/*
- * Copyright (C) 1994-1997, Index Data I/S
+ * Copyright (C) 1994-1999, Index Data
* All rights reserved.
* Sebastian Hammer, Adam Dickmeiss
*
* $Log: recgrs.c,v $
- * Revision 1.10 1997-09-18 08:59:21 adam
+ * 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
#include <stdio.h>
#include <assert.h>
#include <sys/types.h>
-#ifndef WINDOWS
+#ifndef WIN32
#include <unistd.h>
#endif
#include <oid.h>
#include <recctrl.h>
-#include <charmap.h>
#include "grsread.h"
#define GRS_MAX_WORD 512
-static int seqno = 0;
+struct grs_handler {
+ RecTypeGrs type;
+ void *clientData;
+ int initFlag;
+ struct grs_handler *next;
+};
+
+struct grs_handlers {
+ struct grs_handler *handlers;
+};
-static data1_node *read_grs_type (struct grs_read_info *p, const char *type)
+static int read_grs_type (struct grs_handlers *h,
+ struct grs_read_info *p, const char *type,
+ data1_node **root)
{
- static struct {
- char *type;
- data1_node *(*func)(struct grs_read_info *p);
- } tab[] = {
- { "sgml", grs_read_sgml },
- { "regx", grs_read_regx },
- { "marc", grs_read_marc },
- { NULL, NULL }
- };
+ struct grs_handler *gh = h->handlers;
const char *cp = strchr (type, '.');
- int i;
if (cp == NULL || cp == type)
{
}
else
strcpy (p->type, cp+1);
- for (i=0; tab[i].type; i++)
+ for (gh = h->handlers; gh; gh = gh->next)
{
- if (!memcmp (type, tab[i].type, cp-type))
- return (tab[i].func)(p);
+ if (!memcmp (type, gh->type->type, cp-type))
+ {
+ if (!gh->initFlag)
+ {
+ gh->initFlag = 1;
+ gh->clientData = (*gh->type->init)();
+ }
+ p->clientData = gh->clientData;
+ *root = (gh->type->read)(p);
+ gh->clientData = p->clientData;
+ return 0;
+ }
}
- return NULL;
+ return 1;
}
-static void grs_init(void)
+static void grs_add_handler (struct grs_handlers *h, RecTypeGrs t)
{
+ struct grs_handler *gh = (struct grs_handler *) malloc (sizeof(*gh));
+ gh->next = h->handlers;
+ h->handlers = gh;
+ gh->initFlag = 0;
+ gh->clientData = 0;
+ gh->type = t;
}
-static void dumpkeys_word(data1_node *n, struct recExtractCtrl *p,
- data1_att *att)
+static void *grs_init(RecType recType)
{
- const char *b = n->u.data.data;
- int remain;
- const char **map = 0;
-
- remain = n->u.data.len - (b - n->u.data.data);
- if (remain > 0)
- map = (*p->map_chrs_input)(0, &b, remain);
-
- while (map)
- {
- RecWord wrd;
- char buf[GRS_MAX_WORD+1];
- int i, remain;
-
- /* Skip spaces */
- while (map && *map && **map == *CHR_SPACE)
- {
- remain = n->u.data.len - (b - n->u.data.data);
- if (remain > 0)
- map = (*p->map_chrs_input)(0, &b, remain);
- else
- map = 0;
- }
- if (!map)
- break;
- i = 0;
- while (map && *map && **map != *CHR_SPACE)
- {
- const char *cp = *map;
+ struct grs_handlers *h = (struct grs_handlers *) malloc (sizeof(*h));
+ h->handlers = 0;
- while (i < GRS_MAX_WORD && *cp)
- buf[i++] = *(cp++);
- remain = n->u.data.len - (b - n->u.data.data);
- if (remain > 0)
- map = (*p->map_chrs_input)(0, &b, remain);
- else
- map = 0;
- }
- if (!i)
- return;
- buf[i] = '\0';
- (*p->init)(&wrd); /* set defaults */
- wrd.which = Word_String;
- wrd.seqno = seqno++;
- wrd.u.string = buf;
- wrd.attrSet = att->parent->ordinal;
- wrd.attrUse = att->locals->local;
- (*p->add)(&wrd);
- }
+ grs_add_handler (h, recTypeGrs_sgml);
+ grs_add_handler (h, recTypeGrs_regx);
+#if HAVE_TCL_H
+ grs_add_handler (h, recTypeGrs_tcl);
+#endif
+ grs_add_handler (h, recTypeGrs_marc);
+ return h;
}
