Rasa Programming Tutorial

This repository contains a collection of tutorials that will help you understand the Rasa framework. Basic concepts of Rasa (e.g., Intent, Entity, Response, Action, Slot, Form) will be explained in a series of examples. Several chatbot interfaces are also included in this repository. Note that all the demos here are built and tested with Rasa 2.5.

This project is built upon the following repositories:

• https://github.com/RasaHQ/rasa-2.x-form-examples
• https://github.com/scalableminds/chatroom
• https://github.com/vishwaspuri/Interactive-Voice-Response

Steps to use

• Before starting the installation process, please ensure that you have conda/virtualenv installed.

• Firstly we install all the dependencies:

  1. Clone the repository

  git clone https://github.com/QingyuGuo/rasa-2.5-tutorial.git

  2. Change into project directory:

  cd rasa-2.5-tutorial/

  3. Create python environment

  conda create --name <environment-name> python=3.8 

  4. Activate environment

  conda activate <environment-name>

  5. Install dependencies

  pip install -r requirements.txt

Rasa Examples

0. Initial Project

It helps to understand basic structures of Rasa project.

Code can be found in the ./chatbot/00-rasa-init folder.

1. Custom Actions

It helps to understand custom actions. In this simple bot we show how to get the current time.

Code can be found in the ./chatbot/01-actions folder.

2. Slots

It helps to understand slots, the memory of the bot. In this simple bot we show how they work by keeping track of a users name.

Code can be found in the ./chatbot/02-slots folder.

3. Entities

It helps to understand how to extract entities from a sentence. In this simple bot we show how to recognize the name of the user from an utterance.

Code can be found in the ./chatbot/03-entities folder.

4. Forms

If we want to query multiple things from the user, it may be best to use forms instead of custom actions. Luckily for us, we can use our RulePolicy to help us out.

Code can be found in the ./chatbot/04-forms folder.

5. Wellness bot

It will be helpful to understand the concepts we have learned in the above examples. We build a bot to ask users status and goals, and could be further modified for providing activity recommendation.

Code can be found in the ./chatbot/05-wellness-bot folder.

Web Interface

1. Chatroom

The project ./UI/chatroom provides a basic interface for interacting with bots in the webpage, which supports text and voice as input. Please refer to https://github.com/scalableminds/chatroom for more details.

• In your Rasa bot setup, make sure to include the Rasa REST channel in your credentials.yml file:

rest:
  # pass

• Install the dependencies for web application

cd UI/chatroom
# install dependencies if you have not installed
yarn install

• Usage - You need to open 3 terminal/shell windows:

Terminal-1: For Rasa server. Depending on your setup you might need to add CORS headers, e.g. --cors "*".

# change to chatbot directory (just an example)
cd chatbot/05-wellness-bot
# Run Rasa server
rasa run --credentials ./credentials.yml  --enable-api --auth-token XYZ123 --model ./models --endpoints ./endpoints.yml --cors "*"

Terminal-2: Run Rasa action server if you need customized actions

# change to chatbot directory (just an example)
cd chatbot/05-wellness-bot
# Run Rasa action server
rasa run actions

Terminal 3: For web application

cd UI/chatroom
# run the local host
yarn serve

Open http://localhost:8080/demo.html in your browser.

2. Interactive-Voice-Response

• The original project: https://github.com/vishwaspuri/Interactive-Voice-Response I make slight modifications here.

• Usage - You need to open 3 terminal/shell windows:

Terminal-1: For Rasa server

# change to chatbot directory (just an example)
cd chatbot/05-wellness-bot
# run RASA core server
rasa run --enable-api --verbose

Terminal-2: For RASA actions server

# change to chatbot directory (just an example)
cd chatbot/05-wellness-bot
# run RASA actions server
rasa run actions

Terminal-3: For web application

# change directory to webapp
cd UI/Interactive-Voice-Response/backend
# run web application server
uvicorn app:app --reload --port 8000

On successfully running these, commands you will be able to run the chatbot in your browser on the address localhost:8000.

Other Resources

• Rasa official introduction: Read Docs
• Rasa 2.x examples: Learn Rasa programming via simple examples
• Rasa chatbot examples: Rasa chatbots
• If you have questions: Rasa Community Forum