def grpc_datadef_to_array(datadef: prediction_pb2.DefaultData) -> np.ndarray:
"""
Convert a SeldonMessage DefaultData to a numpy array.
Parameters
----------
datadef
SeldonMessage DefaultData
Returns
-------
A numpy array
"""
data_type = datadef.WhichOneof("data_oneof")
if data_type == "tensor":
if sys.version_info >= (3, 0):
sz = np.prod(datadef.tensor.shape) # get number of float64 entries
c = datadef.tensor.SerializeToString() # get bytes
# create array from packed entries which are at end of bytes - assumes same endianness
features = np.frombuffer(memoryview(
c[-(sz * 8):]), dtype=np.float64, count=sz, offset=0)
features = features.reshape(datadef.tensor.shape)
else:
# Python 2 version which is slower
features = np.array(datadef.tensor.values).reshape(
datadef.tensor.shape)
elif data_type == "ndarray":
features = np.array(datadef.ndarray)
elif data_type == "tftensor":
features = tf.make_ndarray(datadef.tftensor)
else:
features = np.array([])
return features
def generate_test_requests() -> List[SeldonMessage]:
max_value = 9999
requests = []
inputs = max_value * np.random.rand(1024)
requests.append(
SeldonMessage(data=DefaultData(
tensor=Tensor(shape=inputs.shape, values=inputs))))
return requests
def main():
model = NoJava()
# model = Baseline()
# model = ProtobufEncoding()
model = PayloadPassthrough()
model.load()
# payload = np.array([0.0, 1.1, 2.2, 3.3])
# prediction = model.predict(payload)
# message = SeldonMessage(
# data=DefaultData(tensor=Tensor(shape=[1], values=[0.0, 1.1, 2.2, 3.3]))
# )
# prediction = model.predict_raw(message)
for n in range(100000):
message = SeldonMessage(data=DefaultData(
tensor=Tensor(shape=[1], values=[0.0, 1.1, 2.2, 3.3])))
prediction = model.predict_raw(MessageToJson(message))
print(f"[PYTHON] Prediction was {prediction}")