-/*
- * Copyright (C) 1994-2002, Index Data
- * All rights reserved.
- *
- * $Id: recgrs.c,v 1.51 2002-05-28 21:10:34 adam Exp $
- */
+/* $Id: recgrs.c,v 1.71 2002-12-16 20:27:18 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 <stdio.h>
#include <assert.h>
grs_add_handler (h, recTypeGrs_tcl);
#endif
grs_add_handler (h, recTypeGrs_marc);
-#if YAZ_HAVE_EXPAT
+#if HAVE_EXPAT_H
grs_add_handler (h, recTypeGrs_xml);
#endif
+#if HAVE_PERL
+ grs_add_handler (h, recTypeGrs_perl);
+#endif
return h;
}
xfree (h);
}
+/* *ostrich*
+
+ New function, looking for xpath "element" definitions in abs, by
+ tagpath, using a kind of ugly regxp search.The DFA was built while
+ parsing abs, so here we just go trough them and try to match
+ against the given tagpath. The first matching entry is returned.
+
+ pop, 2002-12-13
+ */
+
+data1_termlist *xpath_termlist_by_tagpath(char *tagpath, data1_node *n)
+{
+ data1_absyn *abs = n->root->u.root.absyn;
+ data1_xpelement *xpe = abs->xp_elements;
+ char *pexpr = malloc(strlen(tagpath)+2);
+ int ok = 0;
+
+ sprintf (pexpr, "%s\n", tagpath);
+ while (xpe)
+ {
+ struct DFA_state **dfaar = xpe->dfa->states;
+ struct DFA_state *s=dfaar[0];
+ struct DFA_tran *t;
+ const char *p;
+ int i;
+ unsigned char c;
+ int start_line = 1;
+
+ c = *pexpr++; t = s->trans; i = s->tran_no;
+ if (c >= t->ch[0] && c <= t->ch[1]) {
+ p = pexpr;
+ do {
+ if ((s = dfaar[t->to])->rule_no &&
+ (start_line || s->rule_nno)) {
+ ok = 1;
+ break;
+ }
+ for (t=s->trans, i=s->tran_no; --i >= 0; t++) {
+ if ((unsigned) *p >= t->ch[0] && (unsigned) *p <= t->ch[1])
+ break;
+ }
+ p++;
+ } while (i >= 0);
+ }
+ pexpr--;
+ if (ok) break;
+ xpe = xpe->next;
+ }
+
+ if (ok) {
+ return xpe->termlists;
+ } else {
+ return NULL;
+ }
+}
+
+/* use
+ 1 start element (tag)
+ 2 end element
+ 3 start attr (and attr-exact)
+ 4 end attr
+
+ 1016 cdata
+ 1015 attr data
+
+ *ostrich*
+
+ Now, if there is a matching xelm described in abs, for the
+ indexed element or the attribute, then the data is handled according
+ to those definitions...
+
+ modified by pop, 2002-12-13
+*/
+
static void index_xpath (data1_node *n, struct recExtractCtrl *p,
int level, RecWord *wrd, int use)
{
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);
+ }
+ else {
+ data1_termlist *tl;
+ int xpdone = 0;
+ flen = 0;
+
+ /* we have to fetch the whole path to the data 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;
+ }
+
+ tag_path_full[flen] = 0;
+
+ /* If we have a matching termlist... */
+ if (tl = xpath_termlist_by_tagpath(tag_path_full, n)) {
+ for (; tl; tl = tl->next) {
+ wrd->reg_type = *tl->structure;
+ /* this is the ! case, so structure is for the xpath index */
+ if (!tl->att) {
+ wrd->attrSet = VAL_IDXPATH;
+ wrd->attrUse = use;
+ (*p->tokenAdd)(wrd);
+ xpdone = 1;
+ /* this is just the old fashioned attribute based index */
+ } else {
+ wrd->attrSet = (int) (tl->att->parent->reference);
+ wrd->attrUse = tl->att->locals->local;
+ (*p->tokenAdd)(wrd);
+ }
+ }
+ }
+ /* xpath indexing is done, if there was no termlist given,
+ or no ! attribute... */
+ if (!xpdone) {
+ wrd->attrSet = VAL_IDXPATH;
+ wrd->attrUse = use;
+ wrd->reg_type = 'w';
+ (*p->tokenAdd)(wrd);
+ }
}
break;
case DATA1N_tag:
+ flen = 0;
for (nn = n; nn; nn = nn->parent)
{
if (nn->which == DATA1N_tag)
tag_path_full[flen++] = '/';
}
else if (nn->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++] = '/';
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)
