X-Git-Url: http://git.indexdata.com/?p=idzebra-moved-to-github.git;a=blobdiff_plain;f=recctrl%2Frecgrs.c;h=c41906fcac47312827e1d40e42329b5fd140d7bf;hp=597488814a7d1111713ca8664146e286b37edb6e;hb=896c0427df9d8eff5de6a1735dcd992e067df844;hpb=15ded665c8ca37e3c7a99ce48c2e78acdb727bd6 diff --git a/recctrl/recgrs.c b/recctrl/recgrs.c index 5974888..c41906f 100644 --- a/recctrl/recgrs.c +++ b/recctrl/recgrs.c @@ -1,237 +1,26 @@ -/* - * Copyright (C) 1994-2001, Index Data - * All rights reserved. - * - * $Log: recgrs.c,v $ - * Revision 1.46 2002-04-13 18:16:43 adam - * More XPATH work; common sequence numbers for extract keys - * - * Revision 1.45 2002/04/12 14:40:42 adam - * Work on XPATH - * - * Revision 1.44 2002/04/11 20:09:47 adam - * work on string tag indexing - * - * Revision 1.43 2002/03/21 23:06:36 adam - * Source 'tag' in abs-file - * - * Revision 1.42 2002/02/20 17:30:01 adam - * Work on new API. Locking system re-implemented - * - * Revision 1.41 2001/05/22 21:01:47 adam - * Removed print of data1 tree on stdout so that inetd works again. - * - * Revision 1.40 2001/03/29 21:31:31 adam - * Fixed "record begin" for Tcl filter. - * - * Revision 1.39 2000/12/05 19:09:15 adam - * Fixed problem where indexer could crash if abstract syntax was undefined. - * - * Revision 1.38 2000/12/05 14:44:58 adam - * Fixed minor bug that could cause zmbol to break it data were emitted - * with not parent tags. - * - * Revision 1.37 2000/12/05 12:22:53 adam - * Termlist source implemented (so that we can index values of XML/SGML - * attributes). - * - * Revision 1.36 2000/12/05 10:01:44 adam - * Fixed bug regarding user-defined attribute sets. - * - * Revision 1.35 2000/11/29 15:21:31 adam - * Fixed problem with passwd db. - * - * Revision 1.34 2000/02/25 13:24:49 adam - * Fixed bug regarding pointer conversion that showed up on OSF V5. - * - * Revision 1.33 1999/11/30 13:48:04 adam - * Improved installation. Updated for inclusion of YAZ header files. - * - * Revision 1.32 1999/09/07 07:19:21 adam - * Work on character mapping. Implemented replace rules. - * - * Revision 1.31 1999/07/14 10:56:43 adam - * Fixed potential memory leak. - * - * Revision 1.30 1999/07/06 12:26:41 adam - * Retrieval handler obeys schema and handles XML transfer syntax. - * - * Revision 1.29 1999/05/26 07:49:14 adam - * C++ compilation. - * - * Revision 1.28 1999/05/21 12:00:17 adam - * Better diagnostics for extraction process. - * - * Revision 1.27 1999/05/20 12:57:18 adam - * Implemented TCL filter. Updated recctrl system. - * - * Revision 1.26 1999/03/02 16:15:44 quinn - * Added "tagsysno" and "tagrank" directives to zebra.cfg. - * - * Revision 1.25 1999/02/18 15:01:26 adam - * Minor changes. - * - * Revision 1.24 1999/02/02 14:51:28 adam - * Updated WIN32 code specific sections. Changed header. - * - * Revision 1.23 1998/10/18 07:51:10 adam - * Changed one logf call. - * - * Revision 1.22 1998/10/16 08:14:37 adam - * Updated record control system. - * - * Revision 1.21 1998/07/01 09:16:10 adam - * Element localno only added when it's greater than 0. - * - * Revision 1.20 1998/05/20 10:12:26 adam - * Implemented automatic EXPLAIN database maintenance. - * Modified Zebra to work with ASN.1 compiled version of YAZ. - * - * Revision 1.19 1998/03/11 11:19:05 adam - * Changed the way sequence numbers are generated. - * - * Revision 1.18 1998/03/05 08:41:31 adam - * Minor changes. - * - * Revision 1.17 1998/02/10 12:03:06 adam - * Implemented Sort. - * - * Revision 1.16 