#include <yaz/timing.h>
#endif
-
+#include "parameters.h"
#include "pazpar2.h"
#include "eventl.h"
#include "http.h"
struct parameters global_parameters =
{
0, // dump_records
- 0, // debug_mode
- 30, // operations timeout
- 60, // session timeout
- 100,
- 180, // Z39.50 session timeout
- 15 // Connect timeout
+ 0 // debug_mode
};
static void log_xml_doc(xmlDoc *doc)
s->termlists[i].name = nmem_strdup(s->nmem, type);
s->termlists[i].termlist
- = termlist_create(s->nmem, s->expected_maxrecs,
- TERMLIST_HIGH_SCORE);
+ = termlist_create(s->nmem, TERMLIST_HIGH_SCORE);
s->num_termlists = i + 1;
}
termlist_insert(s->termlists[i].termlist, value);
}
-xmlDoc *record_to_xml(struct session_database *sdb, const char *rec)
+static xmlDoc *record_to_xml(struct session_database *sdb, const char *rec)
{
struct database *db = sdb->database;
xmlDoc *rdoc = 0;
}
}
-xmlDoc *normalize_record(struct session_database *sdb, struct session *se,
- const char *rec)
+static xmlDoc *normalize_record(struct session_database *sdb,
+ struct session *se,
+ 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))
+ {
+ yaz_log(YLOG_WARN, "Normalize failed from %s", sdb->database->url);
+ }
+ else
{
- xmlDoc *new = 0;
+ insert_settings_values(sdb, rdoc, se->service);
+ if (global_parameters.dump_records)
{
- 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;
- }
+ yaz_log(YLOG_LOG, "Normalized record from %s",
+ sdb->database->url);
+ log_xml_doc(rdoc);
}
-
- xmlFreeDoc(rdoc);
- rdoc = new;
- }
-
- insert_settings_values(sdb, rdoc, se->service);
-
- if (global_parameters.dump_records)
- {
- yaz_log(YLOG_LOG, "Normalized record from %s",
- sdb->database->url);
- log_xml_doc(rdoc);
}
}
return rdoc;
}
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(se->service->config,
- stylesheets[i])))
- {
- yaz_log(YLOG_FATAL|YLOG_ERRNO, "Unable to load stylesheet: %s",
- stylesheets[i]);
- return -1;
- }
- m = &(*m)->next;
- }
+ sdb->map = normalize_cache_get(se->normalize_cache,
+ se->service, s);
+ if (!sdb->map)
+ return -1;
}
- if (!sdb->map)
- yaz_log(YLOG_WARN, "No Normalization stylesheet for target %s",
- sdb->database->url);
return 0;
}
int subnum;
int subi;
struct database_criterion *new = nmem_malloc(m, sizeof(*new));
- char *eq = strchr(values[i], '=');
- if (!eq)
+ char *eq;
+ if ((eq = strchr(values[i], '=')))
+ new->type = PAZPAR2_STRING_MATCH;
+ else if ((eq = strchr(values[i], '~')))
+ new->type = PAZPAR2_SUBSTRING_MATCH;
+ else
{
- yaz_log(YLOG_WARN, "Missing equal-sign in filter");
+ yaz_log(YLOG_WARN, "Missing equal-sign/tilde in filter");
return 0;
}
*(eq++) = '\0';
}
enum pazpar2_error_code search(struct session *se,
- char *query, char *filter,
+ const char *query,
+ const char *maxrecs,
+ const char *filter,
const char **addinfo)
{
int live_channels = 0;
live_channels = select_targets(se, criteria);
if (live_channels)
{
- int maxrecs = live_channels * global_parameters.toget; // This is buggy!!!
