/* This file is part of Pazpar2.
- Copyright (C) 2006-2008 Index Data
+ Copyright (C) 2006-2009 Index Data
Pazpar2 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
#if HAVE_UNISTD_H
#include <unistd.h>
#endif
-#if HAVE_SYS_SOCKET_H
-#include <sys/socket.h>
-#endif
-#if HAVE_NETDB_H
-#include <netdb.h>
-#endif
#include <signal.h>
#include <ctype.h>
#include <assert.h>
#include <yaz/timing.h>
#endif
-#if HAVE_NETINET_IN_H
-#include <netinet/in.h>
-#endif
-
+#include "parameters.h"
#include "pazpar2.h"
#include "eventl.h"
#include "http.h"
// Note: Some things in this structure will eventually move to configuration
struct parameters global_parameters =
{
- "",
- "",
- "",
- 0,
0, // dump_records
0, // debug_mode
- 30, // operations timeout
- "81",
- "Index Data PazPar2",
- VERSION,
- 60, // session timeout
100,
- MAX_CHUNK,
- 0,
- 0,
- 180, // Z39.50 session timeout
- 15 // Connect timeout
};
+static void log_xml_doc(xmlDoc *doc)
+{
+ FILE *lf = yaz_log_file();
+ xmlChar *result = 0;
+ int len = 0;
+#if LIBXML_VERSION >= 20600
+ xmlDocDumpFormatMemory(doc, &result, &len, 1);
+#else
+ xmlDocDumpMemory(doc, &result, &len);
+#endif
+ if (lf && len)
+ {
+ fwrite(result, 1, len, lf);
+ fprintf(lf, "\n");
+ }
+ xmlFree(result);
+}
+
// Recursively traverse query structure to extract terms.
void pull_terms(NMEM nmem, struct ccl_rpn_node *n, char **termlist, int *num)
{
termlist_insert(s->termlists[i].termlist, value);
}
-xmlDoc *record_to_xml(struct session_database *sdb, Z_External *rec)
+xmlDoc *record_to_xml(struct session_database *sdb, const char *rec)
{
struct database *db = sdb->database;
xmlDoc *rdoc = 0;
- const Odr_oid *oid = rec->direct_reference;
- /* convert response record to XML somehow */
- if (rec->which == Z_External_octet && oid
- && !oid_oidcmp(oid, yaz_oid_recsyn_xml))
- {
- /* xml already */
- rdoc = xmlParseMemory((char*) rec->u.octet_aligned->buf,
- rec->u.octet_aligned->len);
- if (!rdoc)
- {
- yaz_log(YLOG_FATAL, "Non-wellformed XML received from %s",
- db->url);
- return 0;
- }
- }
- else if (rec->which == Z_External_OPAC)
- {
- if (!sdb->yaz_marc)
- {
- yaz_log(YLOG_WARN, "MARC decoding not configured");
- return 0;
- }
- else
- {
- /* OPAC gets converted to XML too */
- WRBUF wrbuf_opac = wrbuf_alloc();
- /* MARCXML inside the OPAC XML. Charset is in effect because we
- use the yaz_marc handle */
- yaz_marc_xml(sdb->yaz_marc, YAZ_MARC_MARCXML);
- yaz_opac_decode_wrbuf(sdb->yaz_marc, rec->u.opac, wrbuf_opac);
-
- rdoc = xmlParseMemory((char*) wrbuf_buf(wrbuf_opac),
- wrbuf_len(wrbuf_opac));
- if (!rdoc)
- {
- yaz_log(YLOG_WARN, "Unable to parse OPAC XML");
- /* Was used to debug bug #1348 */
-#if 0
- FILE *f = fopen("/tmp/opac.xml.txt", "wb");
- if (f)
- {
- fwrite(wrbuf_buf(wrbuf_opac), 1, wrbuf_len(wrbuf_opac), f);
- fclose(f);
- }
-#endif
- }
- wrbuf_destroy(wrbuf_opac);
- }
- }
- else if (oid && yaz_oid_is_iso2709(oid))
- {
- /* ISO2709 gets converted to MARCXML */
- if (!sdb->yaz_marc)
- {
- yaz_log(YLOG_WARN, "MARC decoding not configured");
- return 0;
- }
- else
- {
- xmlNode *res;
- char *buf;
- int len;
-
- if (rec->which != Z_External_octet)
- {
- yaz_log(YLOG_WARN, "Unexpected external branch, probably BER %s",
- db->url);
- return 0;
- }
- buf = (char*) rec->u.octet_aligned->buf;
- len = rec->u.octet_aligned->len;