1998/01/29 13:38:17 adam - * Fixed problem with mapping to record with unknown schema. - * - * Revision 1.15 1998/01/26 10:37:57 adam - * Better diagnostics. - * - * Revision 1.14 1997/11/06 11:41:01 adam - * Implemented "begin variant" for the sgml.regx filter. - * - * Revision 1.13 1997/10/31 12:35:44 adam - * Added a few log statements. - * - * Revision 1.12 1997/10/29 12:02:22 adam - * Using oid_ent_to_oid used instead of the non thread-safe oid_getoidbyent. - * - * Revision 1.11 1997/10/27 14:34:00 adam - * Work on generic character mapping depending on "structure" field - * in abstract syntax file. - * - * Revision 1.10 1997/09/18 08:59:21 adam - * Extra generic handle for the character mapping routines. - * - * Revision 1.9 1997/09/17 12:19:21 adam - * Zebra version corresponds to YAZ version 1.4. - * Changed Zebra server so that it doesn't depend on global common_resource. - * - * Revision 1.8 1997/09/09 13:38:14 adam - * Partial port to WIN95/NT. - * - * Revision 1.7 1997/09/05 15:30:10 adam - * Changed prototype for chr_map_input - added const. - * Added support for C++, headers uses extern "C" for public definitions. - * - * Revision 1.6 1997/09/04 13:54:40 adam - * Added MARC filter - type grs.marc. where syntax refers - * to abstract syntax. New method tellf in retrieve/extract method. - * - * Revision 1.5 1997/07/15 16:29:03 adam - * Initialized dummy variable to keep checker gcc happy. - * - * Revision 1.4 1997/04/30 08:56:08 quinn - * null - * - * Revision 1.2 1996/10/11 16:06:43 quinn - * Revision 1.3 1997/02/24 10:41:50 adam - * Cleanup of code and commented out the "end element-end-record" code. - * - * Revision 1.2 1996/10/11 16:06:43 quinn - * Fixed arguments to nodetogr - * - * Revision 1.1 1996/10/11 10:57:25 adam - * New module recctrl. Used to manage records (extract/retrieval). - * - * Revision 1.29 1996/10/08 10:30:21 quinn - * Fixed type mismatch - * - * Revision 1.28 1996/10/07 16:06:40 quinn - * Added SOIF support - * - * Revision 1.27 1996/06/11 10:54:12 quinn - * Relevance work - * - * Revision 1.26 1996/06/06 12:08:45 quinn - * Added showRecord function - * - * Revision 1.25 1996/06/04 14:18:53 quinn - * Charmap work - * - * Revision 1.24 1996/06/04 13:27:54 quinn - * More work on charmapping - * - * Revision 1.23 1996/06/04 10:19:01 adam - * Minor changes - removed include of ctype.h. - * - * Revision 1.22 1996/06/03 10:15:27 quinn - * Various character-mapping. - * - * Revision 1.21 1996/05/31 13:27:24 quinn - * Character-conversion in phrases, too. - * - * Revision 1.19 1996/05/16 15:31:14 quinn - * a7 - * - * Revision 1.18 1996/05/09 07:28:56 quinn - * Work towards phrases and multiple registers - * - * Revision 1.17 1996/05/01 13:46:37 adam - * First work on multiple records in one file. - * New option, -offset, to the "unread" command in the filter module. - * - * Revision 1.16 1996/01/17 14:57:54 adam - * Prototype changed for reader functions in extract/retrieve. File - * is identified by 'void *' instead of 'int. - * - * Revision 1.15 1996/01/08 19:15:47 adam - * New input filter that works! - * - * Revision 1.14 1995/12/15 12:36:11 adam - * Retrieval calls data1_read_regx when subType is specified. - * - * Revision 1.13 1995/12/15 12:24:43 quinn - * *** empty log message *** - * - * Revision 1.12 1995/12/15 12:20:28 quinn - * *** empty log message *** - * - * Revision 1.11 1995/12/15 12:07:57 quinn - * Changed extraction strategy. - * - * Revision 1.10 1995/12/14 11:10:48 quinn - * Explain