-/* $Id: recgrs.c,v 1.94 2004-11-29 21:45:12 adam Exp $
- Copyright (C) 1995,1996,1997,1998,1999,2000,2001,2002,2003,2004
- Index Data Aps
+/* $Id: recgrs.c,v 1.100 2005-03-05 09:19:15 adam Exp $
+ Copyright (C) 1995-2005
+ Index Data ApS
This file is part of the Zebra server.
#include <stdio.h>
#include <assert.h>
#include <sys/types.h>
-#ifndef WIN32
-#include <unistd.h>
-#include <dlfcn.h>
-#endif
+#include <ctype.h>
-#include <yaz/ylog.h>
+#include <yaz/log.h>
#include <yaz/oid.h>
#include <d1_absyn.h>
const char *tok;
const char *src;
int lookahead;
+ NMEM nmem;
};
static int sp_lex(struct source_parser *sp)
return sp->lookahead;
}
+static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd);
+
+static int sp_range(struct source_parser *sp, data1_node *n, RecWord *wrd)
+{
+ int start, len;
+ RecWord tmp_w;
+
+ /* ( */
+ sp_lex(sp);
+ if (sp->lookahead != '(')
+ return 0;
+ sp_lex(sp); /* skip ( */
+
+ /* 1st arg: string */
+ if (!sp_expr(sp, n, wrd))
+ return 0;
+
+ if (sp->lookahead != ',')
+ return 0;
+ sp_lex(sp); /* skip , */
+
+ /* 2nd arg: start */
+ if (!sp_expr(sp, n, &tmp_w))
+ return 0;
+ start = atoi_n(tmp_w.term_buf, tmp_w.term_len);
+
+ if (sp->lookahead == ',')
+ {
+ sp_lex(sp); /* skip , */
+
+ /* 3rd arg: length */
+ if (!sp_expr(sp, n, &tmp_w))
+ return 0;
+ len = atoi_n(tmp_w.term_buf, tmp_w.term_len);
+ }
+ else
+ len = wrd->term_len;
+
+ /* ) */
+ if (sp->lookahead != ')')
+ return 0;
+ sp_lex(sp);
+
+ if (wrd->term_buf && wrd->term_len)
+ {
+ wrd->term_buf += start;
+ wrd->term_len -= start;
+ if (wrd->term_len > len)
+ wrd->term_len = len;
+ }
+ return 1;
+}
+
+static int sp_first(struct source_parser *sp, data1_node *n, RecWord *wrd)
+{
+ char num_str[20];
+ int min_pos = -1;
+ sp_lex(sp);
+ if (sp->lookahead != '(')
+ return 0;
+ sp_lex(sp); /* skip ( */
+ if (!sp_expr(sp, n, wrd))
+ return 0;
+ while (sp->lookahead == ',')
+ {
+ RecWord search_w;
+ int i;
+ sp_lex(sp); /* skip , */
+
+ if (!sp_expr(sp, n, &search_w))
+ return 0;
+ for (i = 0; i<wrd->term_len; i++)
+ {
+ int j;
+ for (j = 0; j<search_w.term_len && i+j < wrd->term_len; j++)
+ if (wrd->term_buf[i+j] != search_w.term_buf[j])
+ break;
+ if (j == search_w.term_len) /* match ? */
+ {
+ if (min_pos == -1 || i < min_pos)
+ min_pos = i;
+ break;
+ }
+ }
+ }
+ if (sp->lookahead != ')')
+ return 0;
+ sp_lex(sp);
+ if (min_pos == -1)
+ min_pos = 0; /* the default if not found */
+ sprintf(num_str, "%d", min_pos);
+ wrd->term_buf = nmem_strdup(sp->nmem, num_str);
+ wrd->term_len = strlen(wrd->term_buf);
+ return 1;
+}
static int sp_expr(struct source_parser *sp, data1_node *n, RecWord *wrd)
{
{
if (n->which == DATA1N_data)
{
- wrd->string = n->u.data.data;
- wrd->length = n->u.data.len;
+ wrd->term_buf = n->u.data.data;
+ wrd->term_len = n->u.data.len;
}
sp_lex(sp);
}
{
if (n->which == DATA1N_tag)
{
- wrd->string = n->u.tag.tag;
- wrd->length = strlen(n->u.tag.tag);
