zoom: Use udb and query only one searchable
[metaproxy-moved-to-github.git] / src / filter_zoom.cpp
index bce817c..129b027 100644 (file)
@@ -19,6 +19,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include "config.hpp"
 #include "filter_zoom.hpp"
 #include <yaz/zoom.h>
+#include <yaz/srw.h>
 #include <metaproxy/package.hpp>
 #include <metaproxy/util.hpp>
 #include "torus.hpp"
@@ -29,6 +30,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <boost/thread/mutex.hpp>
 #include <boost/thread/condition.hpp>
 #include <yaz/ccl.h>
+#include <yaz/cql.h>
 #include <yaz/oid_db.h>
 #include <yaz/diagbib1.h>
 #include <yaz/log.h>
@@ -41,6 +43,10 @@ namespace yf = mp::filter;
 namespace metaproxy_1 {
     namespace filter {
         struct Zoom::Searchable : boost::noncopyable {
+            std::string authentication;
+            std::string cfAuth;
+            std::string cfProxy;
+            std::string cfSubDb;
             std::string database;
             std::string target;
             std::string query_encoding;
@@ -110,14 +116,12 @@ namespace metaproxy_1 {
         private:
             FrontendPtr get_frontend(mp::Package &package);
             void release_frontend(mp::Package &package);
-            void parse_torus(const xmlNode *ptr);
-
-            std::list<Zoom::SearchablePtr>m_searchables;
+            SearchablePtr parse_torus(const xmlNode *ptr);
 
             std::map<mp::Session, FrontendPtr> m_clients;            
             boost::mutex m_mutex;
             boost::condition m_cond_session_ready;
-            mp::Torus torus;
+            std::string torus_url;
         };
     }
 }
@@ -201,7 +205,7 @@ int yf::Zoom::Backend::get_error(const char **addinfo)
 yf::Zoom::Searchable::Searchable()
 {
     piggyback = true;
-    use_turbomarc = false;
+    use_turbomarc = true;
     ccl_bibset = ccl_qual_mk();
 }
 
@@ -272,10 +276,11 @@ yf::Zoom::Impl::~Impl()
 { 
 }
 
-void yf::Zoom::Impl::parse_torus(const xmlNode *ptr1)
+yf::Zoom::SearchablePtr yf::Zoom::Impl::parse_torus(const xmlNode *ptr1)
 {
+    SearchablePtr notfound;
     if (!ptr1)
-        return ;
+        return notfound;
     for (ptr1 = ptr1->children; ptr1; ptr1 = ptr1->next)
     {
         if (ptr1->type != XML_ELEMENT_NODE)
@@ -296,7 +301,27 @@ void yf::Zoom::Impl::parse_torus(const xmlNode *ptr1)
                     {
                         if (ptr3->type != XML_ELEMENT_NODE)
                             continue;
-                        if (!strcmp((const char *) ptr3->name, "id"))
+                        if (!strcmp((const char *) ptr3->name,
+                                    "authentication"))
+                        {
+                            s->authentication = mp::xml::get_text(ptr3);
+                        }
+                        else if (!strcmp((const char *) ptr3->name,
+                                    "cfAuth"))
+                        {
+                            s->cfAuth = mp::xml::get_text(ptr3);
+                        } 
+                        else if (!strcmp((const char *) ptr3->name,
+                                    "cfProxy"))
+                        {
+                            s->cfProxy = mp::xml::get_text(ptr3);
+                        }  
+                        else if (!strcmp((const char *) ptr3->name,
+                                    "cfSubDb"))
+                        {
+                            s->cfSubDb = mp::xml::get_text(ptr3);
+                        }  
+                        else if (!strcmp((const char *) ptr3->name, "id"))
                         {
                             s->database = mp::xml::get_text(ptr3);
                         }
@@ -341,7 +366,7 @@ void yf::Zoom::Impl::parse_torus(const xmlNode *ptr1)
                         else if (!strcmp((const char *) ptr3->name,
                                          "useTurboMarc"))
                         {
-                            s->use_turbomarc = mp::xml::get_bool(ptr3, false);
+                            ; // useTurboMarc is ignored
                         }
                         else if (!strncmp((const char *) ptr3->name,
                                           "cclmap_", 7))
@@ -353,15 +378,15 @@ void yf::Zoom::Impl::parse_torus(const xmlNode *ptr1)
                     }
                     if (s->database.length() && s->target.length())
                     {
-                        yaz_log(YLOG_LOG, "add db=%s target=%s turbomarc=%s", 
-                                s->database.c_str(), s->target.c_str(),
-                                s->use_turbomarc ? "1" : "0");
-                        m_searchables.push_back(s);
+                        yaz_log(YLOG_LOG, "add db=%s target=%s", 
+                                s->database.c_str(), s->target.c_str());
                     }
+                    return s;
                 }
             }
         }
     }
+    return notfound;
 }
 
