now I am thinking about deployment
what I am thinking is put rasa-core, rasa-nlu, rasa-action as stand-alone service
in the front , it’s flask with jwt authentication
like this
@app.route('/chat')
@jwt_required
def chat():
#handle post request ,
#get user_id
#get message
#call rasa-core service
one option is use requests to post to rasa-core rest api, but that is not elegant
I see some interface like agent.load_from_server
is it possible to build a agent object , and use it to communicate with a remote rasa-core server?
currently I see all use-case of agent , is to load model from local folder , not a remote service
or can I use rasa-core as part of current flask application , like a blueprint?
any suggestions , thanks