{
}
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];
+ int int_len = flen;
+
+ sprintf (attr_tag_path_full, "@%s/%.*s",
+ xp->name, int_len, 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);
+
+ if (xp->value)
+ {
+ /* the same jokes, as with the data nodes ... */
+ data1_termlist *tl;
+ int xpdone = 0;
+
+ wrd->string = xp->value;
+ wrd->length = strlen(xp->value);
+ wrd->reg_type = 'w';
+
+ if (tl = xpath_termlist_by_tagpath(attr_tag_path_full,
+ n)) {
+ for (; tl; tl = tl->next) {
+ wrd->reg_type = *tl->structure;
+ if (!tl->att) {
+ wrd->attrSet = VAL_IDXPATH;
+ wrd->attrUse = 1015;
+ (*p->tokenAdd)(wrd);
+ xpdone = 1;
+ } else {
+ wrd->attrSet = (int) (tl->att->parent->reference);
+ wrd->attrUse = tl->att->locals->local;
+ (*p->tokenAdd)(wrd);
+ }
+ }
+
+ }
+ if (!xpdone) {
+ wrd->attrSet = VAL_IDXPATH;
+ wrd->attrUse = 1015;
+ wrd->reg_type = 'w';
+ (*p->tokenAdd)(wrd);
+ }
+ }
+
+ wrd->attrSet = VAL_IDXPATH;
+ wrd->reg_type = '0';
+ wrd->attrUse = 2;
+ wrd->string = attr_tag_path_full;
+ wrd->length = strlen(attr_tag_path_full);
+ (*p->tokenAdd)(wrd);
+ }
+ }
}
- break;
}
}
{
data1_termlist *tlist = 0;
data1_datatype dtype = DATA1K_string;
+
/*
* cycle up towards the root until we find a tag with an att..
* this has the effect of indexing locally defined tags with
*/
while (!par->u.tag.element)
- if (!par->parent || !(par=get_parent_tag(p->dh, par->parent)))
- break;
+ if (!par->parent || !(par=get_parent_tag(p->dh, par->parent)))
+ break;
if (!par || !(tlist = par->u.tag.element->termlists))
- return;
+ return;
if (par->u.tag.element->tag)
- dtype = par->u.tag.element->tag->kind;
+ dtype = par->u.tag.element->tag->kind;
for (; tlist; tlist = tlist->next)
{
+
char xattr[512];
/* consider source */
wrd->string = 0;
-
+
if (!strcmp (tlist->source, "data") && n->which == DATA1N_data)
{
wrd->string = n->u.data.data;
{
index_termlist (n, n, p, level, wrd);
/* index start tag */
+ assert (n->root->u.root.absyn);
+
if (!n->root->u.root.absyn)
index_xpath (n, p, level, wrd, 1);
+ else if (n->root->u.root.absyn->enable_xpath_indexing)
+ index_xpath (n, p, level, wrd, 1);
}
if (n->child)
{
printf("%*s", level * 4, "");
printf("Data: ");
- if (n->u.data.len > 32)
- printf("'%.24s ... %.6s'\n", n->u.data.data,
+ if (n->u.data.len > 256)
+ printf("'%.240s ... %.6s'\n", n->u.data.data,
n->u.data.data + n->u.data.len-6);
else if (n->u.data.len > 0)
printf("'%.*s'\n", n->u.data.len, n->u.data.data);
index_termlist (par, n, p, level, wrd);
if (!n->root->u.root.absyn)
index_xpath (n, p, level, wrd, 1016);
-
+ else if (n->root->u.root.absyn->enable_xpath_indexing)
+ 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);
+ else if (n->root->u.root.absyn->enable_xpath_indexing)
+ index_xpath (n, p, level, wrd, 2);
}
-
if (p->flagShowRecords && n->which == DATA1N_root)
{
printf("%*s-------------\n\n", level * 4, "");
(*p->schemaAdd)(p, oidtmp);
}
(*p->init)(p, &wrd);
+
return dumpkeys(n, p, 0, &wrd);
}
gri.dh = p->dh;
if (read_grs_type (h, &gri, p->subType, &n))
- return RECCTRL_EXTRACT_ERROR;
+ return RECCTRL_EXTRACT_ERROR_NO_SUCH_FILTER;
if (!n)
return RECCTRL_EXTRACT_EOF;
oe.proto = PROTO_Z3950;
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)
{
data1_free_tree(p->dh, n);
- return RECCTRL_EXTRACT_ERROR;
+ return RECCTRL_EXTRACT_ERROR_GENERIC;
}
data1_free_tree(p->dh, n);
return RECCTRL_EXTRACT_OK;
}
}
+/* Add Zebra info in separate namespace ...