- if (yaz_marc_read_iso2709(sdb->yaz_marc, buf, len) < 0)
- {
- yaz_log(YLOG_WARN, "Failed to decode MARC %s", db->url);
- return 0;
- }
-
- if (yaz_marc_write_xml(sdb->yaz_marc, &res,
- "http://www.loc.gov/MARC21/slim", 0, 0) < 0)
- {
- yaz_log(YLOG_WARN, "Failed to encode as XML %s",
- db->url);
- return 0;
- }
- rdoc = xmlNewDoc((xmlChar *) "1.0");
- xmlDocSetRootElement(rdoc, res);
- }
- }
- else
+ rdoc = xmlParseMemory(rec, strlen(rec));
+
+ if (!rdoc)
{
- char oid_name_buf[OID_STR_MAX];
- const char *oid_name = yaz_oid_to_string_buf(oid, 0, oid_name_buf);
- yaz_log(YLOG_FATAL,
- "Unable to handle record of type %s from %s",
- oid_name, db->url);
+ yaz_log(YLOG_FATAL, "Non-wellformed XML received from %s",
+ db->url);
return 0;
}
-
+
if (global_parameters.dump_records)
{
- FILE *lf = yaz_log_file();
- if (lf)
- {
- yaz_log(YLOG_LOG, "Un-normalized record from %s", db->url);
-#if LIBXML_VERSION >= 20600
- xmlDocFormatDump(lf, rdoc, 1);
-#else
- xmlDocDump(lf, rdoc);
-#endif
- fprintf(lf, "\n");
- }
+ yaz_log(YLOG_LOG, "Un-normalized record from %s", db->url);
+ log_xml_doc(rdoc);
}
+
return rdoc;
}
static void insert_settings_parameters(struct session_database *sdb,
struct session *se, char **parms)
{
- struct conf_service *service = global_parameters.server->service;
+ struct conf_service *service = se->service;
int i;
int nparms = 0;
int offset = 0;
int setting;
if (md->setting == Metadata_setting_parameter &&
- (setting = settings_offset(md->name)) > 0)
+ (setting = settings_offset(service, md->name)) > 0)
{
const char *val = session_setting_oneval(sdb, setting);
if (val && nparms < MAX_XSLT_ARGS)
}
// Add static values from session database settings if applicable
-static void insert_settings_values(struct session_database *sdb, xmlDoc *doc)
+static void insert_settings_values(struct session_database *sdb, xmlDoc *doc,
+ struct conf_service *service)
{
- struct conf_service *service = global_parameters.server->service;
int i;
for (i = 0; i < service->num_metadata; i++)
int offset;
if (md->setting == Metadata_setting_postproc &&
- (offset = settings_offset(md->name)) > 0)
+ (offset = settings_offset(service, md->name)) > 0)
{
const char *val = session_setting_oneval(sdb, offset);
if (val)
}
xmlDoc *normalize_record(struct session_database *sdb, struct session *se,
- Z_External *rec)
+ const char *rec)
{
- struct database_retrievalmap *m;
xmlDoc *rdoc = record_to_xml(sdb, rec);
+
if (rdoc)
{
- for (m = sdb->map; m; m = m->next)
+ char *parms[MAX_XSLT_ARGS*2+1];
+
+ insert_settings_parameters(sdb, se, parms);
+
+ if (normalize_record_transform(sdb->map, &rdoc, (const char **)parms))
{
- xmlDoc *new = 0;
-
- {
- xmlNodePtr root = 0;
- char *parms[MAX_XSLT_ARGS*2+1];
-
- insert_settings_parameters(sdb, se, parms);
-
- new = xsltApplyStylesheet(m->stylesheet, rdoc, (const char **) parms);
- root= xmlDocGetRootElement(new);
- if (!new || !root || !(root->children))
- {
- yaz_log(YLOG_WARN, "XSLT transformation failed from %s",
- sdb->database->url);
- xmlFreeDoc(new);
- xmlFreeDoc(rdoc);
- return 0;
- }
- }
-
- xmlFreeDoc(rdoc);
- rdoc = new;
+ yaz_log(YLOG_WARN, "Normalize failed from %s", sdb->database->url);
}
-
- insert_settings_values(sdb, rdoc);
-
- if (global_parameters.dump_records)
+ else
{
- FILE *lf = yaz_log_file();
