def test_trasformer_pytorch_logger(integration_test_url, project_name,
use_google_oauth):
merlin.set_url(integration_test_url, use_google_oauth=use_google_oauth)
merlin.set_project(project_name)
merlin.set_model("transformer-logger", ModelType.PYTORCH)
model_dir = "test/transformer"
undeploy_all_version()
resource_request = ResourceRequest(1, 1, "100m", "200Mi")
transformer = Transformer(
"gcr.io/kubeflow-ci/kfserving/image-transformer:latest",
resource_request=resource_request,
)
logger = Logger(
model=LoggerConfig(enabled=True, mode=LoggerMode.ALL),
transformer=LoggerConfig(enabled=True, mode=LoggerMode.ALL),
)
with merlin.new_model_version() as v:
merlin.log_pytorch_model(model_dir=model_dir)
endpoint = merlin.deploy(transformer=transformer, logger=logger)
assert endpoint.logger is not None
model_config = endpoint.logger.model
assert model_config is not None
assert model_config.enabled
assert model_config.mode == LoggerMode.ALL
transformer_config = endpoint.logger.transformer
assert transformer_config is not None
assert transformer_config.enabled
assert transformer_config.mode == LoggerMode.ALL
with open(os.path.join("test/transformer", "input.json"), "r") as f:
req = json.load(f)
sleep(5)
resp = requests.post(f"{endpoint.url}", json=req)
assert resp.status_code == 200
assert resp.json() is not None
assert len(resp.json()["predictions"]) == len(req["instances"])
model_endpoint = merlin.serve_traffic({endpoint: 100})
sleep(5)
resp = requests.post(f"{model_endpoint.url}", json=req)
assert resp.status_code == 200
assert resp.json() is not None
assert len(resp.json()["predictions"]) == len(req["instances"])
# Try to undeploy serving model version. It must be fail
with pytest.raises(Exception):
assert merlin.undeploy(v)
# Undeploy other running model version endpoints
undeploy_all_version()
def test_standard_transformer_with_env_vars():
transformer_config_path = os.path.join("test/transformer",
"feast_enricher.yaml")
resource = ResourceRequest(min_replica=1,
max_replica=2,
cpu_request="100m",
memory_request="128Mi")
transformer = StandardTransformer(
config_file=transformer_config_path,
enabled=True,
resource_request=resource,
env_vars={"MODEL_URL": "http://model.default"})
assert transformer.env_vars == {
'MODEL_URL':
"http://model.default",
'STANDARD_TRANSFORMER_CONFIG':
'{"transformerConfig": {"feast": [{"project": "merlin", "entities": [{"name": "merlin_test_driver_id", "valueType": "STRING", "jsonPath": "$.driver_id"}], "features": [{"name": "merlin_test_driver_features:test_int32", "valueType": "INT32", "defaultValue": "0"}, {"name": "merlin_test_driver_features:test_float", "valueType": "FLOAT", "defaultValue": "0.0"}, {"name": "merlin_test_driver_features:test_double", "valueType": "DOUBLE", "defaultValue": "0.0"}, {"name": "merlin_test_driver_features:test_string", "valueType": "STRING", "defaultValue": ""}]}]}}'
}
assert transformer.enabled
assert transformer.command is None
assert transformer.args is None
assert transformer.resource_request.min_replica == 1
assert transformer.resource_request.max_replica == 2
assert transformer.resource_request.cpu_request == "100m"
assert transformer.resource_request.memory_request == "128Mi"
assert transformer.transformer_type == TransformerType.STANDARD_TRANSFORMER
def __init__(self, env: client.Environment):
self._name = env.name
self._cluster = env.cluster
self._is_default = env.is_default
self._default_resource_request = ResourceRequest(
env.default_resource_request.min_replica,
env.default_resource_request.max_replica,
env.default_resource_request.cpu_request,
env.default_resource_request.memory_request)
def test_custom_model_with_artifact(integration_test_url, project_name,
use_google_oauth):
merlin.set_url(integration_test_url, use_google_oauth=use_google_oauth)
merlin.set_project(project_name)
merlin.set_model("custom-w-artifact", ModelType.CUSTOM)
undeploy_all_version()
resource_request = ResourceRequest(1, 1, "1", "1Gi")
model_dir = "test/custom-model"
BST_FILE = "model.bst"
iris = load_iris()
y = iris["target"]
X = iris["data"]
dtrain = xgb.DMatrix(X, label=y)
param = {
"max_depth": 6,
"eta": 0.1,
"silent": 1,
"nthread": 4,
"num_class": 10,
"objective": "multi:softmax",
}
xgb_model = xgb.train(params=param, dtrain=dtrain)
model_file = os.path.join((model_dir), BST_FILE)
xgb_model.save_model(model_file)
with merlin.new_model_version() as v:
v.log_custom_model(
image="ghcr.io/tiopramayudi/custom-predictor-go:v0.2",
model_dir=model_dir)
endpoint = merlin.deploy(v,
resource_request=resource_request,
env_vars={"MODEL_FILE_NAME": BST_FILE})
sleep(5)
resp = requests.post(f"{endpoint.url}", json=request_json)
assert resp.status_code == 200
assert resp.json() is not None
assert resp.json()["predictions"] is not None
model_endpoint = merlin.serve_traffic({endpoint: 100})
sleep(5)
resp = requests.post(f"{model_endpoint.url}", json=request_json)
assert resp.status_code == 200
assert resp.json() is not None
assert resp.json()["predictions"] is not None
# Try to undeploy serving model version. It must be fail
with pytest.raises(Exception):
assert merlin.undeploy(v)
# Undeploy other running model version endpoints
undeploy_all_version()
def test_resource_request_validate():
