def init_feature_vector_graph(vector, query_options, update_stats=False):
try:
from storey import SyncEmitSource
except ImportError as exc:
raise ImportError(f"storey not installed, use pip install storey, {exc}")
feature_set_objects, feature_set_fields = vector.parse_features(
offline=False, update_stats=update_stats
)
graph = _build_feature_vector_graph(
vector, feature_set_fields, feature_set_objects, query_options
)
graph.set_flow_source(SyncEmitSource())
server = create_graph_server(graph=graph, parameters={})
cache = ResourceCache()
index_columns = []
for featureset in feature_set_objects.values():
driver = get_online_target(featureset)
if not driver:
raise mlrun.errors.MLRunInvalidArgumentError(
f"resource {featureset.uri} does not have an online data target"
)
cache.cache_table(featureset.uri, driver.get_table_object())
for key in featureset.spec.entities.keys():
if not vector.spec.with_indexes and key not in index_columns:
index_columns.append(key)
server.init_states(context=None, namespace=None, resource_cache=cache)
server.init_object(None)
return graph, index_columns
def test_v2_mock():
host = create_graph_server(graph=RouterState())
host.graph.add_route(
"my", class_name=ModelTestingClass, model_path="", multiplier=100
)
host.init(None, globals())
logger.info(host.to_yaml())
resp = host.test("/v2/models/my/infer", testdata)
logger.info(f"resp: {resp}")
# expected: source (5) * multiplier (100)
assert resp["outputs"] == 5 * 100, f"wrong health response {resp}"
def init_featureset_graph(source, featureset, namespace, targets=None, return_df=True):
"""create storey ingestion graph/DAG from feature set object"""
cache = ResourceCache()
graph = featureset.spec.graph.copy()
# init targets (and table)
targets = targets or []
_add_data_states(
graph, cache, featureset, targets=targets, source=source, return_df=return_df,
)
server = create_graph_server(graph=graph, parameters={})
server.init(None, namespace, cache)
return graph
def init_feature_vector_graph(vector):
try:
from storey import Source
except ImportError as e:
raise ImportError(f"storey not installed, use pip install storey, {e}")
feature_set_objects, feature_set_fields = vector.parse_features()
graph = _build_feature_vector_graph(vector, feature_set_fields,
feature_set_objects)
graph.set_flow_source(Source())
server = create_graph_server(graph=graph, parameters={})
cache = ResourceCache()
for featureset in feature_set_objects.values():
driver = get_online_target(featureset)
cache.cache_table(featureset.uri, driver.get_table_object())
server.init(None, None, cache)
return graph
def init_feature_vector_graph(vector):
try:
from storey import SyncEmitSource
except ImportError as exc:
raise ImportError(f"storey not installed, use pip install storey, {exc}")
feature_set_objects, feature_set_fields = vector.parse_features(False)
graph = _build_feature_vector_graph(vector, feature_set_fields, feature_set_objects)
graph.set_flow_source(SyncEmitSource())
server = create_graph_server(graph=graph, parameters={})
cache = ResourceCache()
index_columns = []
for featureset in feature_set_objects.values():
driver = get_online_target(featureset)
cache.cache_table(featureset.uri, driver.get_table_object())
for key in featureset.spec.entities.keys():
if not vector.spec.with_indexes and key not in index_columns:
index_columns.append(key)
server.init(None, None, cache)
return graph, index_columns