+ insert_settings_values(sdb, rdoc, se->service);
- if (lf)
+ if (global_parameters.dump_records)
{
yaz_log(YLOG_LOG, "Normalized record from %s",
sdb->database->url);
-#if LIBXML_VERSION >= 20600
- xmlDocFormatDump(lf, rdoc, 1);
-#else
- xmlDocDump(lf, rdoc);
-#endif
- fprintf(lf, "\n");
+ log_xml_doc(rdoc);
}
}
}
return db->settings[offset]->value;
}
-
-
-// Initialize YAZ Map structures for MARC-based targets
-static int prepare_yazmarc(struct session_database *sdb)
-{
- const char *s;
-
- if (!sdb->settings)
- {
- yaz_log(YLOG_WARN, "No settings for %s", sdb->database->url);
- return -1;
- }
- if ((s = session_setting_oneval(sdb, PZ_NATIVESYNTAX))
- && !strncmp(s, "iso2709", 7))
- {
- char *encoding = "marc-8s", *e;
- yaz_iconv_t cm;
-
- // See if a native encoding is specified
- if ((e = strchr(s, ';')))
- encoding = e + 1;
-
- sdb->yaz_marc = yaz_marc_create();
- yaz_marc_subfield_str(sdb->yaz_marc, "\t");
-
- cm = yaz_iconv_open("utf-8", encoding);
- if (!cm)
- {
- yaz_log(YLOG_FATAL,
- "Unable to map from %s to UTF-8 for target %s",
- encoding, sdb->database->url);
- return -1;
- }
- yaz_marc_iconv(sdb->yaz_marc, cm);
- }
- return 0;
-}
-
// Prepare XSLT stylesheets for record normalization
// Structures are allocated on the session_wide nmem to avoid having
// to recompute this for every search. This would lead
}
if ((s = session_setting_oneval(sdb, PZ_XSLT)))
{
- char **stylesheets;
- struct database_retrievalmap **m = &sdb->map;
- int num, i;
char auto_stylesheet[256];
if (!strcmp(s, "auto"))
yaz_log(YLOG_WARN, "No pz:requestsyntax for auto stylesheet");
}
}
- nmem_strsplit(se->session_nmem, ",", s, &stylesheets, &num);
- for (i = 0; i < num; i++)
- {
- (*m) = nmem_malloc(se->session_nmem, sizeof(**m));
- (*m)->next = 0;
- if (!((*m)->stylesheet = conf_load_stylesheet(stylesheets[i])))
- {
- yaz_log(YLOG_FATAL|YLOG_ERRNO, "Unable to load stylesheet: %s",
- stylesheets[i]);
- return -1;
- }
- m = &(*m)->next;
- }
+ sdb->map = normalize_record_create(se->service, s);
}
- if (!sdb->map)
- yaz_log(YLOG_WARN, "No Normalization stylesheet for target %s",
- sdb->database->url);
return 0;
}
"No settings associated with %s", sdb->database->url);
return -1;
}
- if (sdb->settings[PZ_NATIVESYNTAX] && !sdb->yaz_marc)
- {
- if (prepare_yazmarc(sdb) < 0)
- return -1;
- }
if (sdb->settings[PZ_XSLT] && !sdb->map)
{
if (prepare_map(se, sdb) < 0)
int subnum;
int subi;
struct database_criterion *new = nmem_malloc(m, sizeof(*new));
- char *eq = strchr(values[i], '=');
+ char *eq;
+ if ((eq = strchr(values[i], '=')))
+ new->type = PAZPAR2_STRING_MATCH;
+ if ((eq = strchr(values[i], '~')))
+ new->type = PAZPAR2_SUBSTRING_MATCH;
if (!eq)
{
- yaz_log(YLOG_WARN, "Missing equal-sign in filter");
+ yaz_log(YLOG_WARN, "Missing equal-signi/tilde in filter");
return 0;
}
*(eq++) = '\0';
}
enum pazpar2_error_code search(struct session *se,
- char *query, char *filter,
+ const char *query, const char *filter,
const char **addinfo)
{
int live_channels = 0;
se->reclist = 0;
se->num_termlists = 0;
criteria = parse_filter(se->nmem, filter);
- se->requestid++;
live_channels = select_targets(se, criteria);
if (live_channels)
{
- int maxrecs = live_channels * global_parameters.toget;
+ int maxrecs = live_channels * global_parameters.toget; // This is buggy!!!