+ wrd->term_buf = n->u.tag.tag;
+ wrd->term_len = strlen(n->u.tag.tag);
}
sp_lex(sp);
}
else if (sp->len == 4 && !memcmp(sp->tok, "attr", sp->len))
{
+ RecWord tmp_w;
sp_lex(sp);
if (sp->lookahead != '(')
return 0;
sp_lex(sp);
- if (sp->lookahead != 't')
+
+ if (!sp_expr(sp, n, &tmp_w))
return 0;
+ wrd->term_buf = "";
+ wrd->term_len = 0;
if (n->which == DATA1N_tag)
{
data1_xattr *p = n->u.tag.attributes;
- while (p && strlen(p->name) != sp->len &&
- memcmp (p->name, sp->tok, sp->len))
+ while (p && strlen(p->name) != tmp_w.term_len &&
+ memcmp (p->name, tmp_w.term_buf, tmp_w.term_len))
p = p->next;
if (p)
{
- wrd->string = p->value;
- wrd->length = strlen(p->value);
+ wrd->term_buf = p->value;
+ wrd->term_len = strlen(p->value);
}
}
- sp_lex(sp);
if (sp->lookahead != ')')
return 0;
sp_lex(sp);
}
+ else if (sp->len == 5 && !memcmp(sp->tok, "first", sp->len))
+ {
+ return sp_first(sp, n, wrd);
+ }
else if (sp->len == 5 && !memcmp(sp->tok, "range", sp->len))
{
- int start, len;
- sp_lex(sp);
- if (sp->lookahead != '(')
- return 0;
-
- sp_lex(sp);
- sp_expr(sp, n, wrd);
- if (sp->lookahead != ',')
- return 0;
-
- sp_lex(sp);
- if (sp->lookahead != 't')
- return 0;
- start = atoi_n(sp->tok, sp->len);
-
- sp_lex(sp);
- if (sp->lookahead != ',')
- return 0;
-
+ return sp_range(sp, n, wrd);
+ }
+ else if (sp->len > 0 && isdigit(*(unsigned char *)sp->tok))
+ {
+ wrd->term_buf = nmem_malloc(sp->nmem, sp->len);
+ memcpy(wrd->term_buf, sp->tok, sp->len);
+ wrd->term_len = sp->len;
sp_lex(sp);
- if (sp->lookahead != 't')
- return 0;
- len = atoi_n(sp->tok, sp->len);
-
+ }
+ else if (sp->len > 2 && sp->tok[0] == '\'' && sp->tok[sp->len-1] == '\'')
+ {
+ wrd->term_len = sp->len - 2;
+ wrd->term_buf = nmem_malloc(sp->nmem, wrd->term_len);
+ memcpy(wrd->term_buf, sp->tok+1, wrd->term_len);
sp_lex(sp);
- if (sp->lookahead != ')')
- return 0;
-
+ }
+ else
+ {
+ wrd->term_buf = "";
+ wrd->term_len = 0;
sp_lex(sp);
- if (wrd->string && wrd->length)
- {
- wrd->string += start;
- wrd->length -= start;
- if (wrd->length > len)
- wrd->length = len;
- }
}
return 1;
}
-static int sp_parse(data1_node *n, RecWord *wrd, const char *src)
+static struct source_parser *source_parser_create()
+{
+ struct source_parser *sp = xmalloc(sizeof(*sp));
+
+ sp->nmem = nmem_create();
+ return sp;
+}
+
+static void source_parser_destroy(struct source_parser *sp)
+{
+ if (!sp)
+ return;
+ nmem_destroy(sp->nmem);
+ xfree(sp);
+}
+
+static int sp_parse(struct source_parser *sp,
+ data1_node *n, RecWord *wrd, const char *src)
{
- struct source_parser sp;
- sp.len = 0;
- sp.tok = 0;
- sp.src = src;
- sp.lookahead = 0;
- sp_lex(&sp);
-
- return sp_expr(&sp, n, wrd);
+ sp->len = 0;
+ sp->tok = 0;
+ sp->src = src;
+ sp->lookahead = 0;
+ nmem_reset(sp->nmem);
+
+ sp_lex(sp);
+ return sp_expr(sp, n, wrd);
}
int d1_check_xpath_predicate(data1_node *n, struct xpath_predicate *p)
