Rename yaz_use_attribute_create to zget_AttributeList_use_string
[yaz-moved-to-github.git] / src / solr.c
index 8c1e29f..c1c1bba 100644 (file)
@@ -1,11 +1,14 @@
 /* This file is part of the YAZ toolkit.
- * Copyright (C) 1995-2010 Index Data
+ * Copyright (C) Index Data
  * See the file LICENSE for details.
  */
 /**
- * \file srwutil.c
- * \brief Implements SRW/SRU utilities.
+ * \file solr.c
+ * \brief Implements Solr decoding/encoding
  */
+#if HAVE_CONFIG_H
+#include <config.h>
+#endif
 
 #include <stdlib.h>
 #include <assert.h>
 #include <yaz/log.h>
 #include <yaz/facet.h>
 #include <yaz/wrbuf.h>
+#include <yaz/proto.h>
 
 #include "sru-p.h"
 
+#define SOLR_MAX_PARAMETERS  100
+
 #if YAZ_HAVE_XML2
 #include <libxml/parser.h>
 #include <libxml/tree.h>
 
-#define SOLR_MAX_PARAMETERS  100
-
-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 void extract_text_node(xmlNodePtr node, WRBUF wrbuf) {
+    xmlNodePtr child;
+    for (child = node->children; child ; child = child->next)
+    {
+        if (child->type == XML_TEXT_NODE)
+            wrbuf_puts(wrbuf, (const char *) child->content);
+    }
 }
 
-
-static int match_xml_node_attribute(xmlNodePtr ptr, const char *node_name, const char *attribute_name, const char *value)
+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))
+    if (attribute_name)
+    {
+        attribute_value = yaz_element_attribute_value_get(ptr, node_name,
+                                                          attribute_name);
+        if (attribute_value && !strcmp(attribute_value, value))
+            return 1;
+    }
+    else /* No attribute to check */
         return 1;
     return 0;
 }
 
-static void yaz_solr_decode_result_docs(ODR o, xmlNodePtr ptr, Odr_int start, Z_SRW_searchRetrieveResponse *sr) {
+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;
@@ -62,7 +70,8 @@ static void yaz_solr_decode_result_docs(ODR o, xmlNodePtr ptr, Odr_int start, Z_
         if (node->type == XML_ELEMENT_NODE)
             sr->num_records++;
 
-    sr->records = odr_malloc(o, sizeof(*sr->records) * sr->num_records);
+    if (sr->num_records)
+        sr->records = odr_malloc(o, sizeof(*sr->records) * sr->num_records);
 
     for (node = ptr->children; node; node = node->next)
     {
@@ -79,11 +88,8 @@ static void yaz_solr_decode_result_docs(ODR o, xmlNodePtr ptr, Odr_int start, Z_
             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->recordData_buf =
+                odr_strdupn(o, (const char *) buf->content, buf->use);
             record->recordPosition = odr_intdup(o, start + offset + 1);
 
             xmlBufferFree(buf);
@@ -94,16 +100,21 @@ static void yaz_solr_decode_result_docs(ODR o, xmlNodePtr ptr, Odr_int start, Z_
     }
 }
 
-static int  yaz_solr_decode_result(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveResponse *sr) {
+static int 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")) {
+        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")) {
+            }
+            else if (!strcmp((const char *) attr->name, "start"))
+            {
                 start = odr_atoi((const char *) attr->children->content);
             }
         }
@@ -114,53 +125,26 @@ static int  yaz_solr_decode_result(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveRe
     return -1;
 }
 