work - * - * Revision 1.9 1995/12/13 17:14:05 quinn - * *** empty log message *** - * - * Revision 1.8 1995/12/13 15:33:18 quinn - * *** empty log message *** - * - * Revision 1.7 1995/12/13 13:45:39 quinn - * Changed data1 to use nmem. - * - * Revision 1.6 1995/12/04 14:22:30 adam - * Extra arg to recType_byName. - * Started work on new regular expression parsed input to - * structured records. - * - * Revision 1.5 1995/11/28 14:18:37 quinn - * Set output_format. - * - * Revision 1.4 1995/11/21 13:14:49 quinn - * Fixed end-of-data-field problem (maybe). - * - * Revision 1.3 1995/11/15 19:13:09 adam - * Work on record management. - * - */ +/* $Id: recgrs.c,v 1.59 2002-08-02 19:26:56 adam Exp $ + Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002 + Index Data Aps + +This file is part of the Zebra server. + +Zebra is free software; you can redistribute it and/or modify it under +the terms of the GNU General Public License as published by the Free +Software Foundation; either version 2, or (at your option) any later +version. + +Zebra is distributed in the hope that it will be useful, but WITHOUT ANY +WARRANTY; without even the implied warranty of MERCHANTABILITY or +FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License +for more details. + +You should have received a copy of the GNU General Public License +along with Zebra; see the file LICENSE.zebra. If not, write to the +Free Software Foundation, 59 Temple Place - Suite 330, Boston, MA +02111-1307, USA. +*/ + + #include #include @@ -293,7 +82,7 @@ static int read_grs_type (struct grs_handlers *h, static void grs_add_handler (struct grs_handlers *h, RecTypeGrs t) { - struct grs_handler *gh = (struct grs_handler *) malloc (sizeof(*gh)); + struct grs_handler *gh = (struct grs_handler *) xmalloc (sizeof(*gh)); gh->next = h->handlers; h->handlers = gh; gh->initFlag = 0; @@ -303,7 +92,7 @@ static void grs_add_handler (struct grs_handlers *h, RecTypeGrs t) static void *grs_init(RecType recType) { - struct grs_handlers *h = (struct grs_handlers *) malloc (sizeof(*h)); + struct grs_handlers *h = (struct grs_handlers *) xmalloc (sizeof(*h)); h->handlers = 0; grs_add_handler (h, recTypeGrs_sgml); @@ -312,6 +101,9 @@ static void *grs_init(RecType recType) grs_add_handler (h, recTypeGrs_tcl); #endif grs_add_handler (h, recTypeGrs_marc); +#if YAZ_HAVE_EXPAT + grs_add_handler (h, recTypeGrs_xml); +#endif return h; } @@ -324,12 +116,22 @@ static void grs_destroy(void *clientData) gh_next = gh->next; if (gh->initFlag) (*gh->type->destroy)(gh->clientData); - free (gh); + xfree (gh); gh = gh_next; } - free (h); + xfree (h); } +/* use + 1 start element (tag) + 2 end element + 3 start attr (and attr-exact) + 4 end attr + + 1016 cdata + 1015 attr data +*/ + static void index_xpath (data1_node *n, struct recExtractCtrl *p, int level, RecWord *wrd, int use) { @@ -361,7 +163,7 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p, case DATA1N_tag: for (nn = n; nn; nn = nn->parent) { - if (n->which == DATA1N_tag) + if (nn->which == DATA1N_tag) { size_t tlen = strlen(nn->u.tag.tag); if (tlen + flen > (sizeof(tag_path_full)-2)) @@ -370,21 +172,13 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p, flen += tlen; tag_path_full[flen++] = '/'; } - else if (n->which == DATA1N_root) - { - size_t tlen = strlen(nn->u.root.type); - if (tlen + flen > (sizeof(tag_path_full)-2)) - return; - memcpy (tag_path_full + flen, nn->u.root.type, tlen); - flen += tlen; - tag_path_full[flen++] = '/'; + else if (nn->which == DATA1N_root) break; - } } wrd->reg_type = '0'; wrd->string = tag_path_full; wrd->length = flen; - wrd->attrSet = VAL_IDXPATH, + wrd->attrSet = VAL_IDXPATH; wrd->attrUse = use; if (p->flagShowRecords) { @@ -397,9 +191,67 @@ static void index_xpath (data1_node *n, struct recExtractCtrl *p, } else { - (*p->tokenAdd)(wrd); + data1_xattr *xp; + (*p->tokenAdd)(wrd); /* index element pag (AKA tag path) */ + if (use == 1) + { + for (xp = n->u.tag.attributes; xp; xp = xp->next) + { + char comb[512]; + /* attribute (no value) */ + wrd->reg_type = '0'; + wrd->attrUse = 3; + wrd->string = xp->name; + wrd->length = strlen(xp->name); + + wrd->seqno--; + (*p->tokenAdd)(wrd); + + if (xp->value && + strlen(xp->name) + strlen(xp->value) < sizeof(comb)-2) + { + /* attribute value exact */ + strcpy (comb, xp->name); + strcat (comb, "="); + strcat (comb, xp->value); + + wrd->attrUse = 3; + wrd->reg_type = '0'; + wrd->string = comb; + wrd->length = strlen(comb); + wrd->seqno--; + + (*p->tokenAdd)(wrd); + } + } + for (xp = n->u.tag.attributes; xp; xp = xp->next) + { + char attr_tag_path_full[1024]; + + sprintf (attr_tag_path_full, "@%s/%.*s", + xp->name, flen, tag_path_full); + + wrd->reg_type = '0'; + wrd->attrUse = 1; + wrd->string = attr_tag_path_full; + wrd->length = strlen(attr_tag_path_full); + (*p->tokenAdd)(wrd); + + wrd->attrUse = 1015; + wrd->reg_type = 'w'; + wrd->string = xp->value; + wrd->length = strlen(xp->value); + + (*p->tokenAdd)(wrd); + + wrd->reg_type = '0'; + wrd->attrUse = 2; + wrd->string = attr_tag_path_full; + wrd->length = strlen(attr_tag_path_full); + (*p->tokenAdd)(wrd); + } + } } - break; } } @@ -591,6 +443,7 @@ int grs_extract_tree(struct recExtractCtrl *p, data1_node *n) (*p->schemaAdd)(p, oidtmp); } (*p->init)(p, &wrd); + return dumpkeys(n, p, 0, &wrd); } @@ -628,9 +481,14 @@ static int grs_extract_sub(struct grs_handlers *h, struct recExtractCtrl *p, if ((oid_ent_to_oid (&oe, oidtmp))) (*p->schemaAdd)(p, oidtmp); } + + /* ensure our data1 tree is UTF-8 */ + data1_iconv (p->dh, mem, n, "UTF-8", data1_get_encoding(p->dh, n)); + #if 0 data1_pr_tree (p->dh, n, stdout); #endif + (*p->init)(p, &wrd); if (dumpkeys(n, p, 0, &wrd) < 0) { @@ -727,9 +585,33 @@ static int process_comp(data1_handle dh, data1_node *n, Z_RecordComposition *c) } } +static void add_idzebra_info (struct recRetrieveCtrl *p, data1_node *top, + NMEM mem) +{ + const char *idzebra_ns[7]; + + idzebra_ns[0] = "xmlns:idzebra"; + idzebra_ns[1] = "http://www.indexdata.dk/zebra/"; + idzebra_ns[2] = 0; + + data1_tag_add_attr (p->dh, mem, top, idzebra_ns); + + data1_mk_tag_data_int (p->dh, top, "idzebra:size", p->recordSize, + mem); + if (p->score != -1) + data1_mk_tag_data_int (p->dh, top, "idzebra:score", + p->score, mem); + + data1_mk_tag_data_int (p->dh, top, "idzebra:localnumber", p->localno, + mem); + if (p->fname) + data1_mk_tag_data_text(p->dh, top, "idzebra:filename", + p->fname, mem); +} + static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) { - data1_node *node = 0, *onode = 0; + data1_node *node = 0, *onode = 0, *top; data1_node *dnew; data1_maptab *map; int res, selected = 0; @@ -738,6 +620,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) char *tagname; struct grs_handlers *h = (struct grs_handlers *) clientData; int requested_schema = VAL_NONE; + data1_marctab *marctab; + int dummy; mem = nmem_create(); gri.readf = p->readf; @@ -762,12 +646,16 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) nmem_destroy (mem); return 