Comma separated value for pz:recordfilter.
[pazpar2-moved-to-github.git] / src / logic.c
index fabf8e6..c76ba28 100644 (file)
@@ -57,7 +57,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <yaz/timing.h>
 #endif
 
-
+#include "parameters.h"
 #include "pazpar2.h"
 #include "eventl.h"
 #include "http.h"
@@ -68,6 +68,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include "client.h"
 #include "settings.h"
 #include "normalize7bit.h"
+#include "marcmap.h"
 
 #define TERMLIST_HIGH_SCORE 25
 
@@ -76,23 +77,9 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 // 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)
@@ -271,8 +258,17 @@ xmlDoc *normalize_record(struct session_database *sdb, struct session *se,
 
                 insert_settings_parameters(sdb, se, parms);
 
-                new = xsltApplyStylesheet(m->stylesheet, rdoc, (const char **) parms);
-                root= xmlDocGetRootElement(new);
+                if (m->stylesheet)
+                {
+                    new = xsltApplyStylesheet(m->stylesheet, rdoc, (const char **) parms);
+                }
+                else if (m->marcmap)
+                {
+                    new = marcmap_apply(m->marcmap, rdoc);
+                }
+
+                root = xmlDocGetRootElement(new);
+
                 if (!new || !root || !(root->children))
                 {
                     yaz_log(YLOG_WARN, "XSLT transformation failed from %s",
@@ -356,12 +352,30 @@ static int prepare_map(struct session *se, struct session_database *sdb)
         {
             (*m) = nmem_malloc(se->session_nmem, sizeof(**m));
             (*m)->next = 0;
-            if (!((*m)->stylesheet = conf_load_stylesheet(stylesheets[i])))
+            // XSLT
+            if (!strcmp(&stylesheets[i][strlen(stylesheets[i])-4], ".xsl")) 
+            {    
+                (*m)->marcmap = NULL;
+                if (!((*m)->stylesheet = conf_load_stylesheet(se->service, stylesheets[i])))
+                {
+                    yaz_log(YLOG_FATAL|YLOG_ERRNO, "Unable to load stylesheet: %s",
+                            stylesheets[i]);
+                    return -1;
+                }
+            }
+            // marcmap
+            else if (!strcmp(&stylesheets[i][strlen(stylesheets[i])-5], ".mmap"))
             {
-                yaz_log(YLOG_FATAL|YLOG_ERRNO, "Unable to load stylesheet: %s",
-                        stylesheets[i]);
-                return -1;
+                (*m)->stylesheet = NULL;
+               if (!((*m)->marcmap = marcmap_load(stylesheets[i], se->session_nmem)))
+                {
+                    yaz_log(YLOG_FATAL|YLOG_ERRNO, "Unable to load marcmap: %s",
+                            stylesheets[i]);
+                    return -1;
+                }
             }
+
             m = &(*m)->next;
         }
     }
@@ -511,7 +525,7 @@ static struct database_criterion *parse_filter(NMEM m, const char *buf)
 }
 
 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;
@@ -552,7 +566,8 @@ enum pazpar2_error_code search(struct session *se,
         else
         {
             no_working++;
-            if (client_prep_connection(cl))
+            if (client_prep_connection(cl, se->service->z3950_connect_timeout,
+                    se->service->z3950_session_timeout))
                 client_start_search(cl);
         }
     }
@@ -615,6 +630,8 @@ static struct session_database *load_session_database(struct session *se,
 {
     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;
@@ -685,6 +702,7 @@ void destroy_session(struct session *s)
     for (sdb = s->databases; sdb; sdb = sdb->next)
         session_database_destroy(sdb);
     nmem_destroy(s->nmem);
+    service_destroy(s->service);
     wrbuf_destroy(s->wrbuf);
 }
 
@@ -877,52 +895,10 @@ void statistics(struct session *se, struct statistics *stat)
     stat->num_clients = count;
 }
 
-int 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);
-        }
-    }
-    return 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;
@@ -964,8 +940,8 @@ static struct record_metadata *record_metadata_init(
     return rec_md;
 }
 
-const char *get_mergekey(xmlDoc *doc, struct client *cl, int record_no,
-                         struct conf_service *service, NMEM nmem)
+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);
@@ -979,7 +955,7 @@ const char *get_mergekey(xmlDoc *doc, struct client *cl, int record_no,
         const char *norm_str;
         pp2_relevance_token_t prt =
             pp2_relevance_tokenize(
-                global_parameters.server->mergekey_pct,
+                service->mergekey_pct,
                 (const char *) mergekey);
         
         while ((norm_str = pp2_relevance_token_next(prt)))
@@ -1023,7 +999,7 @@ const char *get_mergekey(xmlDoc *doc, struct client *cl, int record_no,
                         const char *norm_str;
                         pp2_relevance_token_t prt =
                             pp2_relevance_tokenize(
-                                global_parameters.server->mergekey_pct,
+                                service->mergekey_pct,
                                 (const char *) value);
                         
                         while ((norm_str = pp2_relevance_token_next(prt)))
@@ -1056,6 +1032,71 @@ const char *get_mergekey(xmlDoc *doc, struct client *cl, int record_no,
     return mergekey_norm;
 }
 
+static const char *str_tok_n(const char *s, const char *delim,
+                             const char **res, size_t *len)
+{
+    *res = s;
+    while (*s && !strchr(delim, *s))
+        s++;
+    *len = s - *res;
+    if (*len == 0)
+        return 0;
+    if (*s && strchr(delim, *s))
+        s++;
+    return s;
+}
+
+/** \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)
+            {
+                const char *s1 = s;
+                size_t len;
+                const char *value;
+                while ((s1 = str_tok_n(s1, ",", &value, &len)) != 0)
+                {
+                    if (len == strlen((const char *)type) &&
+                        !memcmp((const char *) type, s, len))
+                    {
+                        xmlChar *value = xmlNodeGetContent(n);
+                        if (value && *value)
+                        {
+                            xmlFree(value);
+                            match = 1;
+                        }
+                    }
+                }
+                xmlFree(type);
+            }
+        }
+    }
+    return match;
+}
 
 
 /** \brief ingest XML record
@@ -1067,12 +1108,12 @@ const char *get_mergekey(xmlDoc *doc, struct client *cl, int record_no,
 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;
@@ -1083,6 +1124,14 @@ struct record *ingest_record(struct client *cl, const char *rec,
 
     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)
     {
@@ -1100,7 +1149,7 @@ struct record *ingest_record(struct client *cl, const char *rec,
                              &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 */
@@ -1209,7 +1258,7 @@ struct record *ingest_record(struct client *cl, const char *rec,
                                             sizeof(union data_types));
                          
                         prt = pp2_relevance_tokenize(
-                            global_parameters.server->sort_pct,
+                            service->sort_pct,
                             rec_md->data.text.disp);
 
                         pp2_relevance_token_next(prt);