abhishek1
(Abhishek )
September 10, 2019, 8:11am
1
Can anyone tell me how to make multiple agent bots with single rasa project.
I want to run two bot weather bots and stock market bot at a time with two different user and it should interrupt each other at a time.
Conversation flow should be different
1 Like
ganeshv
(Ganesh Iyer)
September 10, 2019, 8:16am
2
Could you give an example of a conversation? Not sure I’m able to picture it correctly.
1 Like
abhishek1
(Abhishek )
September 10, 2019, 8:37am
3
I want to make two different chatbot with single rasa project?
Is it possible ?
If no then How can we run rasa server with different port and action.py too with different port
Manthan
(Manthan)
September 10, 2019, 8:55am
4
Hello @abhishek1 you can use --port 80 at the time of run like
action server:
python3 -m rasa run actions --actions actions --debug --port 5056
Rasa run:
sudo python3 -m rasa run --enable-api --cors “*” --debug --port 80
capgos17
(Ganesh)
September 10, 2019, 8:58am
5
run the different bots in different ports
abhishek1
(Abhishek )
September 10, 2019, 9:24am
6
I did it now
How can we use different bots agent with single server and port
abhishek1
(Abhishek )
September 10, 2019, 9:26am
7
Can you please tell me How can we run multiple bots with single project
Like for bank and hotel two different bot with single rasa project
Ya you can run 2 dialogue model in one server by creating 2 agents and this 2 agents can have different dialogue flow
1 Like
abhishek1
(Abhishek )
September 10, 2019, 10:49am
9
How can we do that can you pleas show an example
cr33dx
(Shashwat)
September 10, 2019, 4:03pm
10
Maybe u can use rasa docker to run 2 different instances of bot.
anurags
(Anurag)
November 28, 2019, 2:09pm
11
Just create separate agents for each model. Use separate tracker_store objects, separate interpreter objects and separate models of course.
Then write a function that returns the correct agent for your use case.
Like this:
agent = Agent.load('./models/dialogue', interpreter=custom_interpreter, action_endpoint=_action_endpoint, tracker_store=custom_tracker_store)
agent2 = Agent.load('./models/dialogue2', interpreter=custom_interpreter2, action_endpoint=_action_endpoint2, tracker_store=custom_tracker_store2)
def foo(integer):
if integer == 0:
return agent
else:
return agent2
Hope this helps if you haven’t figured out already.
@Abhishek did you succeed in implementing a multi-bot set–up?
Would you be willing to share what was your solution?