Adding '(backend)' to addinfo for backend diagnostics
[metaproxy-moved-to-github.git] / src / filter_zoom.cpp
index d381d43..b3ddddb 100644 (file)
@@ -23,6 +23,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <yaz/srw.h>
 #include <metaproxy/package.hpp>
 #include <metaproxy/util.hpp>
+#include <metaproxy/xmlutil.hpp>
 #include "torus.hpp"
 
 #include <libxslt/xsltutils.h>
@@ -60,6 +61,7 @@ namespace metaproxy_1 {
             std::string element_set;
             std::string record_encoding;
             std::string transform_xsl_fname;
+            std::string urlRecipe;
             bool use_turbomarc;
             bool piggyback;
             CCL_bibset ccl_bibset;
@@ -78,16 +80,17 @@ namespace metaproxy_1 {
         public:
             Backend(SearchablePtr sptr);
             ~Backend();
-            void connect(std::string zurl, int *error, const char **addinfo);
+            void connect(std::string zurl, int *error, char **addinfo,
+                         ODR odr);
             void search_pqf(const char *pqf, Odr_int *hits,
-                            int *error, const char **addinfo);
+                            int *error, char **addinfo, ODR odr);
             void search_cql(const char *cql, Odr_int *hits,
-                            int *error, const char **addinfo);
+                            int *error, char **addinfo, ODR odr);
             void present(Odr_int start, Odr_int number, ZOOM_record *recs,
-                         int *error, const char **addinfo);
+                         int *error, char **addinfo, ODR odr);
             void set_option(const char *name, const char *value);
             const char *get_option(const char *name);
-            void get_zoom_error(int *error, const char **addinfo);
+            void get_zoom_error(int *error, char **addinfo, ODR odr);
         };
         class Zoom::Frontend : boost::noncopyable {
             friend class Impl;
@@ -106,7 +109,7 @@ namespace metaproxy_1 {
             Z_Records *get_records(Odr_int start,
                                    Odr_int number_to_present,
                                    int *error,
-                                   const char **addinfo,
+                                   char **addinfo,
                                    Odr_int *number_of_records_returned,
                                    ODR odr, BackendPtr b,
                                    Odr_oid *preferredRecordSyntax,
@@ -121,7 +124,8 @@ namespace metaproxy_1 {
             Impl();
             ~Impl();
             void process(metaproxy_1::Package & package);
-            void configure(const xmlNode * ptr, bool test_only);
+            void configure(const xmlNode * ptr, bool test_only,
+                           const char *path);
         private:
             void configure_local_records(const xmlNode * ptr, bool test_only);
             FrontendPtr get_frontend(mp::Package &package);
@@ -134,6 +138,7 @@ namespace metaproxy_1 {
             std::string torus_url;
             std::map<std::string,std::string> fieldmap;
             std::string xsldir;
+            std::string file_path;
             CCL_bibset bibset;
             std::string element_transform;
             std::string element_raw;
@@ -152,9 +157,10 @@ yf::Zoom::~Zoom()
 {  // must have a destructor because of boost::scoped_ptr
 }
 
-void yf::Zoom::configure(const xmlNode *xmlnode, bool test_only)
+void yf::Zoom::configure(const xmlNode *xmlnode, bool test_only,
+                         const char *path)
 {
-    m_p->configure(xmlnode, test_only);
+    m_p->configure(xmlnode, test_only, path);
 }
 
 void yf::Zoom::process(mp::Package &package) const
@@ -181,34 +187,57 @@ yf::Zoom::Backend::~Backend()
 }
 