-static Z_AttributeList *yaz_solr_use_atttribute_create(ODR o, const char *name) {
-    Z_AttributeList *attributes= (Z_AttributeList *) odr_malloc(o, sizeof(*attributes));
-    Z_AttributeElement ** elements;
-    attributes->num_attributes = 1;
-    /* TODO check on name instead
-    if (!attributes->num_attributes) {
-        attributes->attributes = (Z_AttributeElement**)odr_nullval();
-        return attributes;
-    }
-    */
-    elements = (Z_AttributeElement**) odr_malloc (o, attributes->num_attributes * sizeof(*elements));
-    elements[0] = (Z_AttributeElement*)odr_malloc(o,sizeof(**elements));
-    elements[0]->attributeType = odr_malloc(o, sizeof(*elements[0]->attributeType));
-   *elements[0]->attributeType = 1;
-    elements[0]->attributeSet = odr_nullval();
-    elements[0]->which = Z_AttributeValue_complex;
-    elements[0]->value.complex = (Z_ComplexAttribute *) odr_malloc(o, sizeof(Z_ComplexAttribute));
-    elements[0]->value.complex->num_list = 1;
-    elements[0]->value.complex->list = (Z_StringOrNumeric **) odr_malloc(o, 1 * sizeof(Z_StringOrNumeric *));
-    elements[0]->value.complex->list[0] = (Z_StringOrNumeric *) odr_malloc(o, sizeof(Z_StringOrNumeric));
-    elements[0]->value.complex->list[0]->which = Z_StringOrNumeric_string;
-    elements[0]->value.complex->list[0]->u.string = (Z_InternationalString *) odr_strdup(o, name);
-    elements[0]->value.complex->semanticAction = 0;
-    elements[0]->value.complex->num_semanticAction = 0;
-    attributes->attributes = elements;
-    return attributes;
-}
-
-
-static const char *get_facet_term_count(xmlNodePtr node, int *freq) {
-
-    const char *term = xml_node_attribute_value_get(node, "int", "name");
+static const char *get_facet_term_count(xmlNodePtr node, Odr_int *freq)
+{
+    const char *term = yaz_element_attribute_value_get(node, "int", "name");
     xmlNodePtr child;
     WRBUF wrbuf = wrbuf_alloc();
     if (!term)
         return term;
 
-    for (child = node->children; child ; child = child->next) {
+    for (child = node->children; child ; child = child->next)
+    {
         if (child->type == XML_TEXT_NODE)
-        wrbuf_puts(wrbuf, (const char *) child->content);
+            wrbuf_puts(wrbuf, (const char *) child->content);
     }
-    *freq = atoi(wrbuf_cstr(wrbuf));
+    *freq = odr_atoi(wrbuf_cstr(wrbuf));
     wrbuf_destroy(wrbuf);
     return term;
 }
 
-Z_FacetField *yaz_solr_decode_facet_field(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveResponse *sr)
+Z_FacetField *yaz_solr_decode_facet_field(ODR o, xmlNodePtr ptr,
+                                          Z_SRW_searchRetrieveResponse *sr)
 
 {
     Z_AttributeList *list;
@@ -169,28 +153,26 @@ Z_FacetField *yaz_solr_decode_facet_field(ODR o, xmlNodePtr ptr, Z_SRW_searchRet
     int index = 0;
     xmlNodePtr node;
     // USE attribute
-    const char* name = xml_node_attribute_value_get(ptr, "lst", "name");
-    char *pos = strstr(name, "_exact");
-    /* HACK */
-    if (pos) {
-        pos[0] = 0;
-    }
-    list = yaz_solr_use_atttribute_create(o, name);
-    for (node = ptr->children; node; node = node->next) {
+    const char* name = yaz_element_attribute_value_get(ptr, "lst", "name");
+    list = zget_AttributeList_use_string(o, name);
+    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;
+    for (node = ptr->children; node; node = node->next)
+    {
+        Odr_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);
+        facet_field_term_set(o, facet_field,
+                             facet_term_create_cstr(o, term, count), index);
         index++;
     }
     return facet_field;
 }
 
-static int yaz_solr_decode_facet_counts(ODR o, xmlNodePtr root, Z_SRW_searchRetrieveResponse *sr) {
+static int yaz_solr_decode_facet_counts(ODR o, xmlNodePtr root,
+                                        Z_SRW_searchRetrieveResponse *sr)
+{
     xmlNodePtr ptr;
     for (ptr = root->children; ptr; ptr = ptr->next)
     {
@@ -207,7 +189,9 @@ static int yaz_solr_decode_facet_counts(ODR o, xmlNodePtr root, Z_SRW_searchRetr
             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);
+                facet_list_field_set(o, facet_list,
+                                     yaz_solr_decode_facet_field(o, node, sr),
+                                     num_facets);
                 num_facets++;
             }
             sr->facetList = facet_list;
@@ -217,6 +201,121 @@ static int yaz_solr_decode_facet_counts(ODR o, xmlNodePtr root, Z_SRW_searchRetr
     return 0;
 }
 
