# Copyright 2021 Zilliz. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from typing import List, Dict
from towhee.serve.triton.bls.mock import mock_pb_util
[docs]class MockInferResult:
[docs] def __init__(self, response: 'InferenceResponse'):
self._response = response
def as_numpy(self, name):
for ts in self._response.output_tensors():
if ts.name() == name:
return ts.as_numpy()
return None
[docs]class CallbackWrapper:
[docs] def __init__(self, callback):
self._callback = callback
[docs] def __call__(self, response, err):
if response is None and err is None:
self._callback(None, None)
else:
self._callback(MockInferResult(response), err)
[docs]class MockTritonClient:
'''
Mock client
'''
[docs] def __init__(self, models: Dict):
self._models = models
def infer(self, model_name: str, inputs: List[MockInferInput]) -> 'InferResult':
if model_name not in self._models:
return None
request = mock_pb_util.MockInferenceRequest(
[mock_pb_util.MockTritonPythonBackendTensor(item.name(), item.data())
for item in inputs]
)
responses = self._models[model_name].execute([request])
assert len(responses) == 1
return MockInferResult(responses[0])
def start_stream(self, callback):
self._callback = CallbackWrapper(callback)
def async_stream_infer(self, model_name, inputs: List[MockInferInput]):
assert self._callback is not None
request = mock_pb_util.MockInferenceRequest(
[mock_pb_util.MockTritonPythonBackendTensor(item.name(), item.data())
for item in inputs],
self._callback
)
self._models[model_name].execute([request])
def stop_stream(self):
self._callback = None