+ <root
+ ...
+ <metadata xmlns="http://www.indexdata.dk/zebra/">
+ <size>359</size>
+ <localnumber>447</localnumber>
+ <filename>records/genera.xml</filename>
+ </metadata>
+ </root>
+*/
+
+static void zebra_xml_metadata (struct recRetrieveCtrl *p, data1_node *top,
+ NMEM mem)
+{
+ const char *idzebra_ns[3];
+ const char *i2 = "\n ";
+ const char *i4 = "\n ";
+ data1_node *n;
+
+ idzebra_ns[0] = "xmlns";
+ idzebra_ns[1] = "http://www.indexdata.dk/zebra/";
+ idzebra_ns[2] = 0;
+
+ data1_mk_text (p->dh, mem, i2, top);
+
+ n = data1_mk_tag (p->dh, mem, "idzebra", idzebra_ns, top);
+
+ data1_mk_text (p->dh, mem, "\n", top);
+
+ data1_mk_text (p->dh, mem, i4, n);
+
+ data1_mk_tag_data_int (p->dh, n, "size", p->recordSize, mem);
+
+ if (p->score != -1)
+ {
+ data1_mk_text (p->dh, mem, i4, n);
+ data1_mk_tag_data_int (p->dh, n, "score", p->score, mem);
+ }
+ data1_mk_text (p->dh, mem, i4, n);
+ data1_mk_tag_data_int (p->dh, n, "localnumber", p->localno, mem);
+ if (p->fname)
+ {
+ data1_mk_text (p->dh, mem, i4, n);
+ data1_mk_tag_data_text(p->dh, n, "filename", p->fname, mem);
+ }
+ data1_mk_text (p->dh, mem, i2, n);
+}
+
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;
NMEM mem;
struct grs_read_info gri;
- char *tagname;
+ const 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_mk_tag_data_wd(p->dh, node, "size", mem)))
+ tagname = data1_systag_lookup(node->u.root.absyn, "size", "size");
+ if (tagname &&
+ (dnew = data1_mk_tag_data_wd(p->dh, top, tagname, mem)))
{
dnew->u.data.what = DATA1I_text;
dnew->u.data.data = dnew->lbuf;
sprintf(dnew->u.data.data, "%d", p->recordSize);
dnew->u.data.len = strlen(dnew->u.data.data);
}
-
- tagname = res_get_def(p->res, "tagrank", "rank");
- if (strcmp(tagname, "0") && p->score >= 0 &&
- (dnew = data1_mk_tag_data_wd(p->dh, node, tagname, mem)))
+
+ tagname = data1_systag_lookup(node->u.root.absyn, "rank", "rank");
+ if (tagname && p->score >= 0 &&
+ (dnew = data1_mk_tag_data_wd(p->dh, top, tagname, mem)))
{
logf (LOG_DEBUG, "grs_retrieve: %s", tagname);
dnew->u.data.what = DATA1I_num;
dnew->u.data.len = strlen(dnew->u.data.data);
}
- tagname = res_get_def(p->res, "tagsysno", "localControlNumber");
- if (strcmp(tagname, "0") && p->localno > 0 &&
- (dnew = data1_mk_tag_data_wd(p->dh, node, tagname, mem)))
+ tagname = data1_systag_lookup(node->u.root.absyn, "sysno",
+ "localControlNumber");
+ if (tagname && p->localno > 0 &&
+ (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);
}
sprintf(p, "%d", *ii);
p += strlen(p);
}
- *(p++) = '\0';
-
- if ((dnew = data1_mk_tag_data_wd(dh, node,
+ if ((dnew = data1_mk_tag_data_wd(dh, top,
"schemaIdentifier", mem)))
{
dnew->u.data.what = DATA1I_oid;
switch (p->output_format = (p->input_format != VAL_NONE ?
p->input_format : VAL_SUTRS))
{
- data1_marctab *marctab;
- int dummy;
-
case VAL_TEXT_XML:
+ zebra_xml_metadata (p, top, mem);
+
+#if 0
+ data1_pr_tree (p->dh, node, stdout);
+#endif
+
+ 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;