ChatGPT plugins quickstart with Gist and Gradio Analytics

Get a todo list ChatGPT plugin up and running in under 5 minutes using Python. This plugin is designed to work in conjunction with the ChatGPT plugins documentation. If you do not already have plugin developer access, please join the waitlist.

This version of the plugin includes additional functionality to log requests to a GitHub Gist and visualize the request data using Gradio.

Setup locally

To install the required packages for this plugin, run the following command:

pip install -r requirements.txt

To run the plugin, enter the following command:

python main.py

Once the local server is running:

1. Navigate to https://chat.openai.com.
2. In the Model drop down, select "Plugins" (note, if you don't see it there, you don't have access yet).
3. Select "Plugin store"
4. Select "Develop your own plugin"
5. Enter in localhost:5003 since this is the URL the server is running on locally, then select "Find manifest file".

The plugin should now be installed and enabled! You can start with a question like "What is on my todo list" and then try adding something to it as well!

Using GitHub Gist for Logging

This plugin logs each request to a GitHub Gist. To set this up:

1. Create a new Gist on GitHub. You can do this by going to https://gist.github.com/ and clicking on the "+ New gist" button. The Gist should contain a single CSV file (e.g., requests_log.csv).
2. Get your GitHub token. You can generate a new token by going to https://github.com/settings/tokens and clicking on the "Generate new token" button. Add this token to the code where it says YOUR_GITHUB_TOKEN.
3. Replace GIST_ID and FILENAME in the code with the ID of your Gist and the name of the CSV file, respectively. The Gist ID is the last part of the URL of your Gist (e.g., if your Gist URL is https://gist.github.com/username/1234567890abcdef, the Gist ID is 1234567890abcdef).

The plugin will now log each request to the Gist.

Gradio Interface for Visualizing Request Data

This guide will help you set up a Gradio interface to visualize the request data logged to a GitHub Gist. This is a continuation of the ChatGPT plugins quickstart guide with Gist and Gradio Analytics.

Setup

1. Install Gradio by running pip install gradio in your terminal.

2. Use the following template to create a Gradio interface:

import gradio as gr
import pandas as pd
import matplotlib.pyplot as plt
from matplotlib.dates import DateFormatter
import matplotlib.dates as mdates

def show_data():
    # URL of the raw CSV file in your Gist
    url = "https://gist.githubusercontent.com/YOUR_GITHUB_USERNAME/GIST_ID/raw/requests_log.csv"
    
    # Read the CSV file
    df = pd.read_csv(url)
    
    # Convert the 'Timestamp' column to datetime
    df['Timestamp'] = pd.to_datetime(df['Timestamp'])
    
    # Table
    table = df

    # Chart
    fig, ax = plt.subplots()

    # Group the data by date and request type, and count the number of each type per day
    daily_counts = df.groupby([df['Timestamp'].dt.date, 'Request Type']).size().unstack()

    # Plot the data
    daily_counts.plot(kind='bar', stacked=True, ax=ax)

    # Format the x-axis
    ax.xaxis.set_major_locator(mdates.WeekdayLocator())
    ax.xaxis.set_major_formatter(mdates.DateFormatter('%b %d'))
    
    plt.xticks(rotation=45)
    plt.tight_layout()

    return fig

iface = gr.Interface(fn=show_data, inputs=[], outputs=["plot"])
iface.launch()

3. Replace YOUR_GITHUB_USERNAME and GIST_ID in the url with your GitHub username and the ID of your Gist, respectively.

4. Run the script. This will launch a Gradio interface in your web browser.

You can now visualize your request data in a Gradio interface! The interface displays a bar chart showing the number of each type of request per day. The data is read directly from the CSV file in your Gist.

For a live example of this setup, you can visit this Gradio interface and replace the url value with your own Gist URL.

Setup Remotely

You can also deploy your plugin on a remote server. Here are a few options:

Cloudflare Workers

Cloudflare Workers allow you to run your serverless code on Cloudflare's edge network. You can follow the instructions in the Cloudflare Workers Quick Start Guide to deploy your plugin.

Code Sandbox

CodeSandbox is an online code editor that allows you to create web applications, including server applications. You can use the CodeSandbox Node.js Template to create a new Node.js sandbox and then copy your plugin code into the server.js file.

Replit

Replit is an online code editor that supports many programming languages, including Python. You can create a new Python repl at https://replit.com/ and copy your plugin code into the main.py file. Then, click on the "Run" button to start your plugin. You can access your plugin at the URL shown in the "Web Server" tab.

Remember to replace localhost:5003 with the URL of your remote server when you install and enable the plugin in the ChatGPT interface.

Getting help

If you run into issues or have questions building a plugin, please join our Developer community forum.