+static void yaz_solr_decode_suggestion_values(xmlNodePtr listPptr, WRBUF wrbuf)
+{
+    xmlNodePtr node;
+    for (node = listPptr; node; node= node->next) {
+        if (!strcmp((char*) node->name, "lst")) {
+            xmlNodePtr child;
+            for (child = node->children; child; child= child->next) {
+                if (match_xml_node_attribute(child, "str", "name", "word")) {
+                    wrbuf_puts(wrbuf, "<suggestion>");
+                    extract_text_node(child, wrbuf);
+                    wrbuf_puts(wrbuf, "</suggestion>\n");
+                }
+            }
+        }
+    }
+}
+
+static void yaz_solr_decode_suggestion_lst(xmlNodePtr lstPtr, WRBUF wrbuf)
+{
+    xmlNodePtr node;
+    for (node = lstPtr; node; node= node->next) {
+        if (match_xml_node_attribute(node, "arr", "name", "suggestion")) {
+            yaz_solr_decode_suggestion_values(node->children, wrbuf);
+        }
+    }
+}
+
+static void yaz_solr_decode_misspelled(xmlNodePtr lstPtr, WRBUF wrbuf)
+{
+    xmlNodePtr node;
+    for (node = lstPtr; node; node= node->next)
+    {
+        if (!strcmp((const char*) node->name, "lst")) {
+            const char *misspelled = yaz_element_attribute_value_get(node, "lst", "name");
+            if (misspelled) {
+                wrbuf_printf(wrbuf, "<misspelled term=\"%s\">\n", misspelled);
+                yaz_solr_decode_suggestion_lst(node->children, wrbuf);
+                wrbuf_puts(wrbuf, "</misspelled>\n");
+            }
+        }
+    }
+}
+
+static int yaz_solr_decode_spellcheck(ODR o, xmlNodePtr spellcheckPtr, Z_SRW_searchRetrieveResponse *sr)
+{
+    xmlNodePtr ptr;
+    WRBUF wrbuf = wrbuf_alloc();
+    wrbuf_puts(wrbuf, "");
+    for (ptr = spellcheckPtr->children; ptr; ptr = ptr->next)
+    {
+        if (match_xml_node_attribute(ptr, "lst", "name", "suggestions"))
+        {
+            yaz_solr_decode_misspelled(ptr->children, wrbuf);
+        }
+    }
+    sr->suggestions = odr_strdup(o, wrbuf_cstr(wrbuf));
+    return 0;
+}
+
+static int yaz_solr_decode_scan_result(ODR o, xmlNodePtr ptr,
+                                       Z_SRW_scanResponse *scr)
+{
+    xmlNodePtr node;
+    char *pos;
+    int i = 0;
+
+    /* find the actual list */
+    for (node = ptr->children; node; node = node->next)
+        if (node->type == XML_ELEMENT_NODE) {
+            ptr = node;
+            break;
+        }
+
+    scr->num_terms = 0;
+    for (node = ptr->children; node; node = node->next)
+        if (node->type == XML_ELEMENT_NODE && !strcmp((const char *) node->name, "int"))
+            scr->num_terms++;
+
+    if (scr->num_terms)
+        scr->terms = odr_malloc(o, sizeof(*scr->terms) * scr->num_terms);
+
+    for (node = ptr->children; node; node = node->next)
+    {
+        if (node->type == XML_ELEMENT_NODE && !strcmp((const char *) node->name, "int"))
+        {
+            Z_SRW_scanTerm *term = scr->terms + i;
+
+            Odr_int count = 0;
+            const char *val = get_facet_term_count(node, &count);
+
+            term->numberOfRecords = odr_intdup(o, count);
+
+            /* if val contains a ^ then it is probably term<^>display term so separate them. This is due to
+             * SOLR not being able to encode them into 2 separate attributes.
+             */
+            pos = strchr(val, '^');
+            if (pos != NULL) {
+               term->displayTerm = odr_strdup(o, pos + 1);
+               *pos = '\0';
+               term->value = odr_strdup(o, val);
+               *pos = '^';
+            } else {
+               term->value = odr_strdup(o, val);
+               term->displayTerm = NULL;
+            }
+            term->whereInList = NULL;
+
+            i++;
+        }
+    }
+
+    if (scr->num_terms)
+        return 0;
+    return -1;
+}
 #endif
 