wrd->attrSet = VAL_IDXPATH;
wrd->attrUse = 1;
wrd->reg_type = '0';
- wrd->string = tag_path;
- wrd->length = strlen(tag_path);
+ wrd->term_buf = tag_path;
+ wrd->term_len = strlen(tag_path);
(*p->tokenAdd)(wrd);
if (value) {
wrd->attrUse = 1015;
wrd->reg_type = 'w';
- wrd->string = value;
- wrd->length = strlen(value);
+ wrd->term_buf = value;
+ wrd->term_len = strlen(value);
(*p->tokenAdd)(wrd);
}
wrd->attrUse = 2;
wrd->reg_type = '0';
- wrd->string = tag_path;
- wrd->length = strlen(tag_path);
+ wrd->term_buf = tag_path;
+ wrd->term_len = strlen(tag_path);
(*p->tokenAdd)(wrd);
}
-static void index_xpath (data1_node *n, struct recExtractCtrl *p,
+static void index_xpath (struct source_parser *sp, data1_node *n,
+ struct recExtractCtrl *p,
int level, RecWord *wrd, int use)
{
int i;
yaz_log(YLOG_DEBUG, "index_xpath level=%d use=%d", level, use);
if ((!n->root->u.root.absyn) ||
(n->root->u.root.absyn->enable_xpath_indexing)) {
- termlist_only = 0;
+ termlist_only = 0;
}
switch (n->which)
{
case DATA1N_data:
- wrd->string = n->u.data.data;
- wrd->length = n->u.data.len;
+ wrd->term_buf = n->u.data.data;
+ wrd->term_len = n->u.data.len;
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) {
+ 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;
+ if (tlen + flen > (sizeof(tag_path_full)-2))
+ break;
memcpy (tag_path_full + flen, nn->u.tag.tag, tlen);
flen += tlen;
tag_path_full[flen++] = '/';
}
- else if (nn->which == DATA1N_root) break;
+ else
+ if (nn->which == DATA1N_root)
+ break;
}
tag_path_full[flen] = 0;
/* this is the ! case, so structure is for the xpath index */
memcpy (&wrd_tl, wrd, sizeof(*wrd));
if (tl->source)
- sp_parse(n, &wrd_tl, tl->source);
+ sp_parse(sp, n, &wrd_tl, tl->source);
if (!tl->att) {
wrd_tl.attrSet = VAL_IDXPATH;
wrd_tl.attrUse = use;
int i;
printf("%*sXPath index", (level + 1) * 4, "");
printf (" XData:\"");
- for (i = 0; i<wrd_tl.length && i < 40; i++)
- fputc (wrd_tl.string[i], stdout);
+ for (i = 0; i<wrd_tl.term_len && i < 40; i++)
+ fputc (wrd_tl.term_buf[i], stdout);
fputc ('"', stdout);
- if (wrd_tl.length > 40)
+ if (wrd_tl.term_len > 40)
printf (" ...");
fputc ('\n', stdout);
}
tl->att->name, tl->att->value,
tl->source);
printf (" XData:\"");
- for (i = 0; i<wrd_tl.length && i < 40; i++)
- fputc (wrd_tl.string[i], stdout);
+ for (i = 0; i<wrd_tl.term_len && i < 40; i++)
+ fputc (wrd_tl.term_buf[i], stdout);
fputc ('"', stdout);
- if (wrd_tl.length > 40)
+ if (wrd_tl.term_len > 40)
printf (" ...");
fputc ('\n', stdout);
}
{
size_t tlen = strlen(nn->u.tag.tag);
if (tlen + flen > (sizeof(tag_path_full)-2))
- return;
+ break;
memcpy (tag_path_full + flen, nn->u.tag.tag, tlen);
flen += tlen;
tag_path_full[flen++] = '/';
wrd->reg_type = '0';
- wrd->string = tag_path_full;
- wrd->length = flen;
+ wrd->term_buf = tag_path_full;
+ wrd->term_len = 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);
