Rasa x SQL_POOL_SIZE & SQL_MAX_OVERFLOW

Hi, we are running a rasa server in docker containers and use an independent DB as tracker store, but we always got alarmed that rasa takes up too many connections. I know I can set SQL_POOL_SIZE (default: 50) and SQL_MAX_OVERFLOW (default: 100), but none of them seems to be working when using docker...

Rasa x connect to cloud DB get error when SSL enabled

Hi, I’m trying to config a cloud PostgreSQL (in Azure) instead of localhost when deploy a rasa server by using docker-compose, here is my docker-compose database-credentials:

x-database-credentials: &database-credentials
  DB_HOST: "rasa-dev.postgres.database.azure.com"
  DB_PORT: "5432"
  DB_USER: ...

Issue with handling conditional slot logic and validations in the same action in Rasa 2.0

Thanks for your reply, Sam @SamS. Sorry, I didn’t make myself clear; I didn’t mean validate_slot function not update the tracker, but it shouldn’t validate the slots I set from previous actions and already in the tracker. What happens here is:

The bot makes an API call and gets the user name: name/...

Issue with handling conditional slot logic and validations in the same action in Rasa 2.0

Hi, Sam, thanks for your reply. I tried this solution, nearly working but still not fully meet my requirement; there’s probably misunderstanding in my code:

async def run(self, dispatcher, tracker, domain):
    events = await self.validate(dispatcher, tracker, domain)
    logger.info("event: {}".format(events))

Thanks for your reply

Hi, Sam, I understand what you are saying; but the problem is all my custom codes are running in prod and we have proved the stability after a long period of user testing process. I can’t repeat this testing process for any changes I made at the moment, so I really want to find a common and safe way to...

Hi, Sam, thanks for your quick reply, here’s my run() method:

def required_slots(self, tracker: Tracker) -> List[Text]:
    name = tracker.get_slot("name")
    name1 = tracker.get_slot("name1")
    if name and (not name1):
        return []
    elif name and name1:
        ...

It would be great if the rasa team can update the doc

since conditional logic is really important and I believe most of us use it a lot in the previous version of rasa, I’m suffering with a different problem:

def validate_name(self, value, dispatcher, tracker, domain):
    PERSON = tracker.get_slot("PERSON")
    if PERSON:
        dispatcher.utter_message(text="Thanks, {}".format(PERSON))
    return {"name": PERSON}

Introducing entity roles and groups

Wow, I love this new feature! Actually, I was trying to integrate Spacy dependency parsing component, and hopefully this entity roles and groups can make my life easier. I still need to relabel nearly all the entities in the training data though.

Solution for dependency parsing in entity extraction

Hi, everyone, I just got some texts, more specifically, a txt file with a lot of unlabeled data. I already got a pretty stable model; is there a way to use rasa NLU to process this txt file first, i.e., do intent classification and entity extraction, by doing this, all I need to do is to review these labels...

Training killed when using DIET config

Yes, I was monitoring the CPU and memory usage during training; I think that’s because CPU usage is too high.

Thanks for your suggestion

I removed both countvectorfeaturizer and unfortunately, it still not working:

- name: ConveRTFeaturizer

Hi, really nice to see the new DIET classifier

I got a problem when training DIET; my process was killed, because of extreme resource starvation, even if I only tried to use light config:

language: en
pipeline:
  - name: ConveRTTokenizer
  - name: ConveRTFeaturizer
  - name: CountVectorsFeaturizer
  - name: CountVectorsFeaturizer
    analyzer: char_wb
    min_ngram: 1
    max_ngram: 4

Custom Featurizer for finetuned BERT features based on SpaCy

Hi, Julian, thanks for your reply, I figured it out, it turns to be there’s no en_trf_xlnetbasecased_lg model in the default(rasa:latest) image. So I’ll need to rewrite a new dockerfile and build a custom image to run BERT models.

Custom response is not working in rasa 1.6

I solved this problem by using a custom action: dispatcher.utter_custom_json(data); still have no idea why the custom payload in the domain file not working in rasa 1.6 though, I made no change.

Connection between Rasa and Rasa X on docker

I am using rasa x on docker and there’s a problem when communicating with slack: rasa-production and rasa-worker use the same port 5005, when I add the endpoint to slack it returns ok, which means there’s reply from 5005; however, there’s no reply when I talk with the bot in slack, rasa x works fine, d...

Connection between Rasa and Rasa X on docker

Hi, I have an issue with the rasa x - share with the guest user function, conversation terminated when the tester is using the UI; I can see the story end up with an unexpected action_listen. I was thinking it’s the error in my story design, but when I use the talk to the bot function in rasa x, it...

Connection between Rasa and Rasa X on docker

Hi, again, I am now having another problem when trying to connect rasa docker service with slack channel; I followed the rasa doc, started the rasa x, and configured the API in slack as well as credentials.yml. My question is: what is the Request URL? I noticed there are rasa production and rasa worker URLs...

Thank you for your reply

I didn’t quiet understand how the rasa works in docker, I wrongly thought that there should be all the config, domain, NLU, stories files in the etc/rasa folder; in fact, rasa on docker only needs models and custom actions to run the production, the bot didn’t react because the pika is not working.