-static void dumpkeys_phrase(data1_node *n, struct recExtractCtrl *p,
- data1_att *att)
+static void grs_destroy(void *clientData)
{
- const char *b = n->u.data.data;
- char buf[GRS_MAX_WORD+1];
- const char **map = 0;
- RecWord wrd;
- int i = 0, remain;
-
- remain = n->u.data.len - (b - n->u.data.data);
- if (remain > 0)
- map = (*p->map_chrs_input)(0, &b, remain);
-
- while (remain > 0 && i < GRS_MAX_WORD)
+ struct grs_handlers *h = (struct grs_handlers *) clientData;
+ struct grs_handler *gh = h->handlers, *gh_next;
+ while (gh)
{
- while (map && *map && **map == *CHR_SPACE)
- {
- remain = n->u.data.len - (b - n->u.data.data);
- if (remain > 0)
- map = (*p->map_chrs_input)(0, &b, remain);
- else
- map = 0;
- }
- if (!map)
- break;
-
- if (i && i < GRS_MAX_WORD)
- buf[i++] = *CHR_SPACE;
- while (map && *map && **map != *CHR_SPACE)
- {
- const char *cp = *map;
-
- if (i >= GRS_MAX_WORD)
- break;
- while (i < GRS_MAX_WORD && *cp)
- buf[i++] = *(cp++);
- remain = n->u.data.len - (b - n->u.data.data);
- if (remain > 0)
- map = (*p->map_chrs_input)(0, &b, remain);
- else
- map = 0;
- }
+ gh_next = gh->next;
+ if (gh->initFlag)
+ (*gh->type->destroy)(gh->clientData);
+ free (gh);
+ gh = gh_next;
}
- if (!i)
- return;
- buf[i] = '\0';
- (*p->init)(&wrd);
- wrd.which = Word_Phrase;
- wrd.seqno = seqno++;
- wrd.u.string = buf;
- wrd.attrSet = att->parent->ordinal;
- wrd.attrUse = att->locals->local;
- (*p->add)(&wrd);
+ free (h);
}
static int dumpkeys(data1_node *n, struct recExtractCtrl *p, int level)
{
+ RecWord wrd;
+ (*p->init)(p, &wrd); /* set defaults */
for (; n; n = n->next)
{
if (p->flagShowRecords) /* display element description to user */
{
data1_node *par = get_parent_tag(p->dh, n);
data1_termlist *tlist = 0;
+ data1_datatype dtype = DATA1K_string;
if (p->flagShowRecords)
{
printf("%*s", level * 4, "");
printf("Data: ");
- if (n->u.data.len > 20)
- printf("'%.20s...'\n", n->u.data.data);
+ if (n->u.data.len > 32)
+ printf("'%.24s ... %.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);
else
*/
while (!par->u.tag.element)
- if (!par->parent || !(par = get_parent_tag(p->dh, par->parent)))
+ if (!par->parent || !(par=get_parent_tag(p->dh, par->parent)))
break;
- if (!par)
- tlist = 0;
- else if (par->u.tag.element->termlists)
- tlist = par->u.tag.element->termlists;
- else
+ if (!par || !(tlist = par->u.tag.element->termlists))
continue;
-
+ if (par->u.tag.element->tag)
+ dtype = par->u.tag.element->tag->kind;
for (; tlist; tlist = tlist->next)
{
if (p->flagShowRecords)
{
- printf("%*sIdx: [", (level + 1) * 4, "");
- switch (tlist->structure)
- {
- case DATA1S_word: printf("w"); break;
- case DATA1S_phrase: printf("p"); break;
- default: printf("?"); break;
- }
- printf("] ");
- printf("%s:%s [%d]\n", tlist->att->parent->name,
- tlist->att->name, tlist->att->value);
+ printf("%*sIdx: [%s]", (level + 1) * 4, "",
+ tlist->structure);
+ printf("%s:%s [%d]\n",
+ tlist->att->parent->name,
+ tlist->att->name, tlist->att->value);
}
- else switch (tlist->structure)
+ else
{
- case DATA1S_word:
- dumpkeys_word(n, p, tlist->att); break;
- case DATA1S_phrase:
- dumpkeys_phrase(n, p, tlist->att); break;
- default:
- logf(LOG_FATAL, "Bad structure type in dumpkeys");
- abort();
+ wrd.reg_type = *tlist->structure;
+ wrd.string = n->u.data.data;
+ wrd.length = n->u.data.len;
+ wrd.attrSet = (int) (tlist->att->parent->reference);