resource_request = ResourceRequest(1, 2, "100m", "128Mi")
resource_request.validate()
resource_request.min_replica = 10
with pytest.raises(
Exception,
match="Min replica must be less or equal to max replica"):
resource_request.validate()
def test_resource_request(integration_test_url, project_name,
use_google_oauth):
merlin.set_url(integration_test_url, use_google_oauth=use_google_oauth)
merlin.set_project(project_name)
merlin.set_model("resource-request", ModelType.XGBOOST)
model_dir = "test/xgboost-model"
BST_FILE = "model.bst"
envs = merlin.list_environment()
assert len(envs) >= 1
default_env = merlin.get_default_environment()
assert default_env is not None
undeploy_all_version()
with merlin.new_model_version() as v:
iris = load_iris()
y = iris['target']
X = iris['data']
dtrain = xgb.DMatrix(X, label=y)
param = {
'max_depth': 6,
'eta': 0.1,
'silent': 1,
'nthread': 4,
'num_class': 10,
'objective': 'multi:softmax'
}
xgb_model = xgb.train(params=param, dtrain=dtrain)
model_file = os.path.join(model_dir, BST_FILE)
xgb_model.save_model(model_file)
# Upload the serialized model to MLP
merlin.log_model(model_dir=model_dir)
resource_request = ResourceRequest(1, 1, "100m", "200Mi")
endpoint = merlin.deploy(v,
environment_name=default_env.name,
resource_request=resource_request)
sleep(5)
resp = requests.post(f"{endpoint.url}", json=request_json)
assert resp.status_code == 200
assert resp.json() is not None
assert len(resp.json()['predictions']) == len(request_json['instances'])
merlin.undeploy(v)
sleep(5)
resp = requests.post(f"{endpoint.url}", json=request_json)
assert resp.status_code == 404
def test_standard_transformer_with_env_vars():
transformer_config_path = os.path.join("test/transformer", "feast_standard_transformer.yaml")
resource = ResourceRequest(min_replica=1, max_replica=2, cpu_request="100m", memory_request="128Mi")
transformer = StandardTransformer(config_file=transformer_config_path,
enabled=True,
resource_request=resource,
env_vars={"MODEL_URL": "http://model.default"})
assert transformer.env_vars == {'MODEL_URL': "http://model.default", 'STANDARD_TRANSFORMER_CONFIG': '{"transformerConfig": {"feast": [{"entities": [{"name": "customer_id", "valueType": "STRING", "jsonPath": "$.customer_id"}], "features": [{"name": "total_booking_1w", "defaultValue": "0.0"}, {"name": "total_booking_4w", "defaultValue": "0.0"}]}, {"entities": [{"name": "merchant_id", "valueType": "STRING", "jsonPath": "$.merchant_id"}], "features": [{"name": "total_completed_order_1w", "defaultValue": "0.0"}, {"name": "avg_completed_order_1w", "defaultValue": "0.0"}]}]}}'}
assert transformer.enabled
assert transformer.command is None
assert transformer.args is None
assert transformer.resource_request.min_replica == 1
assert transformer.resource_request.max_replica == 2
assert transformer.resource_request.cpu_request == "100m"
assert transformer.resource_request.memory_request == "128Mi"
assert transformer.transformer_type == TransformerType.STANDARD_TRANSFORMER
def test_xgboost(integration_test_url, project_name, use_google_oauth):
merlin.set_url(integration_test_url, use_google_oauth=use_google_oauth)
merlin.set_project(project_name)
merlin.set_model("xgboost-sample", ModelType.XGBOOST)
model_dir = "test/xgboost-model"
BST_FILE = "model.bst"
undeploy_all_version()
with merlin.new_model_version() as v:
iris = load_iris()
y = iris["target"]
X = iris["data"]
dtrain = xgb.DMatrix(X, label=y)
param = {
"max_depth": 6,
"eta": 0.1,
"silent": 1,
"nthread": 4,
"num_class": 10,
"objective": "multi:softmax",
}
xgb_model = xgb.train(params=param, dtrain=dtrain)
model_file = os.path.join(model_dir, BST_FILE)
xgb_model.save_model(model_file)
# Upload the serialized model to MLP
merlin.log_model(model_dir=model_dir)
resource_request = ResourceRequest(1, 1, "100m", "200Mi")
endpoint = merlin.deploy(v, resource_request=resource_request)
resp = requests.post(f"{endpoint.url}", json=request_json)
assert resp.status_code == 200
assert resp.json() is not None
assert len(resp.json()["predictions"]) == len(request_json["instances"])
merlin.undeploy(v)
def test_custom_model_without_artifact(integration_test_url, project_name,
use_google_oauth):
merlin.set_url(integration_test_url, use_google_oauth=use_google_oauth)
merlin.set_project(project_name)
merlin.set_model("custom-wo-artifact", ModelType.CUSTOM)
undeploy_all_version()
resource_request = ResourceRequest(1, 1, "1", "1Gi")
with merlin.new_model_version() as v:
v.log_custom_model(image="ghcr.io/tiopramayudi/custom-predictor:v0.2")
endpoint = merlin.deploy(v, resource_request=resource_request)
with open(os.path.join("test/custom-model", "input.json"), "r") as f:
req = json.load(f)
sleep(5)
resp = requests.post(f"{endpoint.url}", json=req)
assert resp.status_code == 200
assert resp.json() is not None
assert resp.json()["predictions"] is not None
model_endpoint = merlin.serve_traffic({endpoint: 100})
sleep(5)
resp = requests.post(f"{model_endpoint.url}", json=req)
assert resp.status_code == 200
assert resp.json() is not None
assert resp.json()["predictions"] is not None
# Try to undeploy serving model version. It must be fail
with pytest.raises(Exception):
assert merlin.undeploy(v)
# Undeploy other running model version endpoints
undeploy_all_version()