- se->reclist = reclist_create(se->nmem, maxrecs);
- se->expected_maxrecs = maxrecs;
+ se->reclist = reclist_create(se->nmem);
}
else
return PAZPAR2_NO_TARGETS;
for (cl = se->clients; cl; cl = client_next_in_session(cl))
{
+ if (maxrecs)
+ client_set_maxrecs(cl, atoi(maxrecs));
if (prepare_session_database(se, client_get_database(cl)) < 0)
- {
- *addinfo = client_get_database(cl)->database->url;
- return PAZPAR2_CONFIG_TARGET;
- }
+ continue;
// Parse query for target
if (client_parse_query(cl, query) < 0)
no_failed++;
else
{
no_working++;
- if (client_prep_connection(cl))
+ if (client_prep_connection(cl, se->service->z3950_operation_timeout,
+ se->service->z3950_session_timeout))
client_start_search(cl);
}
}
-
- // If no queries could be mapped, we signal an error
if (no_working == 0)
{
- *addinfo = "query";
- return PAZPAR2_MALFORMED_PARAMETER_VALUE;
+ if (no_failed > 0)
+ {
+ *addinfo = "query";
+ return PAZPAR2_MALFORMED_PARAMETER_VALUE;
+ }
+ else
+ return PAZPAR2_NO_TARGETS;
}
return PAZPAR2_NO_ERROR;
}
// 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);
+ sdb->map = 0;
}
// Initialize session_database list -- this represents this session's view
{
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
return se->databases;
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);
sdb->map = 0;
}
break;
client_destroy(s->clients);
for (sdb = s->databases; sdb; sdb = sdb->next)
session_database_destroy(sdb);
+ normalize_cache_destroy(s->normalize_cache);
nmem_destroy(s->nmem);
+ service_destroy(s->service);
wrbuf_destroy(s->wrbuf);
}
session->num_termlists = 0;
session->reclist = 0;
session->clients = 0;
- session->expected_maxrecs = 0;
session->session_nmem = nmem;
session->nmem = nmem_create();
session->wrbuf = wrbuf_alloc();
session->watchlist[i].data = 0;
session->watchlist[i].fun = 0;
}
+ session->normalize_cache = normalize_cache_create();
+
return session;
}
}
reclist_sort(s->reclist, sp);
- *total = s->reclist->num_records;
+ *total = reclist_get_num_records(s->reclist);
*sumhits = s->total_hits;
for (i = 0; i < start; i++)
// 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;
}
+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();
- xmlNode *n;
- /* create mergekey based on mergekey attribute from XSL (if any) */
+ /* consider mergekey from XSL first */
xmlChar *mergekey = xmlGetProp(root, (xmlChar *) "mergekey");
if (mergekey)
{
pp2_relevance_token_destroy(prt);
xmlFree(mergekey);
}
- /* append (if any) mergekey=yes metadata values */
- for (n = root->children; n; n = n->next)
+ else
{
- if (n->type != XML_ELEMENT_NODE)
- continue;
- if (!strcmp((const char *) n->name, "metadata"))
+ /* 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 = 0;
- int md_field_id = -1;
-
- xmlChar *type = xmlGetProp(n, (xmlChar *) "type");
-
- if (!type)
- continue;
-
- md_field_id
- = conf_service_metadata_field_id(service,
- (const char *) type);
- if (md_field_id >= 0)
+ struct conf_metadata *ser_md = &service->metadata[field_id];
+ if (ser_md->mergekey != Metadata_mergekey_no)
{
- ser_md = &service->metadata[md_field_id];
- if (ser_md->mergekey == Metadata_mergekey_yes)
+ int r = get_mergekey_from_doc(doc, root, ser_md->name,
+ service, norm_wr);
+ if (r == 0 && ser_md->mergekey == Metadata_mergekey_required)
{
- 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);
-
- 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 mergekey on this one and it is required..
+ Generate unique key instead */
+ wrbuf_rewind(norm_wr);
+ break;
}
}
- xmlFree(type);
}
}
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
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);
const char *mergekey_norm;
xmlChar *type = 0;
xmlChar *value = 0;
root = xmlDocGetRootElement(xdoc);
+ if (!check_record_filter(root, sdb))
+ {
+ yaz_log(YLOG_WARN, "Filtered out record no %d from %s", record_no,
+ sdb->database->url);
+ xmlFreeDoc(xdoc);
+ return 0;
+ }
+
mergekey_norm = get_mergekey(xdoc, cl, record_no, service, se->nmem);
if (!mergekey_norm)
{
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);
if (!cluster)
{
/* no room for record */
xmlFreeDoc(xdoc);
return 0;
}
+ if (global_parameters.dump_records)
+ yaz_log(YLOG_LOG, "Cluster id %s from %s (#%d)", cluster->recid,
+ sdb->database->url, record_no);
relevance_newrec(se->relevance, cluster);
// now parsing XML record and adding data to cluster or record metadata
projects / pazpar2-moved-to-github.git / blobdiff