se->reclist = reclist_create(se->nmem, maxrecs);
se->expected_maxrecs = maxrecs;
}
else
{
no_working++;
- client_prep_connection(cl);
+ if (client_prep_connection(cl, se->service->z3950_operation_timeout,
+ se->service->z3950_session_timeout))
+ client_start_search(cl);
}
}
static void session_init_databases_fun(void *context, struct database *db)
{
struct session *se = (struct session *) context;
+ struct conf_service *service = se->service;
struct session_database *new = nmem_malloc(se->session_nmem, sizeof(*new));
- int num = settings_num();
+ int num = settings_num(service);
int i;
new->database = db;
- new->yaz_marc = 0;
new->map = 0;
new->settings
// Doesn't free memory associated with sdb -- nmem takes care of that
static void session_database_destroy(struct session_database *sdb)
{
- struct database_retrievalmap *m;
-
- for (m = sdb->map; m; m = m->next)
- xsltFreeStylesheet(m->stylesheet);
- if (sdb->yaz_marc)
- yaz_marc_destroy(sdb->yaz_marc);
+ normalize_record_destroy(sdb->map);
+ sdb->map = 0;
}
// Initialize session_database list -- this represents this session's view
void session_init_databases(struct session *se)
{
se->databases = 0;
- predef_grep_databases(se, 0, session_init_databases_fun);
+ predef_grep_databases(se, se->service, 0, session_init_databases_fun);
}
// Probably session_init_databases_fun should be refactored instead of
static struct session_database *load_session_database(struct session *se,
char *id)
{
- struct database *db = find_database(id, 0);
+ struct database *db = find_database(id, 0, se->service);
+
+ resolve_database(db);
session_init_databases_fun((void*) se, db);
// New sdb is head of se->databases list
char *value)
{
struct session_database *sdb = find_session_database(se, dbname);
+ struct conf_service *service = se->service;
struct setting *new = nmem_malloc(se->session_nmem, sizeof(*new));
- int offset = settings_offset_cprefix(setting);
+ int offset = settings_offset_cprefix(service, setting);
if (offset < 0)
{
// (happens when a search starts and client connections are prepared)
switch (offset)
{
- case PZ_NATIVESYNTAX:
- if (sdb->yaz_marc)
- {
- yaz_marc_destroy(sdb->yaz_marc);
- sdb->yaz_marc = 0;
- }
- break;
case PZ_XSLT:
if (sdb->map)
{
- struct database_retrievalmap *m;
- // We don't worry about the map structure -- it's in nmem
- for (m = sdb->map; m; m = m->next)
- xsltFreeStylesheet(m->stylesheet);
+ normalize_record_destroy(sdb->map);
sdb->map = 0;
}
break;
for (sdb = s->databases; sdb; sdb = sdb->next)
session_database_destroy(sdb);
nmem_destroy(s->nmem);
+ service_destroy(s->service);
wrbuf_destroy(s->wrbuf);
}
-struct session *new_session(NMEM nmem)
+struct session *new_session(NMEM nmem, struct conf_service *service)
{
int i;
struct session *session = nmem_malloc(nmem, sizeof(*session));
yaz_log(YLOG_DEBUG, "New Pazpar2 session");
-
+
+ session->service = service;
session->relevance = 0;
session->total_hits = 0;
session->total_records = 0;
+ session->number_of_warnings_unknown_elements = 0;
+ session->number_of_warnings_unknown_metadata = 0;
session->num_termlists = 0;
session->reclist = 0;
- session->requestid = -1;
session->clients = 0;
session->expected_maxrecs = 0;
session->session_nmem = nmem;
}
#endif
-struct record_cluster *show_single(struct session *s, const char *id)
+struct record_cluster *show_single(struct session *s, const char *id,
+ struct record_cluster **prev_r,
+ struct record_cluster **next_r)
{
struct record_cluster *r;
reclist_rewind(s->reclist);
+ *prev_r = 0;
+ *next_r = 0;
while ((r = reclist_read_record(s->reclist)))
+ {
if (!strcmp(r->recid, id))
+ {
+ *next_r = reclist_read_record(s->reclist);
return r;
+ }
+ *prev_r = r;
+ }
return 0;
}
switch (client_get_state(cl))
{
case Client_Connecting: stat->num_connecting++; break;
