Pulse AI - Internet Technologies Project

Developed by
Juled Zaganjori (22206337) Elmer Dema (22211551)

Special thanks to Prof. Dr. Andreas Wölfl for his guidance and support throughout the project.

Deployment through Azure found at pulse-ai.de

About our idea

Pulse is a web-based chatbot that is able to hold a medical conversation with you and performing a disease diagnosis based on the symptoms you provide.

Diagnosis

During the diagnosis, the chatbot will give you a list of precautions, information about the disease, and a emergency number in case you need it. If you are not happy with your diagnosis you can ask for a new one and the process will restart.

Normal conversation

The chatbot is also able to perform a normal medical conversation with you, such as greeting you, asking you how you are doing, saying goodbye, giving advice and recommendations on how to stay healthy.

Other features

Pulse can also give you advice on different topics such as nutrition, exercise, mental health, emergency situations, and more.

Extending the chatbot

The conversation responses can be easily extended by the user by adding more responses to the JSON dataset responses.json.
You provide the keywords, different responses and set the do_action to null.
Advanced features: In case you want to introduce a new action feature you would need to modify the code in the pulse-engine.js file.

OpenAI engine

You can also switch between the pulse engine and the OpenAI engine through the dropdown on the top right corner of the chatbot.

Getting Started

Clone the repository

git clone https://github.com/Dialogue-Developers/pulse.git

Install dependencies

npm install

Run the server

npm start

Editing the chatbot engine

Adding new responses

You can change the responses of the chatbot by editing the responses.json file.
The file contains a list of responses that the chatbot will use to respond to the user.
The responses are divided into several action categories called do_action: null (normal), diagnose, ask_for_disease_info, ask_for_disease_precautions, ask_for_doctor, end_conversation etc.

Adding new diseases

You can add new diseases to the chatbot by editing the diseases.json file.

Editing the UI

You can edit the UI of the chatbot by cloning the repo somewhere else and switching to the dev-front-end branch.

git clone https://github.com/Dialogue-Developers/pulse.git
git checkout dev-front-end

After you are done editing the UI you can push the changes to the dev-front-end branch.

Run the following command to build the app:

npm run build

Copy the build folder and paste it into the public folder of the main branch.
After that you can push the changes to the main branch and the app will be deployed automatically by our Azure pipeline.

Deployment

The app is deployed on Azure and can be found at pulse-ai.de or pulseailive.azurewebsites.net (backup link).
The deployment is done automatically through Azure pipelines.
The pipeline is triggered when a new commit is pushed to the main branch.
The pipeline will build the app and deploy it to the Azure web app.

Responsibilites

Juled Zaganjori

• Pulse engine
• Fallback options
• Deployment
• Keyword spotting for the diagnosis
• OpenAI engine

Elmer Dema

• Design & Layout
• React components
• Keyword spotting for the normal conversation
• Creating the dataset for the knowledge base

Backend built with

• Node.js - JavaScript runtime environment
• Express.js - Web framework for Node.js
• Socket.io - Real-time communication engine
• OpenAI API - API for the OpenAI engine

Frontend built with

• Bootstrap for the essential styling + Custom CSS for the chat UI
• React.js - JavaScript library for building user interfaces

Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.