Jasper Chatbot

A chatbot for my Twitch and Website that uses auto versioning via Commitizen and runs off Rasa.

Documentation

You can visit ReadTheDocs online or you can follow the steps below to generate them locally, most more in depth information is documented here.

Most things below plus more in-depth information is located at the docs above which are also built from the docs dir.

Building Docs Locally

You can cd into the docs dir and setup a python venv and install the requirements.txt file from this dir and then do make html to generate these, this is just using the typical sphinx setup nothing crazy.

Environment Setup

I recommend if running on windows to setup WSL2 and setup ubuntu on it. Then you can clone the repo to this setup and do code . with Visual Studio installed on your windows machine and edit the code,etc. If you are on a Mac some of this will differ depending on Rasa. Rasa up until recently had issues with silicon macs due to a tensorflow issue, this should have been resolved in the most recent version we are using.

First things you will need is ensure you have python 3.7 or higher installed, after having that you should be able to do the following to setup your virtual env:

python3 -m venv .venv

Now we can activate this virtual env and install our requirements.

Linux source .venv/bin/activate

Windows .venv\Scripts\activate

Once activated we can install with: pip install -r requirements.txt

This will install Rasa and Commitizen

Setting Up Action Server

For certain things to work that make external API calls you need to also run the action server, that can be done by using the same previous venv you already setup and activate it:

cd into the actions directory

source .venv/bin/activate

Then install the action server requirements:

pip install -r actions.py

There are also a few env vars/secrets used and we currently use Doppler for this but you can also just setup your own env vars via:

export NASA_API_KEY=XXX - This is the Nasa API where you can get your own key at https://api.nasa.gov/

export GITHUB_TOKEN=XXX - This is a GitHub API token that has access to read action builds and such, you can skip this if not building new intents or responses related to this.

For further information about not having to use these vars please see the docs at https://docs.doppler.com/docs/install-cli for more information and use https://doppler.com/join?invite=524473B9 to sign up for $100 free credit.

Now to run the action server you can do:

cd back to the main root of the repo cd ..

rasa run actions

If running with doppler would be:

doppler run -- rasa run actions

How To Run Locally

After performing the above you can run the bot locally via:

rasa run --cors '*' --debug

How To Deploy

This is deployed via the Rasa helm chart values file located in the rasa dir of DevOps With Brian's LKE Repo.

More info in that location can be found on deploying.