- case Client_Initializing: stat->num_initializing++; break;
- case Client_Searching: stat->num_searching++; break;
- case Client_Presenting: stat->num_presenting++; break;
+ case Client_Working: stat->num_working++; break;
case Client_Idle: stat->num_idle++; break;
case Client_Failed: stat->num_failed++; break;
case Client_Error: stat->num_error++; break;
stat->num_clients = count;
}
-void start_http_listener(void)
-{
- char hp[128] = "";
- struct conf_server *ser = global_parameters.server;
-
- if (*global_parameters.listener_override)
- strcpy(hp, global_parameters.listener_override);
- else
- {
- strcpy(hp, ser->host ? ser->host : "");
- if (ser->port)
- {
- if (*hp)
- strcat(hp, ":");
- sprintf(hp + strlen(hp), "%d", ser->port);
- }
- }
- http_init(hp);
-}
-
-void start_proxy(void)
-{
- char hp[128] = "";
- struct conf_server *ser = global_parameters.server;
-
- if (*global_parameters.proxy_override)
- strcpy(hp, global_parameters.proxy_override);
- else if (ser->proxy_host || ser->proxy_port)
- {
- strcpy(hp, ser->proxy_host ? ser->proxy_host : "");
- if (ser->proxy_port)
- {
- if (*hp)
- strcat(hp, ":");
- sprintf(hp + strlen(hp), "%d", ser->proxy_port);
- }
- }
- else
- return;
-
- http_set_proxyaddr(hp, ser->myurl ? ser->myurl : "");
-}
-
// Master list of connections we're handling events to
-static IOCHAN channel_list = 0;
+static IOCHAN channel_list = 0; /* thread pr */
+
void pazpar2_add_channel(IOCHAN chan)
{
chan->next = channel_list;
return rec_md;
}
-struct record *ingest_record(struct client *cl, Z_External *rec,
+static int get_mergekey_from_doc(xmlDoc *doc, xmlNode *root, const char *name,
+ struct conf_service *service, WRBUF norm_wr)
+{
+ xmlNode *n;
+ int no_found = 0;
+ for (n = root->children; n; n = n->next)
+ {
+ if (n->type != XML_ELEMENT_NODE)
+ continue;
+ if (!strcmp((const char *) n->name, "metadata"))
+ {
+ xmlChar *type = xmlGetProp(n, (xmlChar *) "type");
+ if (!strcmp(name, (const char *) type))
+ {
+ xmlChar *value = xmlNodeListGetString(doc, n->children, 1);
+ if (value)
+ {
+ const char *norm_str;
+ pp2_relevance_token_t prt =
+ pp2_relevance_tokenize(
+ service->mergekey_pct,
+ (const char *) value);
+
+ wrbuf_puts(norm_wr, name);
+ wrbuf_puts(norm_wr, "=");
+ while ((norm_str =
+ pp2_relevance_token_next(prt)))
+ {
+ if (*norm_str)
+ {
+ if (wrbuf_len(norm_wr))
+ wrbuf_puts(norm_wr, " ");
+ wrbuf_puts(norm_wr, norm_str);
+ }
+ }
+ xmlFree(value);
+ pp2_relevance_token_destroy(prt);
+ no_found++;
+ }
+ }
+ xmlFree(type);
+ }
+ }
+ return no_found;
+}
+
+static const char *get_mergekey(xmlDoc *doc, struct client *cl, int record_no,
+ struct conf_service *service, NMEM nmem)
+{
+ char *mergekey_norm = 0;
+ xmlNode *root = xmlDocGetRootElement(doc);
+ WRBUF norm_wr = wrbuf_alloc();
+
+ /* consider mergekey from XSL first */
+ xmlChar *mergekey = xmlGetProp(root, (xmlChar *) "mergekey");
+ if (mergekey)
+ {
+ const char *norm_str;
+ pp2_relevance_token_t prt =
+ pp2_relevance_tokenize(
+ service->mergekey_pct,
+ (const char *) mergekey);
+
+ while ((norm_str = pp2_relevance_token_next(prt)))
+ {
+ if (*norm_str)
+ {
+ if (wrbuf_len(norm_wr))
+ wrbuf_puts(norm_wr, " ");
+ wrbuf_puts(norm_wr, norm_str);
+ }
+ }
+ pp2_relevance_token_destroy(prt);
+ xmlFree(mergekey);
+ }
+ else
+ {
+ /* no mergekey defined in XSL. Look for mergekey metadata instead */
+ int field_id;
+ for (field_id = 0; field_id < service->num_metadata; field_id++)
+ {
+ struct conf_metadata *ser_md = &service->metadata[field_id];
+ if (ser_md->mergekey != Metadata_mergekey_no)
+ {
+ int r = get_mergekey_from_doc(doc, root, ser_md->name,
+ service, norm_wr);
+ if (r == 0 && ser_md->mergekey == Metadata_mergekey_required)
+ {
+ /* no mergekey on this one and it is required..