 
-void yf::Zoom::Backend::get_zoom_error(int *error, const char **addinfo)
+void yf::Zoom::Backend::get_zoom_error(int *error, char **addinfo,
+                                       ODR odr)
 {
     const char *msg = 0;
-    *error = ZOOM_connection_error(m_connection, &msg, addinfo);
+    const char *zoom_addinfo = 0;
+    *error = ZOOM_connection_error(m_connection, &msg, &zoom_addinfo);
     if (*error)
     {
         if (*error >= ZOOM_ERROR_CONNECT)
         {
             // turn ZOOM diagnostic into a Bib-1 2: with addinfo=zoom err msg
             *error = YAZ_BIB1_TEMPORARY_SYSTEM_ERROR;
-            if (addinfo)
-                *addinfo = msg;
+            *addinfo = (char *) odr_malloc(
+                odr, 20 + strlen(msg) + 
+                (zoom_addinfo ? strlen(zoom_addinfo) : 0));
+            strcpy(*addinfo, msg);
+            if (zoom_addinfo)
+            {
+                strcat(*addinfo, ": ");
+                strcat(*addinfo, zoom_addinfo);
+                strcat(*addinfo, " ");
+            }
+        }
+        else
+        {
+            *addinfo = (char *) odr_malloc(
+                odr, 20 + (zoom_addinfo ? strlen(zoom_addinfo) : 0));
+            **addinfo = '\0';
+            if (zoom_addinfo && *zoom_addinfo)
+            {
+                strcpy(*addinfo, zoom_addinfo);
+                strcat(*addinfo, " ");
+            }
+            strcat(*addinfo, "(backend)");
         }
     }
 }
 
 void yf::Zoom::Backend::connect(std::string zurl,
-                                int *error, const char **addinfo)
+                                int *error, char **addinfo,
+                                ODR odr)
 {
     ZOOM_connection_connect(m_connection, zurl.c_str(), 0);
-    get_zoom_error(error, addinfo);
+    get_zoom_error(error, addinfo, odr);
 }
 
 void yf::Zoom::Backend::search_pqf(const char *pqf, Odr_int *hits,
-                                   int *error, const char **addinfo)
+                                   int *error, char **addinfo, ODR odr)
 {
     m_resultset = ZOOM_connection_search_pqf(m_connection, pqf);
-    get_zoom_error(error, addinfo);
+    get_zoom_error(error, addinfo, odr);
     if (*error == 0)
         *hits = ZOOM_resultset_size(m_resultset);
     else
@@ -216,7 +245,7 @@ void yf::Zoom::Backend::search_pqf(const char *pqf, Odr_int *hits,
 }
 
 void yf::Zoom::Backend::search_cql(const char *cql, Odr_int *hits,
-                                   int *error, const char **addinfo)
+                                   int *error, char **addinfo, ODR odr)
 {
     ZOOM_query q = ZOOM_query_create();
 
@@ -224,7 +253,7 @@ void yf::Zoom::Backend::search_cql(const char *cql, Odr_int *hits,
 
     m_resultset = ZOOM_connection_search(m_connection, q);
     ZOOM_query_destroy(q);
-    get_zoom_error(error, addinfo);
+    get_zoom_error(error, addinfo, odr);
     if (*error == 0)
         *hits = ZOOM_resultset_size(m_resultset);
     else
@@ -233,10 +262,10 @@ void yf::Zoom::Backend::search_cql(const char *cql, Odr_int *hits,
 
 void yf::Zoom::Backend::present(Odr_int start, Odr_int number,
                                 ZOOM_record *recs,
-                                int *error, const char **addinfo)
+                                int *error, char **addinfo, ODR odr)
 {
     ZOOM_resultset_records(m_resultset, recs, start, number);
-    get_zoom_error(error, addinfo);
+    get_zoom_error(error, addinfo, odr);
 }
 
 void yf::Zoom::Backend::set_option(const char *name, const char *value)
@@ -400,6 +429,11 @@ yf::Zoom::SearchablePtr yf::Zoom::Impl::parse_torus_record(const xmlNode *ptr)
             s->transform_xsl_fname = mp::xml::get_text(ptr);
         }
         else if (!strcmp((const char *) ptr->name,
+                         "urlRecipe"))
+        {
+            s->urlRecipe = mp::xml::get_text(ptr);
+        }
+        else if (!strcmp((const char *) ptr->name,
                          "useTurboMarc"))
         {
             ; // useTurboMarc is ignored
@@ -463,8 +497,15 @@ void yf::Zoom::Impl::configure_local_records(const xmlNode *ptr, bool test_only)
     }
 }
 