+ for (i = 0; i<wrd->term_len && i < 40; i++)
+ fputc (wrd->term_buf[i], stdout);
if (i == 40)
printf (" ..");
printf("\n");
/* attribute (no value) */
wrd->reg_type = '0';
wrd->attrUse = 3;
- wrd->string = xp->name;
- wrd->length = strlen(xp->name);
+ wrd->term_buf = xp->name;
+ wrd->term_len = strlen(xp->name);
wrd->seqno--;
(*p->tokenAdd)(wrd);
wrd->attrUse = 3;
wrd->reg_type = '0';
- wrd->string = comb;
- wrd->length = strlen(comb);
+ wrd->term_buf = comb;
+ wrd->term_len = strlen(comb);
wrd->seqno--;
(*p->tokenAdd)(wrd);
(tl->att->parent->reference);
wrd->attrUse = tl->att->locals->local;
wrd->reg_type = *tl->structure;
- wrd->string = xp->value;
- wrd->length = strlen(xp->value);
+ wrd->term_buf = xp->value;
+ wrd->term_len = strlen(xp->value);
(*p->tokenAdd)(wrd);
}
}
}
}
-static void index_termlist (data1_node *par, data1_node *n,
+static void index_termlist (struct source_parser *sp, data1_node *par,
+ data1_node *n,
struct recExtractCtrl *p, int level, RecWord *wrd)
{
data1_termlist *tlist = 0;
return;
if (par->u.tag.element->tag)
dtype = par->u.tag.element->tag->kind;
-
+
for (; tlist; tlist = tlist->next)
{
/* consider source */
- wrd->string = 0;
+ wrd->term_buf = 0;
assert(tlist->source);
- sp_parse(n, wrd, tlist->source);
+ sp_parse(sp, n, wrd, tlist->source);
- if (wrd->string)
+ if (wrd->term_buf && wrd->term_len)
{
if (p->flagShowRecords)
{
tlist->att->name, tlist->att->value,
tlist->source);
printf (" XData:\"");
- for (i = 0; i<wrd->length && i < 40; i++)
- fputc (wrd->string[i], stdout);
+ for (i = 0; i<wrd->term_len && i < 40; i++)
+ fputc (wrd->term_buf[i], stdout);
fputc ('"', stdout);
- if (wrd->length > 40)
+ if (wrd->term_len > 40)
printf (" ...");
fputc ('\n', stdout);
}
}
}
-static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level,
- RecWord *wrd)
+static int dumpkeys_r(struct source_parser *sp,
+ data1_node *n, struct recExtractCtrl *p, int level,
+ RecWord *wrd)
{
for (; n; n = n->next)
{
if (n->which == DATA1N_tag)
{
- index_termlist (n, n, p, level, wrd);
+ index_termlist(sp, n, n, p, level, wrd);
/* index start tag */
if (n->root->u.root.absyn)
- index_xpath (n, p, level, wrd, 1);
+ index_xpath(sp, n, p, level, wrd, 1);
}
if (n->child)
- if (dumpkeys(n->child, p, level + 1, wrd) < 0)
+ if (dumpkeys_r(sp, n->child, p, level + 1, wrd) < 0)
return -1;
}
if (par)
- index_termlist (par, n, p, level, wrd);
+ index_termlist(sp, par, n, p, level, wrd);
- index_xpath (n, p, level, wrd, 1016);
+ index_xpath(sp, n, p, level, wrd, 1016);
}
if (n->which == DATA1N_tag)
{
/* index end tag */
- index_xpath (n, p, level, wrd, 2);
+ index_xpath(sp, n, p, level, wrd, 2);
}
if (p->flagShowRecords && n->which == DATA1N_root)
return 0;
}
+static int dumpkeys(data1_node *n, struct recExtractCtrl *p, RecWord *wrd)
+{
+ struct source_parser *sp = source_parser_create();
+ int r = dumpkeys_r(sp, n, p, 0, wrd);
