Decode SOLR Facet response into Z_FacetList
authorDennis Schafroth <dennis@indexdata.com>
Wed, 25 Aug 2010 14:47:50 +0000 (16:47 +0200)
committerDennis Schafroth <dennis@indexdata.com>
Wed, 25 Aug 2010 14:47:50 +0000 (16:47 +0200)
src/solr.c

index a7cdb52..4e0a74a 100644 (file)
 #include <yaz/matchstr.h>
 #include <yaz/yaz-iconv.h>
 #include <yaz/log.h>
+#include <yaz/facet.h>
 
 #include "sru-p.h"
 
 #if YAZ_HAVE_XML2
 #include <libxml/parser.h>
 #include <libxml/tree.h>
+
+
+const char *xml_node_attribute_value_get(xmlNodePtr ptr, const char *node_name, const char *attribute_name) {
+
+    struct _xmlAttr *attr;
+    // check if the node name matches
+    if (strcmp((const char*) ptr->name, node_name))
+        return 0;
+    // check if the attribute name and return the value
+    for (attr = ptr->properties; attr; attr = attr->next)
+        if (attr->children && attr->children->type == XML_TEXT_NODE) {
+            if (!strcmp((const char *) attr->name, attribute_name))
+                return (const char *) attr->children->content;
+        }
+    return 0;
+}
+
+
+static int match_xml_node_attribute(xmlNodePtr ptr, const char *node_name, const char *attribute_name, const char *value)
+{
+    const char *attribute_value;
+    // check if the node name matches
+    if (strcmp((const char*) ptr->name, node_name))
+        return 0;
+    attribute_value = xml_node_attribute_value_get(ptr, node_name, attribute_name);
+    if (attribute_value && !strcmp(attribute_value, value))
+        return 1;
+    return 0;
+}
+
+static void yaz_solr_decode_result_docs(ODR o, xmlNodePtr ptr, Odr_int start, Z_SRW_searchRetrieveResponse *sr) {
+    xmlNodePtr node;
+    int offset = 0;
+    int i = 0;
+
+    sr->num_records = 0;
+    for (node = ptr->children; node; node = node->next)
+        if (node->type == XML_ELEMENT_NODE)
+            sr->num_records++;
+
+    sr->records = odr_malloc(o, sizeof(*sr->records) * sr->num_records);
+
+    for (node = ptr->children; node; node = node->next)
+    {
+        if (node->type == XML_ELEMENT_NODE)
+        {
+            Z_SRW_record *record = sr->records + i;
+            xmlBufferPtr buf = xmlBufferCreate();
+            xmlNode *tmp = xmlCopyNode(node, 1);
+
+            xmlNodeDump(buf, tmp->doc, tmp, 0, 0);
+
+            xmlFreeNode(tmp);
+
+            record->recordSchema = 0;
+            record->recordPacking = Z_SRW_recordPacking_XML;
+            record->recordData_len = buf->use;
+            record->recordData_buf = odr_malloc(o, buf->use + 1);
+            memcpy(record->recordData_buf, buf->content, buf->use);
+            record->recordData_buf[buf->use] = '\0';
+            // TODO Solve the real problem in zoom-sru, that doesnt work with 0-based indexes.
+            // Work-around: Making the recordPosition 1-based.
+            record->recordPosition = odr_intdup(o, start + offset + 1);
+
+            xmlBufferFree(buf);
+
+            offset++;
+            i++;
+        }
