[recInfo_fileType],
371 recordAttr->staticrank,
372 recordAttr->recordSize,
377 *rec_lenp = strlen(rec_str);
379 *rec_bufp = odr_strdup(odr, rec_str);
382 ret = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
389 /* processing special elementsetnames zebra::index:: */
390 if (elemsetname && 0 == strncmp(elemsetname, "index", 5)){
392 int ret = zebra_special_index_fetch(zh, sysno, odr, rec,
394 input_format, output_format,
401 /* processing special elementsetnames zebra::data */
402 if (elemsetname && 0 == strcmp(elemsetname, "data")){
403 struct ZebraRecStream stream;
404 RecordAttr *recordAttr = rec_init_attr(zh->reg->zei, rec);
405 zebra_create_record_stream(zh, &rec, &stream);
406 *output_format = input_format;
407 *rec_lenp = recordAttr->recordSize;
408 *rec_bufp = (char *) odr_malloc(odr, *rec_lenp);
409 stream.readf(&stream, *rec_bufp, *rec_lenp);
410 stream.destroy(&stream);
417 return YAZ_BIB1_SPECIFIED_ELEMENT_SET_NAME_NOT_VALID_FOR_SPECIFIED_;
421 int zebra_record_fetch(ZebraHandle zh, zint sysno, int score,
422 zebra_snippets *hit_snippet, ODR odr,
423 oid_value input_format, Z_RecordComposition *comp,
424 oid_value *output_format,
425 char **rec_bufp, int *rec_lenp, char **basenamep,
429 char *fname, *file_type, *basename;
430 const char *elemsetname;
431 struct ZebraRecStream stream;
432 RecordAttr *recordAttr;
438 elemsetname = yaz_get_esn(comp);
440 /* processing zebra special elementset names of form 'zebra:: */
441 if (elemsetname && 0 == strncmp(elemsetname, "zebra::", 7))
442 return zebra_special_fetch(zh, sysno, score, odr,
444 input_format, output_format,
448 /* processing all other element set names */
449 rec = rec_get(zh->reg->records, sysno);
452 yaz_log(YLOG_WARN, "rec_get fail on sysno=" ZINT_FORMAT, sysno);
454 return YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
458 recordAttr = rec_init_attr(zh->reg->zei, rec);
460 file_type = rec->info[recInfo_fileType];
461 fname = rec->info[recInfo_filename];
462 basename = rec->info[recInfo_databaseName];
463 *basenamep = (char *) odr_malloc (odr, strlen(basename)+1);
464 strcpy (*basenamep, basename);
466 yaz_log(YLOG_DEBUG, "retrieve localno=" ZINT_FORMAT " score=%d",
469 zebra_create_record_stream(zh, &rec, &stream);
473 zebra_snippets *snippet;
474 zebra_rec_keys_t reckeys = zebra_rec_keys_open();
476 struct recRetrieveCtrl retrieveCtrl;
478 retrieveCtrl.stream = &stream;
479 retrieveCtrl.fname = fname;
480 retrieveCtrl.localno = sysno;
481 retrieveCtrl.staticrank = recordAttr->staticrank;
482 retrieveCtrl.score = score;
483 retrieveCtrl.recordSize = recordAttr->recordSize;
484 retrieveCtrl.odr = odr;
485 retrieveCtrl.input_format = retrieveCtrl.output_format = input_format;
486 retrieveCtrl.comp = comp;
487 retrieveCtrl.encoding = zh->record_encoding;
488 retrieveCtrl.diagnostic = 0;
489 retrieveCtrl.addinfo = 0;
490 retrieveCtrl.dh = zh->reg->dh;
491 retrieveCtrl.res = zh->res;
492 retrieveCtrl.rec_buf = 0;
493 retrieveCtrl.rec_len = -1;
494 retrieveCtrl.hit_snippet = hit_snippet;
495 retrieveCtrl.doc_snippet = zebra_snippets_create();
497 zebra_rec_keys_set_buf(reckeys,
498 rec->info[recInfo_delKeys],
499 rec->size[recInfo_delKeys],
501 zebra_rec_keys_to_snippets(zh, reckeys, retrieveCtrl.doc_snippet);
502 zebra_rec_keys_close(reckeys);
505 /* for debugging purposes */
506 yaz_log(YLOG_LOG, "DOC SNIPPET:");
507 zebra_snippets_log(retrieveCtrl.doc_snippet, YLOG_LOG);
508 yaz_log(YLOG_LOG, "HIT SNIPPET:");
509 zebra_snippets_log(retrieveCtrl.hit_snippet, YLOG_LOG);
511 snippet = zebra_snippets_window(retrieveCtrl.doc_snippet,
512 retrieveCtrl.hit_snippet,
515 /* for debugging purposes */
516 yaz_log(YLOG_LOG, "WINDOW SNIPPET:");
517 zebra_snippets_log(snippet, YLOG_LOG);
520 if (!(rt = recType_byName(zh->reg->recTypes, zh->res,
521 file_type, &clientData)))
523 return_code = YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS;
527 (*rt->retrieve)(clientData, &retrieveCtrl);
528 return_code = retrieveCtrl.diagnostic;
530 *output_format = retrieveCtrl.output_format;
531 *rec_bufp = (char *) retrieveCtrl.rec_buf;
532 *rec_lenp = retrieveCtrl.rec_len;
533 *addinfo = retrieveCtrl.addinfo;
536 zebra_snippets_destroy(snippet);
537 zebra_snippets_destroy(retrieveCtrl.doc_snippet);
540 stream.destroy(&stream);
549 * indent-tabs-mode: nil
551 * vim: shiftwidth=4 tabstop=8 expandtab