Multiple Agents and Routing

You are right, I commented it out and I am still getting a 404 error in the browser visiting the url that is spits out. The server is running at http://localhost:5002/login?username=me&password=ekpKs7OPGNa4

Running Rasa X inside of the moodbot example gives me this stacktrace. Process SpawnProcess-1: Traceback (most recent call last): File "/usr/local/Cellar/python/3.6.5_1/Frameworks/Python.framework/Versions/3.6/lib/python3.6/multiprocessing/process.py", line 258, in _bootstrap self.run()

Let me know if this helps you.

You can create different agents and specify which agent you want to send it to, but you need to create your own server and load a different agent per route.

• http://localhost:5003/parse?q=hi&agent=resturant&version=1
• http://localhost:5003/parse?q=hi&agent=resturant&version=2

Running this causes this output: /usr/local/lib/python3.6/site-packages/pykwalify/core.py:99: UnsafeLoaderWarning:

The default ‘Loader’ for ‘load(stream)’ without further arguments can be unsafe.

That is exactly what I have already, could I see what the output would look like?

can you please give me an example for both command line and the python file way of doing that?

AWESOME Thank you!

Generated Story -195929544701750116

• caller_choice_personal_loan
• utter_confirm_personal_loan
• confirm
• utter_calls_are_recorded
• confirm
• action_personal_loan_form
• slot{"requested_slot": "caller_last_name"}
• inform
• action_personal_loan_form

You must add 'action_purchase' to your domain file under actions: Then you must retrain the dialogue model.

Yes, create your own server, and load different agents from different directories. Then create a custom tracker store. You can look at these on the GitHub: Tracker Store Implementation

So first you need to have a trained dialogue model and a trained NLU model. This runs the rasa_core run script.

If I run all of this in a python terminal, it loads the correct tracker from the mongo tracker store.

The tracker store class is meant to be subclassed and not used as is. So it must be Redis or Mongo or InMemory.

Hey Souvik,

So in theory I could do this. Create a tracker with a sender_id. Save that tracker in the database. Create two agents. Instantiate those two agents with the same tracker. When a message comes in, with a sender_id, I pull that tracker from the database and parse which agent to route to.

Could you give me a little more context on what you mean by custom web channel? Also, RedisTrackerStore subclasses TrackerStore. And TrackerStore returns a DialogueStateTracker when it is initialized.

The uuid is for logging and the string after conversations//say is what they are using for the sender_id.

My guess is that when we create a tracker store and save it to the databases, Mongo or Redis, when we load the agent, we can send it messages with a sender_id.

So far for me it has, I have tested having two different conversations at one time and it handles it quite well.