X-Git-Url: http://git.indexdata.com/?p=yaz-moved-to-github.git;a=blobdiff_plain;f=src%2Fsolr.c;h=3458a0b9037c7b8e86e8f6e3b3dc0ca4fa0180c8;hp=4e0a74a4683bc88a066ebdb18f8aa11d4ee01c13;hb=1fbd038d97d6e1bfe97270ecc17cca123bcc435c;hpb=93cf5a348dd93d545c30ffa2f2b1aa4bfee62cbb diff --git a/src/solr.c b/src/solr.c index 4e0a74a..3458a0b 100644 --- a/src/solr.c +++ b/src/solr.c @@ -1,11 +1,14 @@ /* This file is part of the YAZ toolkit. - * Copyright (C) 1995-2010 Index Data + * Copyright (C) 1995-2011 Index Data * See the file LICENSE for details. */ /** - * \file srwutil.c - * \brief Implements SRW/SRU utilities. + * \file solr.c + * \brief Implements SOAP Webservice decoding/encoding */ +#if HAVE_CONFIG_H +#include +#endif #include #include @@ -14,38 +17,28 @@ #include #include #include +#include #include "sru-p.h" +#define SOLR_MAX_PARAMETERS 100 + #if YAZ_HAVE_XML2 #include #include - -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)) + 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; } @@ -92,7 +85,7 @@ static void yaz_solr_decode_result_docs(ODR o, xmlNodePtr ptr, Odr_int start, Z_ } } -static void 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) @@ -100,33 +93,51 @@ static void yaz_solr_decode_result(ODR o, xmlNodePtr ptr, Z_SRW_searchRetrieveRe 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); } } - yaz_solr_decode_result_docs(o, ptr, start, sr); + if (sr->numberOfRecords && *sr->numberOfRecords > 0) + yaz_solr_decode_result_docs(o, ptr, start, sr); + if (sr->numberOfRecords) + return 0; + return -1; } -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) { + const char *term = yaz_element_attribute_value_get(node, "int", "name"); + xmlNodePtr child; + WRBUF wrbuf = wrbuf_alloc(); + if (!term) + return term; -static const char *get_facet_term_count(xmlNodePtr node, int *freq) { - // TODO implement - return 0; + for (child = node->children; child ; child = child->next) { + if (child->type == XML_TEXT_NODE) + wrbuf_puts(wrbuf, (const char *) child->content); + } + *freq = atoi(wrbuf_cstr(wrbuf)); + wrbuf_destroy(wrbuf); + return term; } 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_AttributeList *list; Z_FacetField *facet_field; int num_terms = 0; int index = 0; xmlNodePtr node; + // USE attribute + const char* name = yaz_element_attribute_value_get(ptr, "lst", "name"); + char *pos = strstr(name, "_exact"); + /* HACK */ + if (pos) { + pos[0] = 0; + } + list = yaz_use_attribute_create(o, name); for (node = ptr->children; node; node = node->next) { num_terms++; } @@ -141,7 +152,7 @@ Z_FacetField *yaz_solr_decode_facet_field(ODR o, xmlNodePtr ptr, Z_SRW_searchRet return facet_field; } -static void 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) { @@ -161,16 +172,13 @@ static void yaz_solr_decode_facet_counts(ODR o, xmlNodePtr root, Z_SRW_searchRet facet_list_field_set(o, facet_list, yaz_solr_decode_facet_field(o, node, sr), num_facets); num_facets++; } - sr->facet_list = facet_list; + sr->facetList = facet_list; break; } } + return 0; } -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) @@ -201,19 +209,20 @@ int yaz_solr_decode_response(ODR o, Z_HTTP_Response *hres, Z_SRW_PDU **pdup) } else { - /** look for result node */ + /** look for result (required) and facets node (optional) */ + int rc_result = -1; + int rc_facets = 0; for (ptr = root->children; ptr; ptr = ptr->next) { if (ptr->type == XML_ELEMENT_NODE && !strcmp((const char *) ptr->name, "result")) - 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) - { - ret = -1; + rc_result = yaz_solr_decode_result(o, ptr, sr); + /* 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); } + ret = rc_result + rc_facets; } } if (doc) @@ -226,17 +235,49 @@ 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) { +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 void yaz_solr_encode_facet_list(ODR encode, char **name, char **value, int *i, Z_FacetList *facet_list, int *limit) { + + int index; + for (index = 0; index < facet_list->num; index++) { + yaz_solr_encode_facet_field(encode, name, value, i, facet_list->elements[index], limit); + } } + int yaz_solr_encode_request(Z_HTTP_Request *hreq, Z_SRW_PDU *srw_pdu, ODR encode, const char *charset) { const char *solr_op = 0; //TODO Change. not a nice hard coded, unchecked limit. - int max = 100; - char *name[100], *value[100]; + char *name[SOLR_MAX_PARAMETERS], *value[SOLR_MAX_PARAMETERS]; char *uri_args; char *path; int i = 0; @@ -273,14 +314,17 @@ 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->facet_list) { - int index; - Z_FacetList *facet_list = request->facet_list; + if (request->facetList) { + Z_FacetList *facet_list = request->facetList; + int limit = 0; 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]); - } + 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); + */ + } break; }