Web-based CHATGPT-powered Prompt for Human Activity Recognition (HAR)

Project Overview

This repository contains a web-based application designed to leverage ChatGPT for Human Activity Recognition (HAR) tasks. Follow the instructions below to set up and run the project.

Getting Started

1. Clone the Repository

git clone https://git.scc.kit.edu/chat-gpt-har](https://github.com/suleimanelkhoury/chat-gpt-har.git
cd chat-gpt-har

2. Install Dependencies

pip install -r requirements.txt

3. Configure OpenAI API Key

Open app.py and insert your OpenAI API key:

openai.api_key = "Your_Key"

4. Run the Application

python app.py

Project Structure

• app.py: Contains the ChatGPT integration and the Gradio-based user interface.
• functions.py: Includes all prompts and processing pipelines for the HAR exercises.
• kittheme.py: Provides a custom theme for the Gradio UI, using the Karlsruhe Institute of Technology color scheme.
• requirements.txt: Lists all dependencies required for the project.

User Interface Guide

Model Selection

• Location: Top-right corner.
• Description: Choose your preferred ChatGPT model from a dropdown menu. The default setting is GPT-3.5-turbo.

Important Information Representation

• Function: Input a dataset description to receive a structured and standardized JSON output of the key information.

Related Work Listing

• Function: Enter a paper title and specify the number of related works to display using the "count" option.
  • "Pure GPT" Button: Displays relevant related papers recommended by ChatGPT.
  • "Get Citations With Scholarly" Button: Shows papers that cite the entered paper.
  • "Get References With Plybliometrics" Button: Outputs references from the entered paper (requires a key in pybliometrics.cfg).

Tools and Algorithm Recommendations

• Function: Provides recommended algorithms based on the input.

Data Exploration

• Function: Upload a .txt file from the Human Activity Recognition dataset.
  • "count" Option: Adjust the number of rows used for data exploration.
  • "Data Visualization Prompt" Button: Generates code for visualizing the data.
  • "Descriptive Statistics Prompt" Button: Outputs a statistical analysis of the input data.

Data Format Transformation

• Function: Transforms input data into a structured format.