/*
- * Copyright (C) 1994-2000, Index Data
+ * Copyright (C) 1994-2002, Index Data
* All rights reserved.
- * Sebastian Hammer, Adam Dickmeiss
- *
- * $Log: recgrs.c,v $
- * 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.<syntax> 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 1.57 2002-08-01 09:37:44 adam Exp $
*/
#include <stdio.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;
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);
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;
}
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)
+{
+ int i;
+ char tag_path_full[1024];
+ size_t flen = 0;
+ data1_node *nn;
+
+ switch (n->which)
+ {
+ case DATA1N_data:
+ wrd->reg_type = 'w';
+ wrd->string = n->u.data.data;
+ wrd->length = n->u.data.len;
+ wrd->attrSet = VAL_IDXPATH,
+ wrd->attrUse = use;
+ if (p->flagShowRecords)
+ {
+ printf("%*s data=", (level + 1) * 4, "");
+ for (i = 0; i<wrd->length && i < 8; i++)
+ fputc (wrd->string[i], stdout);
+ printf("\n");
+ }
+ else
+ {
+ (*p->tokenAdd)(wrd);
+ }
+ break;
+ case DATA1N_tag:
+ for (nn = n; nn; nn = nn->parent)
+ {
+ if (nn->which == DATA1N_tag)
+ {
+ size_t tlen = strlen(nn->u.tag.tag);
+ if (tlen + flen > (sizeof(tag_path_full)-2))
+ return;
+ memcpy (tag_path_full + flen, nn->u.tag.tag, 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->attrUse = use;
+ if (p->flagShowRecords)
+ {
+ printf("%*s tag=", (level + 1) * 4, "");
+ for (i = 0; i<wrd->length && i < 40; i++)
+ fputc (wrd->string[i], stdout);
+ if (i == 40)
+ printf (" ..");
+ printf("\n");
+ }
+ else
+ {
+ data1_xattr *xp;
+ (*p->tokenAdd)(wrd);
+
+#if 0
+ for (xp = n->u.tag.attributes; xp; xp = xp->next)
+ {
+ if (use == 1)
+ { /* attribute (no value) */
+ wrd->reg_type = '0';
+ wrd->attrUse = 3;
+ wrd->string = xp->name;
+ wrd->length = strlen(xp->name);
+
+ wrd->seqno--;
+ (*p->tokenAdd)(wrd);
+ }
+ }
+#else
+ for (xp = n->u.tag.attributes; xp; xp = xp->next)
+ {
+ char comb[512];
+
+ if (use == 1)
+ { /* attribute start */
+ wrd->reg_type = '0';
+ wrd->attrUse = 3;
+ wrd->string = xp->name;
+ wrd->length = strlen(xp->name);
+
+ wrd->seqno--;
+ (*p->tokenAdd)(wrd);
+ }
+
+ if (use == 1 && 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);
+
+ /* attribute value phrase */
+
+ wrd->attrUse = 1015;
+ wrd->reg_type = 'w';
+ wrd->string = xp->value;
+ wrd->length = strlen(xp->value);
+
+ (*p->tokenAdd)(wrd);
+ }
+ if (use == 2)
+ {
+ wrd->reg_type = '0';
+ wrd->attrUse = 4;
+ wrd->string = xp->name;
+ wrd->length = strlen(xp->name);
+
+ (*p->tokenAdd)(wrd);
+ }
+ }
+#endif
+ }
+ break;
+ }
}
-static void index_tag (data1_node *par, data1_node *n,
- struct recExtractCtrl *p, int level, RecWord *wrd)
+static void index_termlist (data1_node *par, data1_node *n,
+ struct recExtractCtrl *p, int level, RecWord *wrd)
{
data1_termlist *tlist = 0;
data1_datatype dtype = DATA1K_string;
wrd->string = n->u.data.data;
wrd->length = n->u.data.len;
}
+ else if (!strcmp (tlist->source, "tag") && n->which == DATA1N_tag)
+ {
+ wrd->string = n->u.tag.tag;
+ wrd->length = strlen(n->u.tag.tag);
+ }
else if (sscanf (tlist->source, "attr(%511[^)])", xattr) == 1 &&
n->which == DATA1N_tag)
{
}
}
-static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level)
+static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level,
+ RecWord *wrd)
{
- RecWord wrd;
- (*p->init)(p, &wrd); /* set defaults */
for (; n; n = n->next)
{
if (p->flagShowRecords) /* display element description to user */
if (n->which == DATA1N_root)
{
printf("%*s", level * 4, "");
- printf("Record type: '%s'\n", n->u.root.absyn->name);
+ printf("Record type: '%s'\n", n->u.root.type);
}
else if (n->which == DATA1N_tag)
{
}