+ source_parser_destroy(sp);
+ return r;
+}
+
int grs_extract_tree(struct recExtractCtrl *p, data1_node *n)
{
oident oe;
}
(*p->init)(p, &wrd);
- return dumpkeys(n, p, 0, &wrd);
+ return dumpkeys(n, p, &wrd);
}
static int grs_extract_sub(void *clientData, struct recExtractCtrl *p,
#endif
(*p->init)(p, &wrd);
- if (dumpkeys(n, p, 0, &wrd) < 0)
+ if (dumpkeys(n, p, &wrd) < 0)
{
data1_free_tree(p->dh, n);
return RECCTRL_EXTRACT_ERROR_GENERIC;
}
data1_concat_text(p->dh, mem, node);
- /* 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
#if 0
data1_pr_tree (p->dh, node, stdout);
#endif
-#if YAZ_VERSIONL >= 0x010903L
if (p->comp && p->comp->which == Z_RecordComp_complex &&
p->comp->u.complex->generic &&
p->comp->u.complex->generic->which == Z_Schema_oid &&
if (oe)
requested_schema = oe->value;
}
-#else
- if (p->comp && p->comp->which == Z_RecordComp_complex &&
- p->comp->u.complex->generic && p->comp->u.complex->generic->schema)
- {
- oident *oe = oid_getentbyoid (p->comp->u.complex->generic->schema);
- if (oe)
- requested_schema = oe->value;
- }
-#endif
-
/* If schema has been specified, map if possible, then check that
* we got the right one
*/
p->input_format : VAL_SUTRS))
{
case VAL_TEXT_XML:
-
#if 0
data1_pr_tree (p->dh, node, stdout);
#endif
-
- if (p->encoding)
- data1_iconv (p->dh, mem, node, p->encoding, "UTF-8");
+ /* default output encoding for XML is UTF-8 */
+ data1_iconv (p->dh, mem, node,
+ p->encoding ? p->encoding : "UTF-8",
+ data1_get_encoding(p->dh, node));
if (!(p->rec_buf = data1_nodetoidsgml(p->dh, node, selected,
&p->rec_len)))
}
break;
case VAL_GRS1:
+ data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node));
dummy = 0;
if (!(p->rec_buf = data1_nodetogr(p->dh, node, selected,
p->odr, &dummy)))
p->rec_len = (size_t) (-1);
break;
case VAL_EXPLAIN:
+ /* ensure our data1 tree is UTF-8 */
+ data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node));
+
if (!(p->rec_buf = data1_nodetoexplain(p->dh, node, selected,
p->odr)))
p->diagnostic = 238;
p->rec_len = (size_t) (-1);
break;
case VAL_SUMMARY:
+ /* ensure our data1 tree is UTF-8 */
+ data1_iconv (p->dh, mem, node, "UTF-8", data1_get_encoding(p->dh, node));
if (!(p->rec_buf = data1_nodetosummary(p->dh, node, selected,
p->odr)))
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->encoding)
+ data1_iconv (p->dh, mem, node, p->encoding,
+ data1_get_encoding(p->dh, node));
if (!(p->rec_buf = data1_nodetobuf(p->dh, node, selected,
&p->rec_len)))
p->diagnostic = 238;
}
break;
case VAL_SOIF:
+ if (p->encoding)
+ data1_iconv (p->dh, mem, node, p->encoding,
+ data1_get_encoding(p->dh, node));
if (!(p->rec_buf = data1_nodetosoif(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->encoding)
+ data1_iconv (p->dh, mem, node, p->encoding,
+ data1_get_encoding(p->dh, node));
if (!(p->rec_buf = data1_nodetomarc(p->dh, marctab, node,
selected, &p->rec_len)))
p->diagnostic = 238;