-void yf::Zoom::Impl::configure(const xmlNode *ptr, bool test_only)
+void yf::Zoom::Impl::configure(const xmlNode *ptr, bool test_only,
+                               const char *path)
 {
+    if (path && *path)
+    {
+        file_path = path;
+        if (path[strlen(path)-1] != '/')
+            file_path += "/";
+    }
     for (ptr = ptr->children; ptr; ptr = ptr->next)
     {
         if (ptr->type != XML_ELEMENT_NODE)
@@ -530,6 +571,7 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
     if (m_backend && m_backend->m_frontend_database == database)
         return m_backend;
 
+    const char *sru_proxy = 0;
     std::string db_args;
     std::string torus_db;
     size_t db_arg_pos = database.find(',');
@@ -590,7 +632,7 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
         if (m_p->xsldir.length()) 
             fname = m_p->xsldir + "/" + sptr->transform_xsl_fname;
         else
-            fname = sptr->transform_xsl_fname;
+            fname = m_p->file_path + sptr->transform_xsl_fname;
         xmlDoc *xsp_doc = xmlParseFile(fname.c_str());
         if (!xsp_doc)
         {
@@ -628,19 +670,34 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
     {
         // A CF target
         b->set_option("user", sptr->cfAuth.c_str());
-        if (authentication.length() && db_args.length() == 0)
+        if (db_args.length() == 0)
         {
-            // no database (auth) args specified already.. and the
-            // Torus authentication has it.. Generate the args that CF
-            // understands..
-            size_t found = authentication.find('/');
-            if (found != std::string::npos)
+            if (authentication.length())
             {
-                db_args += "user=" + mp::util::uri_encode(authentication.substr(0, found))
-                    + "&password=" + mp::util::uri_encode(authentication.substr(found+1));
+                // no database (auth) args specified already.. and the
+                // Torus authentication has it.. Generate the args that CF
+                // understands..
+                size_t found = authentication.find('/');
+                if (found != std::string::npos)
+                {
+                    db_args += "user=" + mp::util::uri_encode(authentication.substr(0, found))
+                        + "&password=" + mp::util::uri_encode(authentication.substr(found+1));
+                }
+                else
+                    db_args += "user=" + mp::util::uri_encode(authentication);
             }
-            else
-                db_args += "user=" + mp::util::uri_encode(authentication);
+            if (sptr->cfProxy.length())
+            {
+                if (db_args.length())
+                    db_args += "&";
+                db_args += "proxy=" + mp::util::uri_encode(sptr->cfProxy);
+            }
+        }
+        if (sptr->cfSubDb.length())
+        {
+            if (db_args.length())
+                db_args += "&";
+            db_args += "subdatabase=" + mp::util::uri_encode(sptr->cfSubDb);
         }
     }
     else
@@ -664,6 +721,8 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
                     param_user = value;
                 else if (!strcmp(name, "password"))
                     param_password = value;
+                else if (!strcmp(name, "proxy"))
+                    sru_proxy = value;
                 else
                 {
                     BackendPtr notfound;
@@ -692,18 +751,9 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
                 b->set_option("user", authentication.c_str());
         }
     }
-    if (sptr->cfProxy.length())
-    {
-        if (db_args.length())
-            db_args += "&";
-        db_args += "proxy=" + mp::util::uri_encode(sptr->cfProxy);
-    }
-    if (sptr->cfSubDb.length())
-    {
-        if (db_args.length())
-            db_args += "&";
-        db_args += "subdatabase=" + mp::util::uri_encode(sptr->cfSubDb);
-    }
+
+    if (sru_proxy)
+        b->set_option("proxy", sru_proxy);
 
     std::string url;
     if (sptr->sru.length())
@@ -718,10 +768,7 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
     if (db_args.length())
         url += "," + db_args;
     yaz_log(YLOG_LOG, "url=%s", url.c_str());
-    const char *addinfo_c = 0;
-    b->connect(url, error, &addinfo_c);
-    if (addinfo_c)
-        *addinfo = odr_strdup(odr, addinfo_c);
+    b->connect(url, error, addinfo, odr);
     if (*error == 0)
     {
         m_backend = b;
@@ -732,7 +779,7 @@ yf::Zoom::BackendPtr yf::Zoom::Frontend::get_backend_from_databases(
 Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
                                            Odr_int number_to_present,
                                            int *error,
-                                           const char **addinfo,
+                                           char **addinfo,
                                            Odr_int *number_of_records_returned,
                                            ODR odr,
                                            BackendPtr b,
@@ -743,6 +790,7 @@ Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
     Z_Records *records = 0;
     bool enable_pz2_retrieval = false; // whether target profile is used
     bool enable_pz2_transform = false; // whether XSLT is used as well
+    bool assume_marc8_charset = false;
 