+    }
+}
+
+static void yaz_solr_decode_result(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveResponse *sr) {
+    Odr_int start = 0;
+    struct _xmlAttr *attr;
+    for (attr = ptr->properties; attr; attr = attr->next)
+        if (attr->children && attr->children->type == XML_TEXT_NODE) {
+            if (!strcmp((const char *) attr->name, "numFound")) {
+                sr->numberOfRecords = odr_intdup(o, odr_atoi(
+                        (const char *) attr->children->content));
+            } else if (!strcmp((const char *) attr->name, "start")) {
+                start = odr_atoi((const char *) attr->children->content);
+            }
+        }
+    yaz_solr_decode_result_docs(o, ptr, start, sr);
+}
+
+static Z_AttributeList *yaz_solr_use_atttribute_create(ODR o, const char *name) {
+    // TODO IMPLEMENT
+    return 0;
+}
+
+
+static const char *get_facet_term_count(xmlNodePtr node, int *freq) {
+    // TODO implement
+    return 0;
+}
+
+Z_FacetField *yaz_solr_decode_facet_field(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveResponse *sr)
+{
+    // USE attribute
+    const char* name = xml_node_attribute_value_get(ptr, "lst", "name");
+    Z_AttributeList *list = yaz_solr_use_atttribute_create(o, name);
+    Z_FacetField *facet_field;
+    int num_terms = 0;
+    int index = 0;
+    xmlNodePtr node;
+    for (node = ptr->children; node; node = node->next) {
+        num_terms++;
+    }
+    facet_field = facet_field_create(o, list, num_terms);
+    index = 0;
+    for (node = ptr->children; node; node = node->next) {
+        int count = 0;
+        const char *term = get_facet_term_count(node, &count);
+        facet_field_term_set(o, facet_field, facet_term_create(o, term_create(o, term), count), index);
+        index++;
+    }
+    return facet_field;
+}
+
+static void yaz_solr_decode_facet_counts(ODR o, xmlNodePtr root, Z_SRW_searchRetrieveResponse *sr) {
+    xmlNodePtr ptr;
+    for (ptr = root->children; ptr; ptr = ptr->next)
+    {
+        if (match_xml_node_attribute(ptr, "lst", "name", "facet_fields"))
+        {
+            xmlNodePtr node;
+            Z_FacetList *facet_list;
+            int num_facets = 0;
+            for (node = ptr->children; node; node= node->next)
+            {
+                num_facets++;
+            }
+            facet_list = facet_list_create(o, num_facets);
+            num_facets = 0;
+            for (node = ptr->children; node; node= node->next)
+            {
+                facet_list_field_set(o, facet_list, yaz_solr_decode_facet_field(o, node, sr), num_facets);
+                num_facets++;
+            }
+            sr->facet_list = facet_list;
+            break;
+        }
+    }
+}
+
+static void yaz_solr_decode_facets(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveResponse *sr) {
+    if (match_xml_node_attribute(ptr, "lst", "name", "facet_counts"))
+        yaz_solr_decode_facet_counts(o, ptr->children, sr);
+}
 #endif
 