+ Generate unique key instead */
+ wrbuf_rewind(norm_wr);
+ break;
+ }
+ }
+ }
+ }
+
+ /* generate unique key if none is not generated already or is empty */
+ if (wrbuf_len(norm_wr) == 0)
+ {
+ wrbuf_printf(norm_wr, "%s-%d",
+ client_get_database(cl)->database->url, record_no);
+ }
+ if (wrbuf_len(norm_wr) > 0)
+ mergekey_norm = nmem_strdup(nmem, wrbuf_cstr(norm_wr));
+ wrbuf_destroy(norm_wr);
+ return mergekey_norm;
+}
+
+/** \brief see if metadata for pz:recordfilter exists
+ \param root xml root element of normalized record
+ \param sdb session database for client
+ \retval 0 if there is no metadata for pz:recordfilter
+ \retval 1 if there is metadata for pz:recordfilter
+
+ If there is no pz:recordfilter defined, this function returns 1
+ as well.
+*/
+
+static int check_record_filter(xmlNode *root, struct session_database *sdb)
+{
+ int match = 0;
+ xmlNode *n;
+ const char *s;
+ s = session_setting_oneval(sdb, PZ_RECORDFILTER);
+
+ if (!s || !*s)
+ return 1;
+
+ for (n = root->children; n; n = n->next)
+ {
+ if (n->type != XML_ELEMENT_NODE)
+ continue;
+ if (!strcmp((const char *) n->name, "metadata"))
+ {
+ xmlChar *type = xmlGetProp(n, (xmlChar *) "type");
+ if (type)
+ {
+ size_t len;
+ const char *eq = strchr(s, '=');
+ if (eq)
+ len = eq - s;
+ else
+ len = strlen(s);
+ if (len == strlen((const char *)type) &&
+ !memcmp((const char *) type, s, len))
+ {
+ xmlChar *value = xmlNodeGetContent(n);
+ if (value && *value)
+ {
+ if (!eq || strstr((const char *) value, eq+1))
+ match = 1;
+ }
+ xmlFree(value);
+ }
+ xmlFree(type);
+ }
+ }
+ }
+ return match;
+}
+
+
+/** \brief ingest XML record
+ \param cl client holds the result set for record
+ \param rec record buffer (0 terminated)
+ \param record_no record position (1, 2, ..)
+ \returns resulting record or NULL on failure
+*/
+struct record *ingest_record(struct client *cl, const char *rec,
int record_no)
{
- xmlDoc *xdoc = normalize_record(client_get_database(cl),
- client_get_session(cl), rec);
+ struct session_database *sdb = client_get_database(cl);
+ struct session *se = client_get_session(cl);
+ xmlDoc *xdoc = normalize_record(sdb, se, rec);
xmlNode *root, *n;
struct record *record;
struct record_cluster *cluster;
- struct session *se = client_get_session(cl);
- xmlChar *mergekey, *mergekey_norm;
+ const char *mergekey_norm;
xmlChar *type = 0;
xmlChar *value = 0;
- struct conf_service *service = global_parameters.server->service;
- const char *norm_str = 0;
- pp2_relevance_token_t prt = 0;
- WRBUF norm_wr = 0;
+ struct conf_service *service = se->service;
if (!xdoc)
return 0;
root = xmlDocGetRootElement(xdoc);
- if (!(mergekey = xmlGetProp(root, (xmlChar *) "mergekey")))
+
+ if (!check_record_filter(root, sdb))
{
- yaz_log(YLOG_WARN, "No mergekey found in record");
+ yaz_log(YLOG_WARN, "Filtered out record no %d from %s", record_no,
+ sdb->database->url);
xmlFreeDoc(xdoc);
return 0;
}
-
- record = record_create(se->nmem,
- service->num_metadata, service->num_sortkeys, cl,
- record_no);
- prt = pp2_relevance_tokenize(
- global_parameters.server->mergekey_pct, (const char *) mergekey);
-
-
- norm_wr = wrbuf_alloc();
-
- while ((norm_str = pp2_relevance_token_next(prt)))
+ mergekey_norm = get_mergekey(xdoc, cl, record_no, service, se->nmem);
+ if (!mergekey_norm)
{
- if (*norm_str)
- {
- if (wrbuf_len(norm_wr))
- wrbuf_puts(norm_wr, " ");
- wrbuf_puts(norm_wr, norm_str);
- }
+ yaz_log(YLOG_WARN, "Got no mergekey");
+ xmlFreeDoc(xdoc);
+ return 0;
}
-
- mergekey_norm = (xmlChar *)nmem_strdup(se->nmem, wrbuf_cstr(norm_wr));
- wrbuf_destroy(norm_wr);
-