 void yf::Zoom::Impl::configure(const xmlNode *ptr, bool test_only)
@@ -370,10 +395,6 @@ void yf::Zoom::Impl::configure(const xmlNode *ptr, bool test_only)
     {
         if (ptr->type != XML_ELEMENT_NODE)
             continue;
-        if (!strcmp((const char *) ptr->name, "records"))
-        {
-            parse_torus(ptr);
-        }
         else if (!strcmp((const char *) ptr->name, "torus"))
         {
             std::string url;
@@ -387,13 +408,11 @@ void yf::Zoom::Impl::configure(const xmlNode *ptr, bool test_only)
                         "Bad attribute " + std::string((const char *)
                                                        attr->name));
             }
-            torus.read_searchables(url);
-            xmlDoc *doc = torus.get_doc();
-            if (doc)
-            {
-                xmlNode *ptr = xmlDocGetRootElement(doc);
-                parse_torus(ptr);
-            }
+            torus_url = url;
+        }
+        else if (!strcmp((const char *) ptr->name, "records"))
+        {
+            yaz_log(YLOG_WARN, "records ignored!");
         }
         else
         {
@@ -412,29 +431,28 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
     if (m_backend && m_backend->m_frontend_database == database)
         return m_backend;
 
-    std::list<Zoom::SearchablePtr>::iterator map_s =
-        m_p->m_searchables.begin();
-
-    std::string c_db = mp::util::database_name_normalize(database);
-
-    while (map_s != m_p->m_searchables.end())
+    xmlDoc *doc = mp::get_searchable(m_p->torus_url, database);
+    if (!doc)
     {
-        if (c_db.compare((*map_s)->database) == 0)
-            break;
-        map_s++;
+        *error = YAZ_BIB1_DATABASE_DOES_NOT_EXIST;
+        *addinfo = database.c_str();
+        BackendPtr b;
+        return b;
     }
-    if (map_s == m_p->m_searchables.end())
+    SearchablePtr sptr = m_p->parse_torus(xmlDocGetRootElement(doc));
+    xmlFreeDoc(doc);
+    if (!sptr)
     {
         *error = YAZ_BIB1_DATABASE_DOES_NOT_EXIST;
         *addinfo = database.c_str();
         BackendPtr b;
         return b;
     }
-
+        
     xsltStylesheetPtr xsp = 0;
-    if ((*map_s)->transform_xsl_fname.length())
+    if (sptr->transform_xsl_fname.length())
     {
-        xmlDoc *xsp_doc = xmlParseFile((*map_s)->transform_xsl_fname.c_str());
+        xmlDoc *xsp_doc = xmlParseFile(sptr->transform_xsl_fname.c_str());
         if (!xsp_doc)
         {
             *error = YAZ_BIB1_TEMPORARY_SYSTEM_ERROR;
@@ -453,18 +471,49 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
         }
     }
 
-    SearchablePtr sptr = *map_s;
-
     m_backend.reset();
 
     BackendPtr b(new Backend(sptr));
 
+    std::string cf_parm;
     b->xsp = xsp;
     b->m_frontend_database = database;
+    std::string authentication = sptr->authentication;
 
     if (sptr->query_encoding.length())
         b->set_option("rpnCharset", sptr->query_encoding.c_str());
 
+    if (sptr->cfAuth.length())
+    {
+        b->set_option("user", sptr->cfAuth.c_str());
+        if (authentication.length())
+        {
+            size_t found = authentication.find('/');
+            if (found != std::string::npos)
+            {
+                cf_parm += "user=" + mp::util::uri_encode(authentication.substr(0, found))
+                    + "&password=" + mp::util::uri_encode(authentication.substr(found+1));
+            }
+            else
+                cf_parm += "user=" + mp::util::uri_encode(authentication);
+        }
+    }
+    else if (authentication.length())
+        b->set_option("user", authentication.c_str());
+
+    if (sptr->cfProxy.length())
+    {
+        if (cf_parm.length())
+            cf_parm += "&";
+        cf_parm += "proxy=" + mp::util::uri_encode(sptr->cfProxy);
+    }
+    if (sptr->cfSubDb.length())
+    {
+        if (cf_parm.length())
+            cf_parm += "&";
+        cf_parm += "subdatabase=" + mp::util::uri_encode(sptr->cfSubDb);
+    }
+
     std::string url;
     if (sptr->sru.length())
     {
@@ -472,8 +521,13 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
         b->set_option("sru", sptr->sru.c_str());
     }
     else
+    {
         url = sptr->target;
-
+    }
+    if (cf_parm.length())
+    {
+        url += "," + cf_parm;
+    }
     b->connect(url, error, addinfo);
     if (*error == 0)
     {
@@ -511,7 +565,8 @@ Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
     if (preferredRecordSyntax)
     {
         if (!oid_oidcmp(preferredRecordSyntax, yaz_oid_recsyn_xml)
-            && !strcmp(element_set_name, "pz2"))
+            && element_set_name &&
+            !strcmp(element_set_name, "pz2"))
         {
             if (b->sptr->request_syntax.length())
             {
@@ -678,45 +733,59 @@ void yf::Zoom::Frontend::handle_search(mp::Package &package)
 