 int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
@@ -227,8 +326,9 @@ 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;
-    Z_SRW_PDU *pdu = yaz_srw_get(o, Z_SRW_searchRetrieve_response);
-    Z_SRW_searchRetrieveResponse *sr = pdu->u.response;
+    Z_SRW_PDU *pdu;
+    Z_SRW_searchRetrieveResponse *sr = NULL;
+    Z_SRW_scanResponse *scr = NULL;
 
     if (!doc)
     {
@@ -253,12 +353,28 @@ int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
             for (ptr = root->children; ptr; ptr = ptr->next)
             {
                 if (ptr->type == XML_ELEMENT_NODE &&
-                    !strcmp((const char *) ptr->name, "result"))
+                    !strcmp((const char *) ptr->name, "result")) {
+                        pdu = yaz_srw_get(o, Z_SRW_searchRetrieve_response);
+                        sr = pdu->u.response;
                         rc_result = yaz_solr_decode_result(o, ptr, sr);
+                }
+                if (ptr->type == XML_ELEMENT_NODE &&
+                    match_xml_node_attribute(ptr, "lst", "name", "terms")) {
+                        pdu = yaz_srw_get(o, Z_SRW_scan_response);
+                        scr = pdu->u.scan_response;
+                        rc_result = yaz_solr_decode_scan_result(o, ptr, scr);
+                }
                 /* TODO The check on hits is a work-around to avoid garbled facets on zero results from the SOLR server.
                  * The work-around works because the results is before the facets in the xml. */
-                if (rc_result == 0 && match_xml_node_attribute(ptr, "lst", "name", "facet_counts"))
-                    rc_facets =  yaz_solr_decode_facet_counts(o, ptr, sr);
+                if (sr) {
+                    if (rc_result == 0 &&  *sr->numberOfRecords > 0 &&
+                        match_xml_node_attribute(ptr, "lst", "name", "facet_counts"))
+                            rc_facets =  yaz_solr_decode_facet_counts(o, ptr, sr);
+                    if (rc_result == 0 &&  *sr->numberOfRecords == 0 &&
+                        match_xml_node_attribute(ptr, "lst", "name", "spellcheck"))
+                            rc_facets =  yaz_solr_decode_spellcheck(o, ptr, sr);
+                }
+
             }
             ret = rc_result + rc_facets;
         }
@@ -273,43 +389,89 @@ int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
 #endif
 }
 
-static void yaz_solr_encode_facet_field(ODR encode, char **name, char **value, int *i, Z_FacetField *facet_field, int *limit) {
-      Z_AttributeList *attribute_list = facet_field->attributes;
-      struct yaz_facet_attr attr_values;
-      yaz_facet_attr_init(&attr_values);
-      yaz_facet_attr_get_z_attributes(attribute_list, &attr_values);
-      // TODO do we want to support server decided
-      if (!attr_values.errcode && attr_values.useattr) {
-          WRBUF wrbuf = wrbuf_alloc();
-          wrbuf_puts(wrbuf, (char *) attr_values.useattr);
-          /* Skip date field */
-          if (strcmp("date", attr_values.useattr) != 0)
-              wrbuf_puts(wrbuf, "_exact");
-          yaz_add_name_value_str(encode, name, value, i, "facet.field", odr_strdup(encode, wrbuf_cstr(wrbuf)));
-          if (attr_values.limit > 0) {
-              WRBUF wrbuf2 = wrbuf_alloc();
-              Odr_int olimit;
-              wrbuf_puts(wrbuf2, "f.");
-              wrbuf_puts(wrbuf2, wrbuf_cstr(wrbuf));
-              wrbuf_puts(wrbuf2, ".facet.limit");
-              olimit = attr_values.limit;
-              yaz_add_name_value_int(encode, name, value, i, odr_strdup(encode, wrbuf_cstr(wrbuf2)), &olimit);
-              wrbuf_destroy(wrbuf2);
-          }
-          wrbuf_destroy(wrbuf);
-      }
-}
+static int yaz_solr_encode_facet_field(
+    ODR encode, char **name, char **value, int *i,
+    Z_FacetField *facet_field)
+{
+    Z_AttributeList *attribute_list = facet_field->attributes;
+    struct yaz_facet_attr attr_values;
+    yaz_facet_attr_init(&attr_values);
+    yaz_facet_attr_get_z_attributes(attribute_list, &attr_values);
 
