Likes - Activity - Ajinkz - Rasa Community Forum

Responses from the Rasa Forum

Ajinkz Responses

Message Structure: The tracker store should also contain the actual user message.
Intent with Story: Well, I have an intent which is named “bienvenue” (welcome in French), with the following story:
```
## story_bienvenue
* bienvenue
- action_bienvenue
```
When the chat opens, it sends this to the bot:
```
{ "message" : "/bienvenue" }
```
Issues with Exporting Stories: I think this should be fixed in the latest core release. Do you mind updating to the most recent version on PyPI? There was an issue in the serialization that we fixed - I suspect that it is the same one.
User Experience: I just wanna say you are doing an amazing professional conversation, though I couldn’t help you.
Chatbot in New Languages: I am flattered knowing you believe I am working for Rasa; however, I am just a fan and an active community member. One of my first blogs was about the TensorFlow embedding and I used it to test out a bot in Bengali using Latin alphabets.

I also stitched another one (taking inspiration).
Real-Time Visualization: For real-time conversations or visualizing the training data, you can use the data from the tracker store to visualize on a graph. Libraries like this might help: Decision Tree Visualization.
Using Buttons: So buttons are used to give suggestions or possibly give the next set of actions which the user might ask while in conversation. Here is an example:
```
templates:
utter_greet:
- text: "Hey! How can I help you?"
buttons:
- title: "What can you do?"
  payload: '/bot_functions'
```
Quotes in Payload: I just tried it out without the quotes, and it works. Seems like the quotes are optional after all.
Monitoring by Rasa: Hello. What is the general level of monitoring by the Rasa company of the forum for helping answer tech issues? Just want to ensure we have the correct expectations.
Custom Action Example: From the custom action, you can get the user message with:
```
tracker.latest_message.text
```

Troubleshooting with Rasa Core: You can get this with:

from rasa_core.trackers import EventVerbosity
tracker = agent.tracker_store.get_or_create_tracker(s_id)
state = tracker.current_state(event_verbosity = EventVerbosity.ALL)
print(state)

Using Duckling: Please be aware that Duckling tries to extract as many entity types as possible without providing a ranking. For example, if you specify both number and time, it will extract two entities.
Training Data and Intents: It depends on the number of intents you have. I suggest having around 20 examples per intent as a good starting point.