Formatting Bot Response with Newline etc

I looked for it in GitHub but couldn’t find one, so I must have forgotten to do it ¯_(ツ)_/¯

When you download Duckling from GitHub, there is no program, only source code. You are going to compile the source code with Haskell. You will need to run these two commands from your terminal, as it says in the Readme on GitHub.

$ stack build
$ stack exec duckling-example-exe

“Illegal instruction (core dumped)” indicates that some C code crashed. You could try do debug that using PDB, but in my experience, you are heading a long road that might lead to nowhere.

TensorFlow library was compiled to use AVX instructions, but these aren’t available on your machine.

There are a lot of dependencies what are easier to install and use on Linux. I think you will have a lot of issues when developing in Windows. Should you feel adventurous then you can try Cygwin or Ubuntu on Windows.

You need to install wget first, since it is not native to Windows. I would recommend that you use Linux instead. Here is an introduction.

This tool is absolutely necessary, and it should be running whenever training happens. TBH it should not even be possible to write new stories in Rasa X without validation from this tool. When multiple members of a team are working together it is very easy for one to accidentally create conflicting …

Do you restart your session when you try the second time?

Are you handling the datetime with an entity extractor? E.g. Duckling? I’m seeing a normal utter-response, but you want to use a custom action for dynamic responses when handling time-dependent questions.

Edit: I realised that the responses are saved in a .yaml file, so that’s it. TL:DR, for new lines use \n in custom actions and the yaml markdown formatting in Rasa X. But it might not work for everything.

I’m still looking for the documentation of the syntax for editing responses. Anyone?

I cannot give you the data you are looking for, but there is an NLU sharing initiative by Rasa. Hopefully they will be accepting more languages soon.

The config is rather opaque and complicated. It is something I set up and forget about. I doubt that I will ever want/dare to change the setup from the default components of supervised_embeddings. I just think “how do I add x feature?”, in this case Duckling, so for me the ideal behavior would be so…

I see. Do I have to monitor this page to keep the configuration up to date?

Here are the default components that make up the supervised_embeddings pipeline

Edit: \n works! but \n\n does not work! I’m submitting a bug report later… Yes, from that tab the formatting is the same. Line-break is missing.

Is the same possible in custom actions? E.g.:

actions.py
def run(self, dispatcher: CollectingDispatcher, tracker: Tracker, domain: Dict[Text, Any]) -> List[Dict[Text, Any]]:
#[...] some logic
message = f"hello {username}! \n\n" #\n does not produce new line
message = message + f" today…"

If you are experienced with Docker then use this: Rasa Duckling Docker Image If not, then pull Duckling from GitHub and install Haskell, and compile it yourself:

I just tested this out and it works. I am very pleased! This should make it much easier to develop good bots - especially if it is implemented in Rasa X. Looking forward to the release!

How do you test if stories are contradicting?

I previously tried to pip install psycopg2 but that threw another error, so I’m doing the “right” thing instead:

RUN pip install rasa-x[sql]==0.23.2 --extra-index-url https://pypi.rasa.com/simple

Regarding the official docker image, the instructions on Rasa X downloads and runs this ansible playbook, which I’m wary of.

What prevents you from switching to the official one? I only have one server available for my containerized services. Your official docker image requires a dedicated machine. The install modifies the host machine and messes with my setup.

JWT version first. I added ‘Bearer’ and get this response:

500 internal server error The nginx log does not give much...