     if (start < 0 || number_to_present <= 0)
         return records;
@@ -774,7 +822,14 @@ Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
     if (enable_pz2_retrieval)
     {
         if (b->sptr->request_syntax.length())
+        {
             syntax_name = b->sptr->request_syntax.c_str();
+            const Odr_oid *syntax_oid = 
+                yaz_string_to_oid(yaz_oid_std(), CLASS_RECSYN, syntax_name);
+            if (!oid_oidcmp(syntax_oid, yaz_oid_recsyn_usmarc)
+                || !oid_oidcmp(syntax_oid, yaz_oid_recsyn_opac))
+                assume_marc8_charset = true;
+        }
     }
     else if (preferredRecordSyntax)
         syntax_name =
@@ -791,7 +846,7 @@ Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
 
     b->set_option("elementSetName", element_set_name);
 
-    b->present(start, number_to_present, recs, error, addinfo);
+    b->present(start, number_to_present, recs, error, addinfo, odr);
 
     Odr_int i = 0;
     if (!*error)
@@ -825,19 +880,22 @@ Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
             else if (enable_pz2_retrieval)
             {
                 char rec_type_str[100];
+                const char *record_encoding = 0;
 
-                strcpy(rec_type_str, b->sptr->use_turbomarc ?
-                       "txml" : "xml");
-                // prevent buffer overflow ...
-                if (b->sptr->record_encoding.length() > 0 &&
-                    b->sptr->record_encoding.length() < 
-                    (sizeof(rec_type_str)-20))
+                if (b->sptr->record_encoding.length())
+                    record_encoding = b->sptr->record_encoding.c_str();
+                else if (assume_marc8_charset)
+                    record_encoding = "marc8";
+
+                strcpy(rec_type_str, b->sptr->use_turbomarc ? "txml" : "xml");
+                if (record_encoding)
                 {
                     strcat(rec_type_str, "; charset=");
-                    strcat(rec_type_str, b->sptr->record_encoding.c_str());
+                    strcat(rec_type_str, record_encoding);
                 }
                 
                 int rec_len;
+                xmlChar *xmlrec_buf = 0;
                 const char *rec_buf = ZOOM_record_get(recs[i], rec_type_str,
                                                       &rec_len);
                 if (rec_buf && b->xsp && enable_pz2_transform)
@@ -849,13 +907,39 @@ Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
                         rec_res = xsltApplyStylesheet(b->xsp, rec_doc, 0);
 
                         if (rec_res)
-                            xsltSaveResultToString((xmlChar **) &rec_buf, &rec_len,
+                            xsltSaveResultToString(&xmlrec_buf, &rec_len,
                                                    rec_res, b->xsp);
+                        rec_buf = (const char *) xmlrec_buf;
+                        xmlFreeDoc(rec_doc);
+                        xmlFreeDoc(rec_res);
                     }
                 }
 
                 if (rec_buf)
                 {
+                    xmlDoc *doc = xmlParseMemory(rec_buf, rec_len);
+                    std::string res = 
+                        mp::xml::url_recipe_handle(doc, b->sptr->urlRecipe);
+                    if (res.length())
+                    {
+                        xmlNode *ptr = xmlDocGetRootElement(doc);
+                        while (ptr && ptr->type != XML_ELEMENT_NODE)
+                            ptr = ptr->next;
+                        xmlNode *c = 
+                            xmlNewChild(ptr, 0, BAD_CAST "generated-url", 0);
+                        xmlNode * t = xmlNewText(BAD_CAST res.c_str());
+                        xmlAddChild(c, t);
+
+                        if (xmlrec_buf)
+                            xmlFree(xmlrec_buf);
+
+                        xmlDocDumpMemory(doc, &xmlrec_buf, &rec_len);
+                        rec_buf = (const char *) xmlrec_buf;
+                    }
+                    xmlFreeDoc(doc);
+                }
+                if (rec_buf)
+                {
                     npr = (Z_NamePlusRecord *) odr_malloc(odr, sizeof(*npr));
                     npr->databaseName = odr_database;
                     npr->which = Z_NamePlusRecord_databaseRecord;
@@ -869,6 +953,8 @@ Z_Records *yf::Zoom::Frontend::get_records(Odr_int start,
                         YAZ_BIB1_SYSTEM_ERROR_IN_PRESENTING_RECORDS,
                         rec_type_str);
                 }
+                if (xmlrec_buf)
+                    xmlFree(xmlrec_buf);
             }
             else
             {
@@ -947,19 +1033,17 @@ void yf::Zoom::Frontend::handle_search(mp::Package &package)
     }
 