-static void yaz_solr_encode_facet_list(ODR encode, char **name, char **value, int *i, Z_FacetList *facet_list, int *limit) {
+    if (attr_values.errcode)
+        return -1;
+    if (attr_values.useattr)
+    {
+        WRBUF wrbuf = wrbuf_alloc();
+        yaz_add_name_value_str(encode, name, value, i,
+                               "facet.field",
+                               odr_strdup(encode, attr_values.useattr));
+
+        if (attr_values.limit > 0)
+        {
+            Odr_int v = attr_values.limit;
+            wrbuf_rewind(wrbuf);
+            wrbuf_printf(wrbuf, "f.%s.facet.limit", attr_values.useattr);
+            yaz_add_name_value_int(encode, name, value, i,
+                                   odr_strdup(encode, wrbuf_cstr(wrbuf)),
+                                   &v);
+        }
+        if (attr_values.start > 1)
+        {
+            Odr_int v = attr_values.start - 1;
+            wrbuf_rewind(wrbuf);
+            wrbuf_printf(wrbuf, "f.%s.facet.offset", attr_values.useattr);
+            yaz_add_name_value_int(encode, name, value, i,
+                                   odr_strdup(encode, wrbuf_cstr(wrbuf)),
+                                   &v);
+        }
+        if (attr_values.sortorder == 1)
+        {
+            wrbuf_rewind(wrbuf);
+            wrbuf_printf(wrbuf, "f.%s.facet.sort", attr_values.useattr);
+            yaz_add_name_value_str(encode, name, value, i,
+                                   odr_strdup(encode, wrbuf_cstr(wrbuf)),
+                                   "index");
+        }
+        wrbuf_destroy(wrbuf);
+    }
+    else
+    {
+        if (attr_values.limit > 0)
+        {
+            Odr_int v = attr_values.limit;
+            yaz_add_name_value_int(encode, name, value, i, "facet.limit", &v);
+        }
+        if (attr_values.start > 1)
+        {
+            Odr_int v = attr_values.start - 1;
+            yaz_add_name_value_int(encode, name, value, i, "facet.offset", &v);
+        }
+        if (attr_values.sortorder == 1)
+        {
+            yaz_add_name_value_str(encode, name, value, i, "facet.sort",
+                                   "index");
+        }
+    }
+    return 0;
+}
 
+static int yaz_solr_encode_facet_list(
+    ODR encode, char **name, char **value,
+    int *i, Z_FacetList *facet_list)
+{
     int index;
-    for (index = 0; index < facet_list->num; index++)  {
-        yaz_solr_encode_facet_field(encode, name, value, i, facet_list->elements[index], limit);
+    for (index = 0; index < facet_list->num; index++)
+    {
+        int r = yaz_solr_encode_facet_field(encode, name, value, i,
+                                            facet_list->elements[index]);
+        if (r)
+            return -1;
 
     }
+    return 0;
 }
 
-
 int yaz_solr_encode_request(Z_HTTP_Request *hreq, Z_SRW_PDU *srw_pdu,
                             ODR encode, const char *charset)
 {
@@ -318,29 +480,22 @@ int yaz_solr_encode_request(Z_HTTP_Request *hreq, Z_SRW_PDU *srw_pdu,
     char *name[SOLR_MAX_PARAMETERS], *value[SOLR_MAX_PARAMETERS];
     char *uri_args;
     char *path;
+    char *q;
+    char *pos;
+    char *cp;
+    const char *path_args = 0;
     int i = 0;
 
-    z_HTTP_header_add_basic_auth(encode, &hreq->headers, 
+    z_HTTP_header_add_basic_auth(encode, &hreq->headers,
                                  srw_pdu->username, srw_pdu->password);
-
-    switch (srw_pdu->which)
+    if (srw_pdu->which == 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", request->query.pqf);
-            break;
-        case Z_SRW_query_type_cql:
-            yaz_add_name_value_str(encode, name, value, &i,
-                                   "q", request->query.cql);
-            break;
-        default:
+        if (!srw_pdu->u.request->query)
             return -1;
-        }
+        /* not considering query type here ! */
+        yaz_add_name_value_str(encode, name, value, &i, "q", request->query);
         if (srw_pdu->u.request->startRecord)
         {
             Odr_int start = *request->startRecord - 1;
@@ -352,33 +507,108 @@ int yaz_solr_encode_request(Z_HTTP_Request *hreq, Z_SRW_PDU *srw_pdu,
         yaz_add_name_value_str(encode, name, value, &i,
                                "fl", request->recordSchema);
 