 int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
@@ -29,7 +181,6 @@ int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
     xmlDocPtr doc = xmlParseMemory(content_buf, content_len);
     int ret = 0;
     xmlNodePtr ptr = 0;
-    Odr_int start = 0;
     Z_SRW_PDU *pdu = yaz_srw_get(o, Z_SRW_searchRetrieve_response);
     Z_SRW_searchRetrieveResponse *sr = pdu->u.response;
 
@@ -55,7 +206,9 @@ int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
             {
                 if (ptr->type == XML_ELEMENT_NODE &&
                     !strcmp((const char *) ptr->name, "result"))
-                    break;
+                        yaz_solr_decode_result(o, ptr, sr);
+                if (match_xml_node_attribute(ptr, "lst", "name", "facet_counts"))
+                        yaz_solr_decode_facet_counts(o, ptr, sr);
             }
             if (!ptr)
             {
@@ -63,67 +216,6 @@ int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
             }
         }
     }
-    if (ptr)
-    {   /* got result node */
-        struct _xmlAttr *attr;
-
-        for (attr = ptr->properties; attr; attr = attr->next)
-            if (attr->children && attr->children->type == XML_TEXT_NODE)
-            {
-                if (!strcmp((const char *) attr->name, "numFound"))
-                {
-                    sr->numberOfRecords =
-                        odr_intdup(o, 
-                                   odr_atoi(
-                                       (const char *) attr->children->content));
-                }
-                else if (!strcmp((const char *) attr->name, "start"))
-                {
-                    start = odr_atoi((const char *) attr->children->content);
-                }
-            }
-    }
-    if (ptr)
-    {
-        xmlNodePtr node;
-        int offset = 0;
-        int i = 0;
-
-        sr->num_records = 0;
-        for (node = ptr->children; node; node = node->next)
-            if (node->type == XML_ELEMENT_NODE)
-                sr->num_records++;
-
-        sr->records = odr_malloc(o, sizeof(*sr->records) * sr->num_records);
-
-        for (node = ptr->children; node; node = node->next)
-        {
-            if (node->type == XML_ELEMENT_NODE)
-            {
-                Z_SRW_record *record = sr->records + i;
-                xmlBufferPtr buf = xmlBufferCreate();
-                xmlNode *tmp = xmlCopyNode(node, 1);
-
-                xmlNodeDump(buf, tmp->doc, tmp, 0, 0);
-
-                xmlFreeNode(tmp);
-
-                record->recordSchema = 0;
-                record->recordPacking = Z_SRW_recordPacking_XML;
-                record->recordData_len = buf->use;
-                record->recordData_buf = odr_malloc(o, buf->use + 1);
-                memcpy(record->recordData_buf, buf->content, buf->use);
-                record->recordData_buf[buf->use] = '\0';
-                // TODO Solve the real problem: Making the recordPosition 1-based due to "funny" code in zoom-sru
-                record->recordPosition = odr_intdup(o, start + offset + 1);
-
-                xmlBufferFree(buf);
-
-                offset++;
-                i++;
-            }
-        }
-    }
     if (doc)
         xmlFreeDoc(doc);
     if (ret == 0)
@@ -134,11 +226,17 @@ int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
 #endif
 }
 
+void solr_encode_facet_field(ODR encode, Z_FacetField *facet_field) {
+
+}
+
 int yaz_solr_encode_request(Z_HTTP_Request *hreq, Z_SRW_PDU *srw_pdu,
                             ODR encode, const char *charset)
 {
     const char *solr_op = 0;
-    char *name[30], *value[30];
+    //TODO Change. not a nice hard coded, unchecked limit.
+    int max = 100;
+    char *name[100], *value[100];
     char *uri_args;
     char *path;
     int i = 0;
@@ -148,33 +246,44 @@ int yaz_solr_encode_request(Z_HTTP_Request *hreq, Z_SRW_PDU *srw_pdu,
 
     switch (srw_pdu->which)
     {
-    case Z_SRW_searchRetrieve_request:
+    case Z_SRW_searchRetrieve_request: {
+        Z_SRW_searchRetrieveRequest *request = srw_pdu->u.request;
         solr_op = "select";
-        
         switch(srw_pdu->u.request->query_type)
         {
         case Z_SRW_query_type_pqf:
             yaz_add_name_value_str(encode, name, value, &i,
-                                   "q", srw_pdu->u.request->query.pqf);
+                                   "q", request->query.pqf);
             break;
         case Z_SRW_query_type_cql:
             yaz_add_name_value_str(encode, name, value, &i,
-                                   "q", srw_pdu->u.request->query.cql);
+                                   "q", request->query.cql);
             break;
         default:
             return -1;
         }
         if (srw_pdu->u.request->startRecord)
         {
-            Odr_int start = *srw_pdu->u.request->startRecord - 1;
+            Odr_int start = *request->startRecord - 1;
             yaz_add_name_value_int(encode, name, value, &i,
                                    "start", &start);
         }
         yaz_add_name_value_int(encode, name, value, &i,
-                               "rows", srw_pdu->u.request->maximumRecords);
+                               "rows", request->maximumRecords);
         yaz_add_name_value_str(encode, name, value, &i,
-                               "fl", srw_pdu->u.request->recordSchema);
+                               "fl", request->recordSchema);
+
+        if (request->facet_list) {
+            int index;
+            Z_FacetList *facet_list = request->facet_list;
+            yaz_add_name_value_str(encode, name, value, &i, "facet", "true");
+            for (index = 0; index < facet_list->num; index++) {
+                //TODO impl
+                //solr_encode_facet_field(encode, name, value, &i, facet_list->elements[index]);
+            }
+        }
         break;
+    }
     default:
         return -1;
     }