- pp2_relevance_token_destroy(prt);
+ record = record_create(se->nmem,
+ service->num_metadata, service->num_sortkeys, cl,
+ record_no);
- xmlFree(mergekey);
-
cluster = reclist_insert(se->reclist,
- global_parameters.server->service,
+ service,
record, (char *) mergekey_norm,
&se->total_merged);
if (global_parameters.dump_records)
yaz_log(YLOG_LOG, "Cluster id %s from %s (#%d)", cluster->recid,
- client_get_database(cl)->database->url, record_no);
+ sdb->database->url, record_no);
if (!cluster)
{
/* no room for record */
return 0;
}
relevance_newrec(se->relevance, cluster);
-
-
- // now parsing XML record and adding data to cluster or record metadata
- for (n = root->children; n; n = n->next)
- {
- if (type)
- xmlFree(type);
- if (value)
- xmlFree(value);
- type = value = 0;
-
- if (n->type != XML_ELEMENT_NODE)
- continue;
- if (!strcmp((const char *) n->name, "metadata"))
- {
- struct conf_metadata *ser_md = 0;
- struct conf_sortkey *ser_sk = 0;
- struct record_metadata **wheretoput = 0;
- struct record_metadata *rec_md = 0;
- int md_field_id = -1;
- int sk_field_id = -1;
-
- type = xmlGetProp(n, (xmlChar *) "type");
- value = xmlNodeListGetString(xdoc, n->children, 1);
-
- if (!type || !value || !*value)
- continue;
-
- md_field_id
- = conf_service_metadata_field_id(service, (const char *) type);
- if (md_field_id < 0)
- {
- yaz_log(YLOG_WARN,
- "Ignoring unknown metadata element: %s", type);
- continue;
- }
-
- ser_md = &service->metadata[md_field_id];
-
- if (ser_md->sortkey_offset >= 0){
- sk_field_id = ser_md->sortkey_offset;
- ser_sk = &service->sortkeys[sk_field_id];
- }
-
- // non-merged metadata
- rec_md = record_metadata_init(se->nmem, (char *) value,
- ser_md->type);
- if (!rec_md)
- {
- yaz_log(YLOG_WARN, "bad metadata data '%s' for element '%s'",
- value, type);
- continue;
- }
- rec_md->next = record->metadata[md_field_id];
- record->metadata[md_field_id] = rec_md;
-
- // merged metadata
- rec_md = record_metadata_init(se->nmem, (char *) value,
- ser_md->type);
- wheretoput = &cluster->metadata[md_field_id];
-
- // and polulate with data:
- // assign cluster or record based on merge action
- if (ser_md->merge == Metadata_merge_unique)
- {
- struct record_metadata *mnode;
- for (mnode = *wheretoput; mnode; mnode = mnode->next)
- if (!strcmp((const char *) mnode->data.text.disp,
- rec_md->data.text.disp))
- break;
- if (!mnode)
- {
- rec_md->next = *wheretoput;
- *wheretoput = rec_md;
- }
- }
- else if (ser_md->merge == Metadata_merge_longest)
- {
- if (!*wheretoput
- || strlen(rec_md->data.text.disp)
- > strlen((*wheretoput)->data.text.disp))
- {
- *wheretoput = rec_md;
- if (ser_sk)
- {
- const char *sort_str = 0;
- int skip_article =
- ser_sk->type == Metadata_sortkey_skiparticle;
-
- if (!cluster->sortkeys[sk_field_id])
- cluster->sortkeys[sk_field_id] =
- nmem_malloc(se->nmem,
- sizeof(union data_types));
+
+ // now parsing XML record and adding data to cluster or record metadata
+ for (n = root->children; n; n = n->next)
+ {
+ pp2_relevance_token_t prt;
+ if (type)
+ xmlFree(type);
+ if (value)
+ xmlFree(value);
+ type = value = 0;
+
+ if (n->type != XML_ELEMENT_NODE)
+ continue;
+ if (!strcmp((const char *) n->name, "metadata"))
+ {
+ struct conf_metadata *ser_md = 0;
+ struct conf_sortkey *ser_sk = 0;
+ struct record_metadata **wheretoput = 0;
+ struct record_metadata *rec_md = 0;
+ int md_field_id = -1;
+ int sk_field_id = -1;
+
+ type = xmlGetProp(n, (xmlChar *) "type");
+ value = xmlNodeListGetString(xdoc, n->children, 1);
+
+ if (!type || !value || !*value)
+ continue;
+
+ md_field_id
