ELM API Keys and Noteable

Objective: to demonstrate the features for generative AI in Jupyter notebooks and give a simple example of using the Jupyternaut generative AI chatbot in Noteable.

The Noteable service available for courses at the University of Edinburgh provides access to computational notebooks, including JupyterLab and its new AI features.  

JupyterAI is an addition to the Jupyter ecosystem that introduces generative AI capabilities in JupyterLab and Jupyter notebooks. It empowers users to generate and explain code, troubleshoot errors, summarise content, and engage with data files through natural language prompts.  

Jupyter AI is engineered with a focus on ethical AI and data privacy, providing users the autonomy to choose LLMs that best meet their criteria and transparency in how data is utilised. Grounded in the open-source ethos, Jupyter AI only engages with an LLM upon direct user request. 

Jupyternaut, the Jupyter AI feature that provides a chatbot assistant, offers conversational problem-solving and code generation:  

 

Language Model

Jupyternaut is an AI chatbot option available in JupyterLab.

You can bring an LLM API key (such as an ELM API key) and start interacting with the chatbot directly in JupyterLab. 

 

The table below can help academics understand and decide which Language Models (LLMs) to use within Noteable, specifically through Jupyternaut and Jupyter AI:

 

LLM Option Capabilities & Use Cases Integration in Noteable
GPT-3.5 Turbo - Strong language processing and coding assistance. - Suitable for generating text and debugging code. - Use for interactive coding support. - Provides real-time suggestions for Jupyter notebooks.
GPT-4 - Enhanced reasoning and complex problem solving. - Better at generating detailed content. - For assignments requiring detailed analysis and content creation. - Supports complex problem-solving tasks.
GPT-4o - Offers fast response times with detailed output. - Capable of handling comprehensive AI tasks. - Mimics the experience of AI use in professional coding environments.
o1/o3 Models - Focus on efficiency and cost-effectiveness. - Optimised for structured, repetitive tasks. - Suitable for real-time interactive sessions. - Automates text-related tasks like summarisation and classification.

 

Integration Tips

  • ELM API Keys: Utilise these to integrate the OpenAI LLMs within Noteable for model interaction.
  • Guidance for Students: Supply clear instructions on using AI tools effectively in coursework.
  • Customisation: Adjust model use to fit course needs, ensuring alignment with learning outcomes.
  • Real-world Skills: Design activities that enhance skills relevant to industry practices.

 

Case Study: 

Introduction:

This case study outlines the implementation of Jupyter AI within Noteable for a University of Edinburgh   course in Geophysics, exploring how educators can get started with Jupyter’s AI tools to enrich the learning experience and connect students with real-world AI applications.

 

Objective:

The goal was to integrate Jupyter’s AI tools into the curriculum to simulate a real-world coding environment, thereby preparing students with practical skills in AI usage. The focus was on demonstrating how AI tools can be used for problem-solving, bug fixes, and methodological research.

 

Outcomes:

Students engaged actively with AI tools, gaining hands-on experience in AI-assisted coding.

The integration of JupyterAI supported students' ability to explore and apply new methods and solve real-time problems efficiently.

Getting started with Jupyter AI and tools like Jupyternaut in Noteable involves understanding the functionalities, integrating them using ELM API keys, and appropriately introducing them to the curriculum. 

 

Getting Started:

Understanding Jupyter AI and Jupyternaut:

Jupyter AI is an extension of JupyterLab that introduces generative AI capabilities, such as code generation, error troubleshooting, and data file engagement through natural language prompts.

Jupyternaut, a chatbot feature, provides conversational problem-solving and code generation, allowing students to interact with AI models in an intuitive way.

To access Jupyternaut in Noteable, launch one of the JupyterLab notebook options, such as the Standard Python 3 notebook, and click on the chat icon in the left side bar: 

 

Noteable Interface
Standard Python

 

1. Access the Jupyter AI chatbot on the left-hand sidebar. 

2. Access the settings to select your desired model 

 

Accessing your Jupyter AI model

 

2. Add your API key and scroll down to Save. 

3. You are now able to interact with Jupyternaut AI chatbot using your chosen AI model and API key. 

 

Adding your API Key

 

Integrating with Noteable:

ELM API Keys: 

Secure ELM API keys are provided by the University of Edinburgh under an enterprise agreement with OpenAI, which can be used for integrating OpenAI LLMs within Noteable to interact with OpenAI models.

You can request an API Key HERE.

This article was published on