     int error = 0;
-    char *addinfo_s = 0;
+    char *addinfo = 0;
     std::string db(sr->databaseNames[0]);
-    BackendPtr b = get_backend_from_databases(db, &error, &addinfo_s, odr);
+    BackendPtr b = get_backend_from_databases(db, &error, &addinfo, odr);
     if (error)
     {
         apdu_res = 
-            odr.create_searchResponse(
-                apdu_req, error, addinfo_s);
+            odr.create_searchResponse(apdu_req, error, addinfo);
         package.response() = apdu_res;
         return;
     }
 
-    const char *addinfo_c = 0;
     b->set_option("setname", "default");
 
     Odr_int hits = 0;
@@ -1002,14 +1086,14 @@ void yf::Zoom::Frontend::handle_search(mp::Package &package)
         if (cn_error)
         {
             // hopefully we are getting a ptr to a index+relation+term node
-            addinfo_c = 0;
+            addinfo = 0;
             if (cn_error->which == CQL_NODE_ST)
-                addinfo_c = cn_error->u.st.index;
+                addinfo = cn_error->u.st.index;
 
             apdu_res = 
                 odr.create_searchResponse(apdu_req, 
                                           YAZ_BIB1_UNSUPP_USE_ATTRIBUTE,
-                                          addinfo_c);
+                                          addinfo);
             package.response() = apdu_res;
             return;
         }
@@ -1103,7 +1187,7 @@ void yf::Zoom::Frontend::handle_search(mp::Package &package)
         if (status == 0)
         {
             yaz_log(YLOG_LOG, "search CQL: %s", wrbuf_cstr(wrb));
-            b->search_cql(wrbuf_cstr(wrb), &hits, &error, &addinfo_c);
+            b->search_cql(wrbuf_cstr(wrb), &hits, &error, &addinfo, odr);
         }
         
         wrbuf_destroy(wrb);
@@ -1120,7 +1204,7 @@ void yf::Zoom::Frontend::handle_search(mp::Package &package)
     else
     {
         yaz_log(YLOG_LOG, "search PQF: %s", wrbuf_cstr(pqf_wrbuf));
-        b->search_pqf(wrbuf_cstr(pqf_wrbuf), &hits, &error, &addinfo_c);
+        b->search_pqf(wrbuf_cstr(pqf_wrbuf), &hits, &error, &addinfo, odr);
         wrbuf_destroy(pqf_wrbuf);
     }
     
@@ -1132,10 +1216,10 @@ void yf::Zoom::Frontend::handle_search(mp::Package &package)
     
     Odr_int number_of_records_returned = 0;
     Z_Records *records = get_records(
-        0, number_to_present, &error, &addinfo_c,
+        0, number_to_present, &error, &addinfo,
         &number_of_records_returned, odr, b, sr->preferredRecordSyntax,
         element_set_name);
-    apdu_res = odr.create_searchResponse(apdu_req, error, addinfo_c);
+    apdu_res = odr.create_searchResponse(apdu_req, error, addinfo);
     if (records)
     {
         apdu_res->u.searchResponse->records = records;
@@ -1173,7 +1257,7 @@ void yf::Zoom::Frontend::handle_present(mp::Package &package)
         element_set_name = comp->u.simple->u.generic;
     Odr_int number_of_records_returned = 0;
     int error = 0;
-    const char *addinfo = 0;
+    char *addinfo = 0;
     Z_Records *records = get_records(
         *pr->resultSetStartPoint - 1, *pr->numberOfRecordsRequested,
         &error, &addinfo, &number_of_records_returned, odr, m_backend,