}
- if (n->child)
- if (dumpkeys(n->child, p, level + 1) < 0)
- return -1;
-
if (n->which == DATA1N_tag)
{
- index_tag (n, n, p, level, &wrd);
+ index_termlist (n, n, p, level, wrd);
+ /* index start tag */
+ if (!n->root->u.root.absyn)
+ index_xpath (n, p, level, wrd, 1);
}
+ if (n->child)
+ if (dumpkeys(n->child, p, level + 1, wrd) < 0)
+ return -1;
+
+
if (n->which == DATA1N_data)
{
data1_node *par = get_parent_tag(p->dh, n);
}
if (par)
- index_tag (par, n, p, level, &wrd);
+ index_termlist (par, n, p, level, wrd);
+ if (!n->root->u.root.absyn)
+ index_xpath (n, p, level, wrd, 1016);
+
}
+
+ if (n->which == DATA1N_tag)
+ {
+ /* index end tag */
+ if (!n->root->u.root.absyn)
+ index_xpath (n, p, level, wrd, 2);
+ }
+
+
if (p->flagShowRecords && n->which == DATA1N_root)
{
printf("%*s-------------\n\n", level * 4, "");
{
oident oe;
int oidtmp[OID_SIZE];
+ RecWord wrd;
oe.proto = PROTO_Z3950;
oe.oclass = CLASS_SCHEMA;
- oe.value = n->u.root.absyn->reference;
-
- if ((oid_ent_to_oid (&oe, oidtmp)))
- (*p->schemaAdd)(p, oidtmp);
+ if (n->u.root.absyn)
+ {
+ oe.value = n->u.root.absyn->reference;
+
+ if ((oid_ent_to_oid (&oe, oidtmp)))
+ (*p->schemaAdd)(p, oidtmp);
+ }
+ (*p->init)(p, &wrd);
- return dumpkeys(n, p, 0);
+ return dumpkeys(n, p, 0, &wrd);
}
static int grs_extract_sub(struct grs_handlers *h, struct recExtractCtrl *p,
struct grs_read_info gri;
oident oe;
int oidtmp[OID_SIZE];
+ RecWord wrd;
gri.readf = p->readf;
gri.seekf = p->seekf;
return RECCTRL_EXTRACT_EOF;
oe.proto = PROTO_Z3950;
oe.oclass = CLASS_SCHEMA;
- oe.value = n->u.root.absyn->reference;
- if ((oid_ent_to_oid (&oe, oidtmp)))
- (*p->schemaAdd)(p, oidtmp);
+#if 0
+ if (!n->u.root.absyn)
+ return RECCTRL_EXTRACT_ERROR;
+#endif
+ if (n->u.root.absyn)
+ {
+ oe.value = n->u.root.absyn->reference;
+ 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 (dumpkeys(n, p, 0) < 0)
+#if 0
+ data1_pr_tree (p->dh, n, stdout);
+#endif
+
+ (*p->init)(p, &wrd);
+ if (dumpkeys(n, p, 0, &wrd) < 0)
{
data1_free_tree(p->dh, n);
return RECCTRL_EXTRACT_ERROR;
}
}
+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;
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;
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;
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;
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)
* the overlap of schema and formatting which is inherent in the MARC
* family)
*/
- logf (LOG_DEBUG, "grs_retrieve: syntax mapping");
- for (map = node->u.root.absyn->maptabs; map; map = map->next)
- {
- if (map->target_absyn_ref == p->input_format)
- {
- onode = node;
- if (!(node = data1_map_record(p->dh, onode, map, mem)))
- {
- p->diagnostic = 14;
- nmem_destroy (mem);
- return 0;
- }
- break;
- }
- }
- logf (LOG_DEBUG, "grs_retrieve: schemaIdentifier");
+ yaz_log (LOG_DEBUG, "grs_retrieve: syntax mapping");
+ if (node->u.root.absyn)
+ for (map = node->u.root.absyn->maptabs; map; map = map->next)
+ {
+ if (map->target_absyn_ref == p->input_format)
+ {
+ onode = node;
+ if (!(node = data1_map_record(p->dh, onode, map, mem)))
+ {
+ p->diagnostic = 14;
+ nmem_destroy (mem);
+ return 0;
+ }
+ break;
+ }
+ }
+ yaz_log (LOG_DEBUG, "grs_retrieve: schemaIdentifier");
if (node->u.root.absyn &&
node->u.root.absyn->reference != VAL_NONE &&
p->input_format == VAL_GRS1)
}
*(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);
else if (p->comp && !res)
selected = 1;
+#if 0
+ data1_pr_tree (p->dh, node, stdout);
+#endif
logf (LOG_DEBUG, "grs_retrieve: transfer syntax mapping");
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;
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;
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;