     Odr_int hits = 0;
     Z_Query *query = sr->query;
+    WRBUF ccl_wrbuf = 0;
+    WRBUF pqf_wrbuf = 0;
+
     if (query->which == Z_Query_type_1 || query->which == Z_Query_type_101)
     {
         // RPN
-        WRBUF w = wrbuf_alloc();
-        yaz_rpnquery_to_wrbuf(w, query->u.type_1);
-
-        b->search_pqf(wrbuf_cstr(w), &hits, &error, &addinfo);
-
-        wrbuf_destroy(w);
+        pqf_wrbuf = wrbuf_alloc();
+        yaz_rpnquery_to_wrbuf(pqf_wrbuf, query->u.type_1);
     }
     else if (query->which == Z_Query_type_2)
     {
         // CCL
-        WRBUF w = wrbuf_alloc();
-        wrbuf_write(w, (const char *) query->u.type_2->buf,
+        ccl_wrbuf = wrbuf_alloc();
+        wrbuf_write(ccl_wrbuf, (const char *) query->u.type_2->buf,
                     query->u.type_2->len);
-        int cerror, cpos;
-        struct ccl_rpn_node *cn;
-        cn = ccl_find_str(b->sptr->ccl_bibset, wrbuf_cstr(w), &cerror, &cpos);
-        wrbuf_destroy(w);
-
-        if (!cn)
+    }
+    else if (query->which == Z_Query_type_104 &&
+             query->u.type_104->which == Z_External_CQL)
+    {
+        // CQL
+        const char *cql = query->u.type_104->u.cql;
+        CQL_parser cp = cql_parser_create();
+        int r = cql_parser_string(cp, cql);
+        if (r)
         {
-            char *addinfo = odr_strdup(odr, ccl_err_msg(cerror));
+            cql_parser_destroy(cp);
+            apdu_res = 
+                odr.create_searchResponse(apdu_req, 
+                                          YAZ_BIB1_MALFORMED_QUERY,
+                                          "CQL syntax error");
+            package.response() = apdu_res;
+            return;
+        }
+        struct cql_node *cn = cql_parser_result(cp);
+        char ccl_buf[1024];
 
+        r = cql_to_ccl_buf(cn, ccl_buf, sizeof(ccl_buf));
+        yaz_log(YLOG_LOG, "cql_to_ccl_buf returned %d", r);
+        if (r == 0)
+        {
+            ccl_wrbuf = wrbuf_alloc();
+            wrbuf_puts(ccl_wrbuf, ccl_buf);
+        }
+        cql_parser_destroy(cp);
+        if (r)
+        {
             apdu_res = 
                 odr.create_searchResponse(apdu_req, 
                                           YAZ_BIB1_MALFORMED_QUERY,
-                                          addinfo);
+                                          "CQL to CCL conversion error");
             package.response() = apdu_res;
             return;
         }
-        w = wrbuf_alloc();
-        ccl_pquery(w, cn);
-        
-        b->search_pqf(wrbuf_cstr(w), &hits, &error, &addinfo);
-        
-        ccl_rpn_delete(cn);
-        wrbuf_destroy(w);
     }
     else
     {
@@ -725,6 +794,36 @@ void yf::Zoom::Frontend::handle_search(mp::Package &package)
         package.response() = apdu_res;
         return;
     }
+
+    if (ccl_wrbuf)
+    {
+        // CCL to PQF
+        assert(pqf_wrbuf == 0);
+        int cerror, cpos;
+        struct ccl_rpn_node *cn;
+        cn = ccl_find_str(b->sptr->ccl_bibset, wrbuf_cstr(ccl_wrbuf),
+                          &cerror, &cpos);
+        wrbuf_destroy(ccl_wrbuf);
+        if (!cn)
+        {
+            char *addinfo = odr_strdup(odr, ccl_err_msg(cerror));
+
+            apdu_res = 
+                odr.create_searchResponse(apdu_req, 
+                                          YAZ_BIB1_MALFORMED_QUERY,
+                                          addinfo);
+            package.response() = apdu_res;
+            return;
+        }
+        pqf_wrbuf = wrbuf_alloc();
+        ccl_pquery(pqf_wrbuf, cn);
+        ccl_rpn_delete(cn);
+    }
+    
+    assert(pqf_wrbuf);
+    b->search_pqf(wrbuf_cstr(pqf_wrbuf), &hits, &error, &addinfo);
+    
+    wrbuf_destroy(pqf_wrbuf);
     
     const char *element_set_name = 0;
     Odr_int number_to_present = 0;