Rasa Receptionist Bot (Streamlined)V2

Description

This Rasa chatbot serves as a smart virtual receptionist in an office setting, handling various types of visitors and purposes of visit, including deliveries, meetings, door opening, and other basic functions such as telling time.

Screenshots

Examples of the bot running in Terminal:

Asking for time:

Delivery:

Interview:

Meeting:

Servicing:

Tech/framework used

• Rasa open source

Files in Rasa Folder

A brief rundown of the purpose of some of the files, which are closely interconnected with one another.

• actions.py This file contains all written Python code of custom actions.

• config.yml This file contains configurations for Rasa NLU and Rasa Core, such as tokenizers and policies.

• credentials.yml This file contains the credentials for the voice & chat platforms which the bot is using.

• data

  • nlu.md This file contains all the intents and their variations.
  • stories.md This file contains different variations of paths that a user might go through.

• domain.yml This file contains the actions, utters, entities, forms, slots, responses and buttons that are used.

• endpoints.yml This file contains the different endpoints the bot can use.

• models This folder contains model(s) which are trained.

Conversational Workflow

Main paths of workflow:

graph LR
Start[Welcome Message] -- I am here to make a delivery --> A(Delivery)
Start[Welcome Message] -- I am here for a meeting --> B(Meeting)
Start[Welcome Message] -- I am here for an interview --> E(Interview)
Start[Welcome Message] -- This is my first day --> F(New Staff)
Start[Welcome Message] -- I am here for servicing --> G(Servicing)

Loading

Full workflow and functions:

Installation

Rasa Quick Installation

You can install Rasa Open Source using pip (requires Python 3.6 or 3.7). It is strongly recommended to create a virtual environment.

$ pip3 install rasa

For full installation guide, check out the Rasa Documentation.

Running locally

The bot can be run on Terminal locally. The bot calls the action server via the url specified under endpoints.yml.

1. Run Rasa Shell In a Terminal tab in your bot directory, enter this command line to run rasa:

rasa shell

Alternatively, to see debugging information in Terminal as the program runs, you can append the debug property:

rasa shell --debug

2. Run Action Server Running the action server is necessary to utilise the custom actions writtten in actions.py. To run the action server, open up another tab in Terminal and enter this command line:

rasa run actions 

Demo:

Comparison with Web-based bot

This project is a streamlined version of another web-based Rasa receptionist bot we built with more comprehensive features and a different scope, which is hosted on a front-end user interface on this webpage. (Link updated as of 03/07/2020)

	Streamlined (this version)	Web-based
Functionality	Helps visitors to find staff	Helps visitors find staff, course registration, room booking, office map, notifications and attendance
Input	Voice input only	Voice, text, button clicks
User Interface	Game-based design	React
Casual Conversation	User can ask for date and time, directions to the MRT, and about Yoozoo	No casual conversation
Name Extraction	Asks for staff's name only	Asks for both visitors' and staff's names
Incorrect Extraction	Bot repeats process	Bot requests user to type in chat bubble
Bot Response	Unique bot responses depending on purpose of visitor	For simplicity, bot response is the same regardless of visitor's purpose
For services	Does not require staff's name to get someone to open the door	Still require staff's name

Credits

While working on this project, we referenced the Rasa Official Documentation.

Furthermore, we would like to express our gratitude to our mentors Ng Wai Foong and San Lin for their patience and guidance as we work on this project.

Thanks for reading :)