+ = conf_service_metadata_field_id(service, (const char *) type);
+ if (md_field_id < 0)
+ {
+ if (se->number_of_warnings_unknown_metadata == 0)
+ {
+ yaz_log(YLOG_WARN,
+ "Ignoring unknown metadata element: %s", type);
+ }
+ se->number_of_warnings_unknown_metadata++;
+ continue;
+ }
+
+ ser_md = &service->metadata[md_field_id];
+
+ if (ser_md->sortkey_offset >= 0){
+ sk_field_id = ser_md->sortkey_offset;
+ ser_sk = &service->sortkeys[sk_field_id];
+ }
+
+ // non-merged metadata
+ rec_md = record_metadata_init(se->nmem, (char *) value,
+ ser_md->type);
+ if (!rec_md)
+ {
+ yaz_log(YLOG_WARN, "bad metadata data '%s' for element '%s'",
+ value, type);
+ continue;
+ }
+ wheretoput = &record->metadata[md_field_id];
+ while (*wheretoput)
+ wheretoput = &(*wheretoput)->next;
+ *wheretoput = rec_md;
+
+ // merged metadata
+ rec_md = record_metadata_init(se->nmem, (char *) value,
+ ser_md->type);
+ wheretoput = &cluster->metadata[md_field_id];
+
+ // and polulate with data:
+ // assign cluster or record based on merge action
+ if (ser_md->merge == Metadata_merge_unique)
+ {
+ struct record_metadata *mnode;
+ for (mnode = *wheretoput; mnode; mnode = mnode->next)
+ if (!strcmp((const char *) mnode->data.text.disp,
+ rec_md->data.text.disp))
+ break;
+ if (!mnode)
+ {
+ rec_md->next = *wheretoput;
+ *wheretoput = rec_md;
+ }
+ }
+ else if (ser_md->merge == Metadata_merge_longest)
+ {
+ if (!*wheretoput
+ || strlen(rec_md->data.text.disp)
+ > strlen((*wheretoput)->data.text.disp))
+ {
+ *wheretoput = rec_md;
+ if (ser_sk)
+ {
+ const char *sort_str = 0;
+ int skip_article =
+ ser_sk->type == Metadata_sortkey_skiparticle;
+
+ if (!cluster->sortkeys[sk_field_id])
+ cluster->sortkeys[sk_field_id] =
+ nmem_malloc(se->nmem,
+ sizeof(union data_types));
- prt = pp2_relevance_tokenize(
- global_parameters.server->sort_pct,
- rec_md->data.text.disp);
+ prt = pp2_relevance_tokenize(
+ service->sort_pct,
+ rec_md->data.text.disp);
- pp2_relevance_token_next(prt);
+ pp2_relevance_token_next(prt);
- sort_str = pp2_get_sort(prt, skip_article);
+ sort_str = pp2_get_sort(prt, skip_article);
- cluster->sortkeys[sk_field_id]->text.disp =
- rec_md->data.text.disp;
- if (!sort_str)
- {
- sort_str = rec_md->data.text.disp;
- yaz_log(YLOG_WARN,
- "Could not make sortkey. Bug #1858");
- }
- cluster->sortkeys[sk_field_id]->text.sort =
- nmem_strdup(se->nmem, sort_str);
+ cluster->sortkeys[sk_field_id]->text.disp =
+ rec_md->data.text.disp;
+ if (!sort_str)
+ {
+ sort_str = rec_md->data.text.disp;
+ yaz_log(YLOG_WARN,
+ "Could not make sortkey. Bug #1858");
+ }
+ cluster->sortkeys[sk_field_id]->text.sort =
+ nmem_strdup(se->nmem, sort_str);
#if 0
- yaz_log(YLOG_LOG, "text disp=%s",
- cluster->sortkeys[sk_field_id]->text.disp);
- yaz_log(YLOG_LOG, "text sort=%s",
- cluster->sortkeys[sk_field_id]->text.sort);
+ yaz_log(YLOG_LOG, "text disp=%s",
+ cluster->sortkeys[sk_field_id]->text.disp);
+ yaz_log(YLOG_LOG, "text sort=%s",
+ cluster->sortkeys[sk_field_id]->text.sort);
#endif
- pp2_relevance_token_destroy(prt);
+ pp2_relevance_token_destroy(prt);
}
}
}
type = value = 0;
}
else
- yaz_log(YLOG_WARN,
- "Unexpected element %s in internal record", n->name);
+ {
+ if (se->number_of_warnings_unknown_elements == 0)
+ yaz_log(YLOG_WARN,
+ "Unexpected element in internal record: %s", n->name);
+ se->number_of_warnings_unknown_elements++;
+ }
}
if (type)
xmlFree(type);
/*
* Local variables:
* c-basic-offset: 4
+ * c-file-style: "Stroustrup"
* indent-tabs-mode: nil
* End:
* vim: shiftwidth=4 tabstop=8 expandtab
*/
+
projects / pazpar2-moved-to-github.git / blobdiff