Rasa Webchat & Online Trainer

This project is a fork of the WebChat widget from MrBot-Ai. We developed a custom online trainer where you can train and build your bots stories. Furthermore this project could be served with docker in order to test the widget.

Structure

• docker contains all docker-compose files, there are several to start the webchat in different modes.
• nginx contains configuration for the http-server nginx.
• src contains the whole source code of the widget.
• static contains static files, like: index.html. Would be used to start the webchat in prod mode.
• static-online-training contains static files, like: index.html. Would be used to start the webchat in trainer mode.
• static-dev contains static files, like: index.html. Would be used for local development.

Other files or directories are from the main project, for detailed documentation on this code please search on MrBot-Ai.

Deploy and run the project

locally

To run this project locally you need to install node.js and npm as a package manager. To install all dependency please execute the following command. This could eventually throw some errors due to some tests that we will not fix.

npm install

Afterwards you can run the local development server by the command:

npm run dev

Build

To deploy and run this project docker is mandatory, you would need to install docker as well as docker stack or docker compose. The project would be deployed with docker build. It will be tagged with our registry name and the project name.

docker build -t docker.nexus.gpchatbot.archi-lab.io/chatbot/webchat .

Run

This project can be started in different modes. The modes differs in configuration and function.

• local local setup for docker, will run the default configuration of the project.

    docker-compose -f docker/docker-compose.yaml -f docker/docker-compose.local.yaml up -d

• prod production setup, will run the webchat widget for customers.

    docker-compose -f docker/docker-compose.yaml -f docker/docker-compose.prod.yaml up -d

• trainer trainer setup, will run the online trainer widget to train and build new stories.

    docker-compose -f docker/docker-compose.yaml -f docker/docker-compose.trainer.yaml up -d