/* This file is part of the YAZ toolkit.
- * Copyright (C) 1995-2011 Index Data
+ * Copyright (C) Index Data
* See the file LICENSE for details.
*/
/**
#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>
-#endif
-static void extract_text_node(xmlNodePtr node, WRBUF wrbuf) {
+static void extract_text_node(xmlNodePtr node, WRBUF wrbuf)
+{
xmlNodePtr child;
for (child = node->children; child ; child = child->next)
{
for (node = ptr->children; node; node = node->next)
if (node->type == XML_ELEMENT_NODE)
sr->num_records++;
-
+
if (sr->num_records)
sr->records = odr_malloc(o, sizeof(*sr->records) * sr->num_records);
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);
{
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);
xmlNodePtr node;
// USE attribute
const char* name = yaz_element_attribute_value_get(ptr, "lst", "name");
- list = yaz_use_attribute_create(o, 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);
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")) {
+ 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")) {
+ 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")) {
+ 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) {
+ 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");
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)
{
#if YAZ_HAVE_XML2
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)
{
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 && *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);
+ 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;
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)
return -1;
if (attr_values.useattr)
{
WRBUF wrbuf = wrbuf_alloc();
- wrbuf_puts(wrbuf, (char *) attr_values.useattr);
yaz_add_name_value_str(encode, name, value, i,
"facet.field",
- odr_strdup(encode, wrbuf_cstr(wrbuf)));
+ odr_strdup(encode, attr_values.useattr));
+
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;
+ 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(wrbuf2)),
- &olimit);
- wrbuf_destroy(wrbuf2);
+ 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;
}
facet_list->elements[index]);
if (r)
return -1;
-
+
}
return 0;
}
ODR encode, const char *charset)
{
const char *solr_op = 0;
- //TODO Change. not a nice hard coded, unchecked limit.
- char *name[SOLR_MAX_PARAMETERS], *value[SOLR_MAX_PARAMETERS];
+ char **name, **value;
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,
+ int no_parms = 20; /* safe upper limit of args without extra_args */
+ Z_SRW_extra_arg *ea;
+
+ for (ea = srw_pdu->extra_args; ea; ea = ea->next)
+ no_parms++;
+ name = (char **) odr_malloc(encode, sizeof(*name) * no_parms);
+ value = (char **) odr_malloc(encode, sizeof(*value) * no_parms);
+
+ for (ea = srw_pdu->extra_args; ea; ea = ea->next)
+ {
+ name[i] = ea->name;
+ value[i] = ea->value;
+ i++;
+ }
+
+ z_HTTP_header_add_basic_auth(encode, &hreq->headers,
srw_pdu->username, srw_pdu->password);
if (srw_pdu->which == 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;
- }
+ yaz_add_name_value_str(encode, name, value, &i, "defType", "lucene");
+ yaz_add_name_value_str(encode, name, value, &i, "q", request->query);
if (srw_pdu->u.request->startRecord)
{
Odr_int start = *request->startRecord - 1;
yaz_add_name_value_str(encode, name, value, &i,
"fl", request->recordSchema);
+ 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;
return -1;
}
}
- else
- return -1;
-
- if (srw_pdu->extra_args)
+ else if (srw_pdu->which == Z_SRW_scan_request)
{
- Z_SRW_extra_arg *ea = srw_pdu->extra_args;
- for (; ea && i < SOLR_MAX_PARAMETERS; ea = ea->next)
+ 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"))
{
- name[i] = ea->name;
- value[i] = ea->value;
- i++;
+ 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));
}
+ 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);
}
+ else
+ return -1;
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,
projects / yaz-moved-to-github.git / blobdiff