def do_suggest(query_params, suggest_params):
search_params = suggest_params["search"]
range_result = _get_predicate_ranges(query_params, search_params)
if is_result_empty(range_result):
message = _(u"Either the predicate {0} does not exists or it does not have any rdfs:range defined in the triplestore")
message = message.format(search_params["target"])
raise HTTPError(400, message)
classes = _validate_class_restriction(query_params, range_result)
graphs = _validate_graph_restriction(query_params, range_result)
indexes = ["semantica." + uri_to_slug(graph) for graph in graphs]
search_fields = list(set(_get_search_fields(query_params, suggest_params) + LABEL_PROPERTIES))
response_params = suggest_params.get("response", {})
response_fields = _get_response_fields(
query_params,
response_params,
classes,
LABEL_PROPERTIES)
# request_body = _build_body_query(
# query_params,
# search_params,
# classes,
# search_fields,
# response_fields)
analyze_response = run_analyze(search_params["pattern"], settings.ES_ANALYZER, indexes)
tokens = analyze_response["tokens"]
request_body = _build_body_query_compatible_with_uatu_and_es_19_in_envs(
query_params,
tokens,
classes,
search_fields,
response_fields,
search_params["pattern"]
)
# Sorting in ES is done using memory. From the docs [1]:
# "When sorting, the relevant sorted field values are loaded into memory.
# This means that per shard, there should be enough memory to contain them"
# Currently Globo.com ES servers don't have enough memory to load all data
# During the 30th October 2013, a query using sort caused Split-brain and all ES shards were down.
# Therefore, we need to think twice and use sort cleverly to avoid problems
# [1] http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-request-sort.html
assert not "sort" in request_body # Read comments above
elasticsearch_result = run_search(request_body, indexes=indexes)
class_fields = response_params.get("class_fields", [])
total_items = elasticsearch_result["hits"]["total"]
if total_items:
items = _build_items(query_params, elasticsearch_result, LABEL_PROPERTIES, class_fields)
response = build_json(items, total_items, query_params)
else:
response = {}
return response
def do_suggest(query_params, suggest_params):
search_params = suggest_params["search"]
range_result = _get_predicate_ranges(query_params, search_params)
if is_result_empty(range_result):
message = _(
u"Either the predicate {0} does not exists or it does not have any rdfs:range defined in the triplestore"
)
message = message.format(search_params["target"])
raise HTTPError(400, message)
classes = _validate_class_restriction(query_params, range_result)
graphs = _validate_graph_restriction(query_params, range_result)
indexes = ["semantica." + uri_to_slug(graph) for graph in graphs]
search_fields = list(
set(
_get_search_fields(query_params, suggest_params) +
LABEL_PROPERTIES))
response_params = suggest_params.get("response", {})
response_fields = _get_response_fields(query_params, response_params,
classes, LABEL_PROPERTIES)
# request_body = _build_body_query(
# query_params,
# search_params,
# classes,
# search_fields,
# response_fields)
analyze_response = run_analyze(search_params["pattern"],
settings.ES_ANALYZER, indexes)
tokens = analyze_response["tokens"]
request_body = _build_body_query_compatible_with_uatu_and_es_19_in_envs(
query_params, tokens, classes, search_fields, response_fields,
search_params["pattern"])
# Sorting in ES is done using memory. From the docs [1]:
# "When sorting, the relevant sorted field values are loaded into memory.
# This means that per shard, there should be enough memory to contain them"
# Currently Globo.com ES servers don't have enough memory to load all data
# During the 30th October 2013, a query using sort caused Split-brain and all ES shards were down.
# Therefore, we need to think twice and use sort cleverly to avoid problems
# [1] http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-request-sort.html
assert not "sort" in request_body # Read comments above
elasticsearch_result = run_search(request_body, indexes=indexes)
class_fields = response_params.get("class_fields", [])
total_items = elasticsearch_result["hits"]["total"]
if total_items:
items = _build_items(query_params, elasticsearch_result,
LABEL_PROPERTIES, class_fields)
response = build_json(items, total_items, query_params)
else:
response = {}
return response
def delete_instance(query_params):
dependants_result_dict = query_dependants(query_params)
if not is_result_empty(dependants_result_dict):
values = [item['dependant']['value'] for item in dependants_result_dict['results']['bindings']]
str_values = ", ".join(values)
raise HTTPError(409, log_message=_(u"Cannot exclude instance because of the dependencies: {0}").format(str_values))
query_result_dict = query_delete(query_params)
if some_triples_deleted(query_result_dict, query_params['graph_uri']):
return True
def get_instance(query_params):
"""
Given a URI, verify that the type corresponds to the class being passed as a parameter
Retrieve all properties and objects of this URI (subject)
"""
if must_retrieve_graph_and_class_uri(query_params):
query_result_dict = get_class_and_graph(query_params)
bindings = query_result_dict['results']['bindings']
query_params["graph_uri"] = extract_graph_uri(bindings)
query_params["class_uri"] = extract_class_uri(bindings)
query_result_dict = query_all_properties_and_objects(query_params)
if is_result_empty(query_result_dict):
return None
else:
class_schema = get_class.get_cached_schema(query_params)
return assemble_instance_json(query_params, query_result_dict,
class_schema)
def delete_instance(query_params):
dependants_result_dict = query_dependants(query_params)
if not is_result_empty(dependants_result_dict):
values = [
item['dependant']['value']
for item in dependants_result_dict['results']['bindings']
]
str_values = ", ".join(values)
raise HTTPError(
409,
log_message=_(
u"Cannot exclude instance because of the dependencies: {0}").
format(str_values))
query_result_dict = query_delete(query_params)
if some_triples_deleted(query_result_dict, query_params['graph_uri']):
return True
def get_instance(query_params):
"""
Given a URI, verify that the type corresponds to the class being passed as a parameter
Retrieve all properties and objects of this URI (subject)
"""
if must_retrieve_graph_and_class_uri(query_params):
query_result_dict = get_class_and_graph(query_params)
bindings = query_result_dict['results']['bindings']
query_params["graph_uri"] = extract_graph_uri(bindings)
query_params["class_uri"] = extract_class_uri(bindings)
query_result_dict = query_all_properties_and_objects(query_params)
if is_result_empty(query_result_dict):
return None
else:
class_schema = get_class.get_cached_schema(query_params)
return assemble_instance_json(query_params,
query_result_dict,
class_schema)