def estimate(messages, inference_type):
if not isinstance(messages, list):
messages = [messages]
instances = list(map(lambda message: json.loads(message), messages))
source_ids = list(
map(lambda instance: instance.pop('source_id'), instances))
source_timestamps = list(
map(lambda instance: instance.pop('source_timestamp'), instances))
if inference_type == 'local':
estimated_weights = inference.estimate_local(instances)
elif inference_type == 'cmle':
estimated_weights = inference.estimate_cmle(instances)
else:
estimated_weights = 'NA'
for i in range(len(instances)):
instance = instances[i]
instance['estimated_weight'] = estimated_weights[i]
instance['source_id'] = source_ids[i]
instance['source_timestamp'] = source_timestamps[i]
instance['predict_timestamp'] = datetime.utcnow().strftime(
'%Y-%m-%d %H:%M:%S')
logging.info(instances)
return instances
def estimate(instance, inference_type):
"""
Estimate the baby weight given an instance
If inference type is 'cmle', the model API deployed on Coud ML Engine is called,
otherwise, the local model is called
Agrs:
instance: dictionary of values representing the input features of the instance
inference_type: cane be 'local or 'cmle'
Returns:
int - baby weight estimate from the model
"""
# pop the actual weight if exists
weight_pounds = instance.pop('weight_pounds', 'NA')
if inference_type == 'local':
estimated_weights = inference.estimate_local([instance])
elif inference_type == 'cmle':
estimated_weights = inference.estimate_cmle([instance])
else:
estimated_weights = 'NA'
instance['estimated_weight'] = estimated_weights[0]
instance['weight_pounds'] = weight_pounds[0]
return instance
def estimate(messages, inference_type):
if not isinstance(messages, list):
messages = [messages]
instances = list(
map(lambda message: json.loads(message),
messages)
)
source_ids = list(
map(lambda instance: instance.pop('source_id'),
instances)
)
source_timestamps = list(
map(lambda instance: instance.pop('source_timestamp'),
instances)
)
if inference_type == 'local':
estimated_weights = inference.estimate_local(instances)
elif inference_type == 'cmle':
estimated_weights = inference.estimate_cmle(instances)
else:
estimated_weights = 'NA'
for i in range(len(instances)):
instance = instances[i]
instance['estimated_weight'] = estimated_weights[i]
instance['source_id'] = source_ids[i]
instance['source_timestamp'] = source_timestamps[i]
instance['predict_timestamp'] = datetime.utcnow().strftime('%Y-%m-%d %H:%M:%S')
logging.info(instances)
return instances
'plurality': 1.0,
'gestation_weeks': 39,
'mother_married': 'True',
'cigarette_use': 'False',
'alcohol_use': 'False'
}
]
print("")
print("Inference Type:{}".format(INFERENCE_TYPE))
print("")
time_start = datetime.utcnow()
print("Inference started at {}".format(time_start.strftime("%H:%M:%S")))
print(".......................................")
for i in range(10):
if INFERENCE_TYPE == 'local':
output = inference.estimate_local(instances)
else:
output = inference.estimate_cmle(instances)
print(output)
time_end = datetime.utcnow()
print(".......................................")
print("Inference finished at {}".format(time_end.strftime("%H:%M:%S")))
print("")
time_elapsed = time_end - time_start
print("Inference elapsed time: {} seconds".format(time_elapsed.total_seconds()))
'plurality': 1.0,
'gestation_weeks': 39,
'mother_married': 'True',
'cigarette_use': 'False',
'alcohol_use': 'False'
}
]
print("")
print("Inference Type:{}".format(INFERENCE_TYPE))
print("")
time_start = datetime.utcnow()
print("Inference started at {}".format(time_start.strftime("%H:%M:%S")))
print(".......................................")
for i in range(10):
if INFERENCE_TYPE == 'local':
output = inference.estimate_local(instances)
else:
output = inference.estimate_cmle(instances)
print(output)
time_end = datetime.utcnow()
print(".......................................")
print("Inference finished at {}".format(time_end.strftime("%H:%M:%S")))
print("")
time_elapsed = time_end - time_start
print("Inference elapsed time: {} seconds".format(time_elapsed.total_seconds()))