1 /* This file is part of the YAZ toolkit.
2 * Copyright (C) 1995-2011 Index Data
3 * See the file LICENSE for details.
7 * \brief Implements SOAP Webservice decoding/encoding
16 #include <yaz/matchstr.h>
17 #include <yaz/yaz-iconv.h>
19 #include <yaz/facet.h>
20 #include <yaz/wrbuf.h>
24 #define SOLR_MAX_PARAMETERS 100
27 #include <libxml/parser.h>
28 #include <libxml/tree.h>
30 const char *xml_node_attribute_value_get(xmlNodePtr ptr, const char *node_name, const char *attribute_name) {
32 struct _xmlAttr *attr;
33 // check if the node name matches
34 if (strcmp((const char*) ptr->name, node_name))
36 // check if the attribute name and return the value
37 for (attr = ptr->properties; attr; attr = attr->next)
38 if (attr->children && attr->children->type == XML_TEXT_NODE) {
39 if (!strcmp((const char *) attr->name, attribute_name))
40 return (const char *) attr->children->content;
46 static int match_xml_node_attribute(xmlNodePtr ptr, const char *node_name, const char *attribute_name, const char *value)
48 const char *attribute_value;
49 // check if the node name matches
50 if (strcmp((const char*) ptr->name, node_name))
53 attribute_value = xml_node_attribute_value_get(ptr, node_name, attribute_name);
54 if (attribute_value && !strcmp(attribute_value, value))
57 else /* No attribute to check */
62 static void yaz_solr_decode_result_docs(ODR o, xmlNodePtr ptr, Odr_int start, Z_SRW_searchRetrieveResponse *sr) {
68 for (node = ptr->children; node; node = node->next)
69 if (node->type == XML_ELEMENT_NODE)
72 sr->records = odr_malloc(o, sizeof(*sr->records) * sr->num_records);
74 for (node = ptr->children; node; node = node->next)
76 if (node->type == XML_ELEMENT_NODE)
78 Z_SRW_record *record = sr->records + i;
79 xmlBufferPtr buf = xmlBufferCreate();
80 xmlNode *tmp = xmlCopyNode(node, 1);
82 xmlNodeDump(buf, tmp->doc, tmp, 0, 0);
86 record->recordSchema = 0;
87 record->recordPacking = Z_SRW_recordPacking_XML;
88 record->recordData_len = buf->use;
89 record->recordData_buf = odr_malloc(o, buf->use + 1);
90 memcpy(record->recordData_buf, buf->content, buf->use);
91 record->recordData_buf[buf->use] = '\0';
92 // TODO Solve the real problem in zoom-sru, that doesnt work with 0-based indexes.
93 // Work-around: Making the recordPosition 1-based.
94 record->recordPosition = odr_intdup(o, start + offset + 1);
104 static int yaz_solr_decode_result(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveResponse *sr) {
106 struct _xmlAttr *attr;
107 for (attr = ptr->properties; attr; attr = attr->next)
108 if (attr->children && attr->children->type == XML_TEXT_NODE) {
109 if (!strcmp((const char *) attr->name, "numFound")) {
110 sr->numberOfRecords = odr_intdup(o, odr_atoi(
111 (const char *) attr->children->content));
113 else if (!strcmp((const char *) attr->name, "start")) {
114 start = odr_atoi((const char *) attr->children->content);
117 if (sr->numberOfRecords && *sr->numberOfRecords > 0)
118 yaz_solr_decode_result_docs(o, ptr, start, sr);
119 if (sr->numberOfRecords)
124 static const char *get_facet_term_count(xmlNodePtr node, int *freq) {
126 const char *term = xml_node_attribute_value_get(node, "int", "name");
128 WRBUF wrbuf = wrbuf_alloc();
132 for (child = node->children; child ; child = child->next) {
133 if (child->type == XML_TEXT_NODE)
134 wrbuf_puts(wrbuf, (const char *) child->content);
136 *freq = atoi(wrbuf_cstr(wrbuf));
137 wrbuf_destroy(wrbuf);
141 Z_FacetField *yaz_solr_decode_facet_field(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveResponse *sr)
144 Z_AttributeList *list;
145 Z_FacetField *facet_field;
150 const char* name = xml_node_attribute_value_get(ptr, "lst", "name");
151 char *pos = strstr(name, "_exact");
156 list = yaz_use_atttribute_create(o, name);
157 for (node = ptr->children; node; node = node->next) {
160 facet_field = facet_field_create(o, list, num_terms);
162 for (node = ptr->children; node; node = node->next) {
164 const char *term = get_facet_term_count(node, &count);
165 facet_field_term_set(o, facet_field, facet_term_create(o, term_create(o, term), count), index);
171 static int yaz_solr_decode_facet_counts(ODR o, xmlNodePtr root, Z_SRW_searchRetrieveResponse *sr) {
173 for (ptr = root->children; ptr; ptr = ptr->next)
175 if (match_xml_node_attribute(ptr, "lst", "name", "facet_fields"))
178 Z_FacetList *facet_list;
180 for (node = ptr->children; node; node= node->next)
184 facet_list = facet_list_create(o, num_facets);
186 for (node = ptr->children; node; node= node->next)
188 facet_list_field_set(o, facet_list, yaz_solr_decode_facet_field(o, node, sr), num_facets);
191 sr->facetList = facet_list;
200 int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup)
203 const char *content_buf = hres->content_buf;
204 int content_len = hres->content_len;
205 xmlDocPtr doc = xmlParseMemory(content_buf, content_len);
208 Z_SRW_PDU *pdu = yaz_srw_get(o, Z_SRW_searchRetrieve_response);
209 Z_SRW_searchRetrieveResponse *sr = pdu->u.response;
217 xmlNodePtr root = xmlDocGetRootElement(doc);
222 else if (strcmp((const char *) root->name, "response"))
228 /** look for result (required) and facets node (optional) */
231 for (ptr = root->children; ptr; ptr = ptr->next)
233 if (ptr->type == XML_ELEMENT_NODE &&
234 !strcmp((const char *) ptr->name, "result"))
235 rc_result = yaz_solr_decode_result(o, ptr, sr);
236 /* TODO The check on hits is a work-around to avoid garbled facets on zero results from the SOLR server.