0; } + /* ensure our data1 tree is UTF-8 */ + data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node)); + #if 0 data1_pr_tree (p->dh, node, stdout); #endif + top = data1_get_root_tag (p->dh, node); + logf (LOG_DEBUG, "grs_retrieve: size"); - if ((dnew = data1_insert_taggeddata(p->dh, node, node, - "size", mem))) + if ((dnew = data1_mk_tag_data_wd(p->dh, top, "size", mem))) { dnew->u.data.what = DATA1I_text; dnew->u.data.data = dnew->lbuf; @@ -777,7 +665,7 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) tagname = res_get_def(p->res, "tagrank", "rank"); if (strcmp(tagname, "0") && p->score >= 0 && - (dnew = data1_insert_taggeddata(p->dh, node, node, tagname, mem))) + (dnew = data1_mk_tag_data_wd(p->dh, top, tagname, mem))) { logf (LOG_DEBUG, "grs_retrieve: %s", tagname); dnew->u.data.what = DATA1I_num; @@ -788,15 +676,18 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) tagname = res_get_def(p->res, "tagsysno", "localControlNumber"); if (strcmp(tagname, "0") && p->localno > 0 && - (dnew = data1_insert_taggeddata(p->dh, node, node, tagname, mem))) + (dnew = data1_mk_tag_data_wd(p->dh, top, tagname, mem))) { logf (LOG_DEBUG, "grs_retrieve: %s", tagname); dnew->u.data.what = DATA1I_text; dnew->u.data.data = dnew->lbuf; + sprintf(dnew->u.data.data, "%d", p->localno); dnew->u.data.len = strlen(dnew->u.data.data); } - +#if 0 + data1_pr_tree (p->dh, node, stdout); +#endif if (p->comp && p->comp->which == Z_RecordComp_complex && p->comp->u.complex->generic && p->comp->u.complex->generic->schema) @@ -839,7 +730,7 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) * the overlap of schema and formatting which is inherent in the MARC * family) */ - logf (LOG_DEBUG, "grs_retrieve: syntax mapping"); + yaz_log (LOG_DEBUG, "grs_retrieve: syntax mapping"); if (node->u.root.absyn) for (map = node->u.root.absyn->maptabs; map; map = map->next) { @@ -855,7 +746,7 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) break; } } - logf (LOG_DEBUG, "grs_retrieve: schemaIdentifier"); + yaz_log (LOG_DEBUG, "grs_retrieve: schemaIdentifier"); if (node->u.root.absyn && node->u.root.absyn->reference != VAL_NONE && p->input_format == VAL_GRS1) @@ -884,8 +775,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) } *(p++) = '\0'; - if ((dnew = data1_insert_taggeddata(dh, node, node, - "schemaIdentifier", mem))) + if ((dnew = data1_mk_tag_data_wd(dh, node, + "schemaIdentifier", mem))) { dnew->u.data.what = DATA1I_oid; dnew->u.data.data = (char *) nmem_malloc(mem, p - tmp); @@ -915,10 +806,12 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) switch (p->output_format = (p->input_format != VAL_NONE ? p->input_format : VAL_SUTRS)) { - data1_marctab *marctab; - int dummy; - case VAL_TEXT_XML: + add_idzebra_info (p, top, mem); + + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, "UTF-8"); + if (!(p->rec_buf = data1_nodetoidsgml(p->dh, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -952,6 +845,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->rec_len = (size_t) (-1); break; case VAL_SUTRS: + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, "UTF-8"); if (!(p->rec_buf = data1_nodetobuf(p->dh, node, selected, &p->rec_len))) p->diagnostic = 238; @@ -988,6 +883,8 @@ static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p) p->diagnostic = 238; break; } + if (p->encoding) + data1_iconv (p->dh, mem, node, p->encoding, "UTF-8"); if (!(p->rec_buf = data1_nodetomarc(p->dh, marctab, node, selected, &p->rec_len))) p->diagnostic = 238;