Hello, I’ve been eager to try the MultiProjectImporter so I can have a chatbot which functionalities can be enabled or disable easily based on the client needs. I’ve read some of the forum posts here, there are a user who said that one agent is enough to handle a lot of intents at the same time and there is also another one that said putting everything on single instance will have performance issues. For my case, I’m planning to try both with the objective of having a chatbot module that users can easily enable or disable the functionality of the chatbot. The former seems like can be achieved by MultiProjectImporter and having a UI for user to choose which functionalities to include in the root config.yml, while the latter needs to have multiple agents with one agent that decide which agent to choose for the response and seems like having those multiple agents with single tracker store is one of the solution.
This time, I would like to try the former first. Using the folder structure like in the documentations, combining the training data and domain seems to be working perfectly, however the rasa run actions command doesn’t seem to run all of the possible actions in the projects folder. Is there any solution for this kind of setting?