237 * The work-around works because the results is before the facets in the xml. */
238 if (rc_result == 0 && *sr->numberOfRecords > 0 && match_xml_node_attribute(ptr, "lst", "name", "facet_counts"))
239 rc_facets = yaz_solr_decode_facet_counts(o, ptr, sr);
241 ret = rc_result + rc_facets;
254 static void yaz_solr_encode_facet_field(ODR encode, char **name, char **value, int *i, Z_FacetField *facet_field, int *limit) {
255 Z_AttributeList *attribute_list = facet_field->attributes;
256 struct yaz_facet_attr attr_values;
257 yaz_facet_attr_init(&attr_values);
258 yaz_facet_attr_get_z_attributes(attribute_list, &attr_values);
259 // TODO do we want to support server decided
260 if (!attr_values.errcode && attr_values.useattr) {
261 WRBUF wrbuf = wrbuf_alloc();
262 wrbuf_puts(wrbuf, (char *) attr_values.useattr);
263 /* Skip date field */
264 if (strcmp("date", attr_values.useattr) != 0)
265 wrbuf_puts(wrbuf, "_exact");
266 yaz_add_name_value_str(encode, name, value, i, "facet.field", odr_strdup(encode, wrbuf_cstr(wrbuf)));
267 if (attr_values.limit > 0) {
268 WRBUF wrbuf2 = wrbuf_alloc();
270 wrbuf_puts(wrbuf2, "f.");
271 wrbuf_puts(wrbuf2, wrbuf_cstr(wrbuf));
272 wrbuf_puts(wrbuf2, ".facet.limit");
273 olimit = attr_values.limit;
274 yaz_add_name_value_int(encode, name, value, i, odr_strdup(encode, wrbuf_cstr(wrbuf2)), &olimit);
275 wrbuf_destroy(wrbuf2);
277 wrbuf_destroy(wrbuf);
281 static void yaz_solr_encode_facet_list(ODR encode, char **name, char **value, int *i, Z_FacetList *facet_list, int *limit) {
284 for (index = 0; index < facet_list->num; index++) {
285 yaz_solr_encode_facet_field(encode, name, value, i, facet_list->elements[index], limit);
291 int yaz_solr_encode_request(Z_HTTP_Request *hreq, Z_SRW_PDU *srw_pdu,
292 ODR encode, const char *charset)
294 const char *solr_op = 0;
295 //TODO Change. not a nice hard coded, unchecked limit.
296 char *name[SOLR_MAX_PARAMETERS], *value[SOLR_MAX_PARAMETERS];
301 z_HTTP_header_add_basic_auth(encode, &hreq->headers,
302 srw_pdu->username, srw_pdu->password);
304 switch (srw_pdu->which)
306 case Z_SRW_searchRetrieve_request: {
307 Z_SRW_searchRetrieveRequest *request = srw_pdu->u.request;
309 switch(srw_pdu->u.request->query_type)
311 case Z_SRW_query_type_pqf:
312 yaz_add_name_value_str(encode, name, value, &i,
313 "q", request->query.pqf);
315 case Z_SRW_query_type_cql:
316 yaz_add_name_value_str(encode, name, value, &i,
317 "q", request->query.cql);
322 if (srw_pdu->u.request->startRecord)
324 Odr_int start = *request->startRecord - 1;
325 yaz_add_name_value_int(encode, name, value, &i,
328 yaz_add_name_value_int(encode, name, value, &i,
329 "rows", request->maximumRecords);
330 yaz_add_name_value_str(encode, name, value, &i,
331 "fl", request->recordSchema);
333 if (request->facetList) {
334 Z_FacetList *facet_list = request->facetList;
336 yaz_add_name_value_str(encode, name, value, &i, "facet", "true");
337 yaz_add_name_value_str(encode, name, value, &i, "facet.mincount", "1");
338 yaz_solr_encode_facet_list(encode, name, value, &i, facet_list, &limit);
341 yaz_add_name_value_int(encode, name, value, &i, "facet.limit", &olimit);
351 yaz_array_to_uri(&uri_args, encode, name, value);
353 hreq->method = "GET";
356 odr_malloc(encode, strlen(hreq->path) +
357 strlen(uri_args) + strlen(solr_op) + 4);
359 sprintf(path, "%s/%s?%s", hreq->path, solr_op, uri_args);
362 z_HTTP_header_add_content_type(encode, &hreq->headers,
363 "text/xml", charset);
371 * c-file-style: "Stroustrup"
372 * indent-tabs-mode: nil
374 * vim: shiftwidth=4 tabstop=8 expandtab