-        if (request->facetList) {
+        switch(srw_pdu->u.request->sort_type)
+        {
+        case Z_SRW_sort_type_none:
+            break;
+        case Z_SRW_sort_type_sort:
+            yaz_add_name_value_str(encode, name, value, &i, "sort",
+                                   srw_pdu->u.request->sort.sortKeys);
+            break;
+        }
+        if (request->facetList)
+        {
             Z_FacetList *facet_list = request->facetList;
-            int limit = 0;
             yaz_add_name_value_str(encode, name, value, &i, "facet", "true");
             yaz_add_name_value_str(encode, name, value, &i, "facet.mincount", "1");
-            yaz_solr_encode_facet_list(encode, name, value, &i, facet_list, &limit);
-            /*
-            olimit = limit;
-            yaz_add_name_value_int(encode, name, value, &i, "facet.limit", &olimit);
-             */
-
+            if (yaz_solr_encode_facet_list(encode, name, value, &i, facet_list))
+                return -1;
+        }
+    }
+    else if (srw_pdu->which == Z_SRW_scan_request) {
+        Z_SRW_scanRequest *request = srw_pdu->u.scan_request;
+        solr_op = "terms";
+        if (!srw_pdu->u.scan_request->scanClause)
+            return -1;
+        if (!strcmp(srw_pdu->u.scan_request->queryType, "pqf"))
+        {
+            yaz_add_name_value_str(encode, name, value, &i,
+                                   "terms.fl", request->scanClause);
+            yaz_add_name_value_str(encode, name, value, &i,
+                                   "terms.lower", request->scanClause);
+        }
+        else if (!strcmp(srw_pdu->u.scan_request->queryType, "cql"))
+        {
+            q = request->scanClause;
+            pos = strchr(q, ':');
+            if (pos != NULL) {
+                yaz_add_name_value_str(encode, name, value, &i,
+                                       "terms.lower", odr_strdup(encode, pos + 1));
+                *pos = '\0';
+                yaz_add_name_value_str(encode, name, value, &i,
+                                       "terms.fl", odr_strdup(encode, q));
+                *pos = ':';
+            } else {
+                yaz_add_name_value_str(encode, name, value, &i,
+                                       "terms.lower", odr_strdup(encode, q));
+            }
         }
-        break;
+        else
+            return -1;
+        yaz_add_name_value_str(encode, name, value, &i,
+                               "terms.sort", "index");
+        yaz_add_name_value_int(encode, name, value, &i,
+                               "terms.limit", request->maximumTerms);
     }
-    default:
+    else
         return -1;
+
+    if (srw_pdu->extra_args)
+    {
+        Z_SRW_extra_arg *ea = srw_pdu->extra_args;
+        for (; ea && i < SOLR_MAX_PARAMETERS; ea = ea->next)
+        {
+            name[i] = ea->name;
+            value[i] = ea->value;
+            i++;
+        }
     }
-    name[i] = 0;
+
+    name[i++] = 0;
+
     yaz_array_to_uri(&uri_args, encode, name, value);
-    
+
     hreq->method = "GET";
-    
+
     path = (char *)
         odr_malloc(encode, strlen(hreq->path) +
-                   strlen(uri_args) + strlen(solr_op) + 4);
+                   strlen(uri_args) + strlen(solr_op) + 5);
 
-    sprintf(path, "%s/%s?%s", hreq->path, solr_op, uri_args);
+    cp = strchr(hreq->path, '#');
+    if (cp)
+        *cp = '\0';
+    cp = strchr(hreq->path, '?');
+    if (cp)
+    {
+        *cp = '\0'; /* args in path */
+        path_args = cp + 1;
+    }
+    strcpy(path, hreq->path);
+    cp = strrchr(path, '/');
+    if (cp)
+    {
+        if (!strcmp(cp, "/select") || !strcmp(cp, "/"))
+            *cp = '\0';
+    }
+    strcat(path, "/");
+    strcat(path, solr_op);
+    strcat(path, "?");
+    if (path_args)
+    {
+        strcat(path, path_args);
+        strcat(path, "&");
+    }
+    strcat(path, uri_args);
     hreq->path = path;
 
     z_HTTP_header_add_content_type(encode, &hreq->headers,