Use python API to retrain nlu model, but cannot reload the lastest one

hi, im holding a nlu http server using django, and it works well, but everytime i retrain the nlu model, it won’t work unless i restart the http server (but when i use python -m rasa_nlu.server to run the http server, it can load the lastest model everytime). the code is as below.

interpreter=Interpreter.load('./train/models/default/faq', builder)

then i take a look at the rasa_nlu.server source code. i found that if i wrote like this as below, it will load lastest model.

data_router = DataRouter(
request_params = {'q': request.GET.get('q'), 'project': 'faq'}
data = data_router.extract(request_params)
res = data_router.parse(data)

so my question is , it that a reason to not use this in python API? and is this risk to write code like above? thanks in advance!!