+ wrd.attrUse = tlist->att->locals->local;
+ (*p->addWord)(&wrd);
}
}
}
return 0;
}
-static int grs_extract(struct recExtractCtrl *p)
+int grs_extract_tree(struct recExtractCtrl *p, data1_node *n)
+{
+ oident oe;
+ int oidtmp[OID_SIZE];
+
+ oe.proto = PROTO_Z3950;
+ oe.oclass = CLASS_SCHEMA;
+ oe.value = n->u.root.absyn->reference;
+
+ if ((oid_ent_to_oid (&oe, oidtmp)))
+ (*p->addSchema)(p, oidtmp);
+
+ return dumpkeys(n, p, 0);
+}
+
+static int grs_extract(void *clientData, struct recExtractCtrl *p)
{
data1_node *n;
- NMEM mem = nmem_create();
+ NMEM mem;
struct grs_read_info gri;
- seqno = 0;
+ oident oe;
+ int oidtmp[OID_SIZE];
+ struct grs_handlers *h = (struct grs_handlers *) clientData;
+ mem = nmem_create ();
gri.readf = p->readf;
gri.seekf = p->seekf;
gri.tellf = p->tellf;
gri.mem = mem;
gri.dh = p->dh;
- n = read_grs_type (&gri, p->subType);
+ if (read_grs_type (h, &gri, p->subType, &n))
+ return RECCTRL_EXTRACT_ERROR;
if (!n)
- return -1;
+ 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->addSchema)(p, oidtmp);
+
if (dumpkeys(n, p, 0) < 0)
{
data1_free_tree(p->dh, n);
- return -2;
+ return RECCTRL_EXTRACT_ERROR;
}
data1_free_tree(p->dh, n);
nmem_destroy(mem);
- return 0;
+ return RECCTRL_EXTRACT_OK;
}
/*
return 26; /* fix */
}
if (espec)
+ {
+ logf (LOG_DEBUG, "Element: Espec-1 match");
return data1_doespec1(dh, n, espec);
+ }
else
+ {
+ logf (LOG_DEBUG, "Element: all match");
return -1;
+ }
}
-static int grs_retrieve(struct recRetrieveCtrl *p)
+static int grs_retrieve(void *clientData, struct recRetrieveCtrl *p)
{
data1_node *node = 0, *onode = 0;
- data1_node *new;
+ data1_node *dnew;
data1_maptab *map;
int res, selected = 0;
- NMEM mem = nmem_create();
+ NMEM mem;
struct grs_read_info gri;
+ char *tagname;
+ struct grs_handlers *h = (struct grs_handlers *) clientData;
+ mem = nmem_create();
gri.readf = p->readf;
gri.seekf = p->seekf;
gri.tellf = p->tellf;
gri.mem = mem;
gri.dh = p->dh;
- node = read_grs_type (&gri, p->subType);
+ logf (LOG_DEBUG, "grs_retrieve");
+ if (read_grs_type (h, &gri, p->subType, &node))
+ {
+ p->diagnostic = 14;
+ nmem_destroy (mem);
+ return 0;
+ }
if (!node)
{
- p->diagnostic = 2;
+ p->diagnostic = 14;
+ nmem_destroy (mem);
return 0;
}
- if (p->score >= 0 && (new =
+ logf (LOG_DEBUG, "grs_retrieve: size");
+ if ((dnew = data1_insert_taggeddata(p->dh, node, node,
+ "size", 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_insert_taggeddata(p->dh, node,
- node, "rank",
+ node, tagname,
mem)))
{
- new->u.data.what = DATA1I_num;
- new->u.data.data = new->u.data.lbuf;
- sprintf(new->u.data.data, "%d", p->score);
- new->u.data.len = strlen(new->u.data.data);
+ logf (LOG_DEBUG, "grs_retrieve: %s", tagname);
+ dnew->u.data.what = DATA1I_num;
+ dnew->u.data.data = dnew->lbuf;
+ sprintf(dnew->u.data.data, "%d", p->score);
+ dnew->u.data.len = strlen(dnew->u.data.data);
}
- if ((new = data1_insert_taggeddata(p->dh, node, node,
- "localControlNumber", mem)))
+
+ 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)))
{
- new->u.data.what = DATA1I_text;
- new->u.data.data = new->u.data.lbuf;
- sprintf(new->u.data.data, "%d", p->localno);
- new->u.data.len = strlen(new->u.data.data);
+ 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);
}
+
+ logf (LOG_DEBUG, "grs_retrieve: schemaIdentifier");
if (p->input_format == VAL_GRS1 && node->u.root.absyn &&
node->u.root.absyn->reference != VAL_NONE)
{
oident oe;
Odr_oid *oid;
+ int oidtmp[OID_SIZE];
oe.proto = PROTO_Z3950;
oe.oclass = CLASS_SCHEMA;
oe.value = node->u.root.absyn->reference;
- if ((oid = oid_getoidbyent(&oe)))
+ if ((oid = oid_ent_to_oid (&oe, oidtmp)))
{
char tmp[128];
data1_handle dh = p->dh;
}
*(p++) = '\0';
- if ((new = data1_insert_taggeddata(dh, node, node,
+ if ((dnew = data1_insert_taggeddata(dh, node, node,
"schemaIdentifier", mem)))
{
- new->u.data.what = DATA1I_oid;
- new->u.data.data = nmem_malloc(mem, p - tmp);
- memcpy(new->u.data.data, tmp, p - tmp);
- new->u.data.len = p - tmp;
+ dnew->u.data.what = DATA1I_oid;
+ dnew->u.data.data = (char *) nmem_malloc(mem, p - tmp);
+ memcpy(dnew->u.data.data, tmp, p - tmp);
+ dnew->u.data.len = p - tmp;
}
}
}
+ logf (LOG_DEBUG, "grs_retrieve: schema mapping");
/*
* Does the requested format match a known schema-mapping? (this reflects
* the overlap of schema and formatting which is inherent in the MARC
if (!(node = data1_map_record(p->dh, onode, map, mem)))
{
p->diagnostic = 14;
+ nmem_destroy (mem);
return 0;
}
-
break;
}
+ logf (LOG_DEBUG, "grs_retrieve: element spec");
if (p->comp && (res = process_comp(p->dh, node, p->comp)) > 0)
{
p->diagnostic = res;
else if (p->comp && !res)
selected = 1;
+ logf (LOG_DEBUG, "grs_retrieve: transfer syntax mapping");
switch (p->output_format = (p->input_format != VAL_NONE ?
p->input_format : VAL_SUTRS))
{
dummy = 0;
if (!(p->rec_buf = data1_nodetogr(p->dh, node, selected,
p->odr, &dummy)))
- p->diagnostic = 2; /* this should be better specified */
+ p->diagnostic = 238; /* not available in requested syntax */
else
- p->rec_len = -1;
+ p->rec_len = (size_t) (-1);
break;
case VAL_EXPLAIN:
if (!(p->rec_buf = data1_nodetoexplain(p->dh, node, selected,
p->odr)))
- p->diagnostic = 2; /* this should be better specified */
+ p->diagnostic = 238;
else
- p->rec_len = -1;
+ p->rec_len = (size_t) (-1);
break;
case VAL_SUMMARY:
if (!(p->rec_buf = data1_nodetosummary(p->dh, node, selected,
p->odr)))
- p->diagnostic = 2;
+ p->diagnostic = 238;
else
- p->rec_len = -1;
+ p->rec_len = (size_t) (-1);
break;
case VAL_SUTRS:
if (!(p->rec_buf = data1_nodetobuf(p->dh, node, selected,
(int*)&p->rec_len)))
+ p->diagnostic = 238;
+ else
{
- p->diagnostic = 2;
- break;
+ char *new_buf = (char*) odr_malloc (p->odr, p->rec_len);
+ memcpy (new_buf, p->rec_buf, p->rec_len);
+ p->rec_buf = new_buf;
}
break;
case VAL_SOIF:
if (!(p->rec_buf = data1_nodetosoif(p->dh, node, selected,
(int*)&p->rec_len)))
+ p->diagnostic = 238;
+ else
{
- p->diagnostic = 2;
- break;
+ char *new_buf = (char*) odr_malloc (p->odr, p->rec_len);
+ memcpy (new_buf, p->rec_buf, p->rec_len);
+ p->rec_buf = new_buf;
}
break;
default:
+ if (!node->u.root.absyn)
+ {
+ p->diagnostic = 238;
+ break;
+ }
for (marctab = node->u.root.absyn->marc; marctab;
marctab = marctab->next)
if (marctab->reference == p->input_format)
break;
if (!marctab)
{
- p->diagnostic = 227;
+ p->diagnostic = 238;
break;
}
if (!(p->rec_buf = data1_nodetomarc(p->dh, marctab, node,
selected,
(int*)&p->rec_len)))
+ p->diagnostic = 238;
+ else
{
- p->diagnostic = 2;
- break;
+ char *new_buf = (char*) odr_malloc (p->odr, p->rec_len);
+ memcpy (new_buf, p->rec_buf, p->rec_len);
+ p->rec_buf = new_buf;
}
}
if (node)
{
"grs",
grs_init,
+ grs_destroy,
grs_extract,
grs_retrieve
};
projects / idzebra-moved-to-github.git / blobdiff