Kite Now Integrates with Jupyter
We’re thrilled to release a Kite integration for JupyterLab, plus special support for teams using JupyterHub. Data scientists can now get useful completions as they type in Jupyter notebooks, powered by Kite’s deep learning.
Check out the new completions experience in JupyterLab below.
Follow the instructions here to install Kite for JupyterLab.
Let’s dive into the details.
Kite automatically gives you longer completions sorted by relevance, without requiring you to run a single cell in your notebook or press `tab` to make completions appear.
Here’s a side-by-side comparison for JupyterLab’s native completions vs. Kite completions:
JupyterLab native completions are alphabetically sorted and have no documentation.
Whereas Kite’s completions are ranked by relevance, and Kite also shows you documentation for the highlighted completion. 1The documentation panel can be toggled on and off with a Jupyter console command.
Kite then turbocharges the JupyterLab experience with additional ML-powered completions features and over 100,000 Python docs just a click away in the Kite Copilot:
Kite can complete up to multiple lines of code at a time, reducing the time you spend writing repetitive, boilerplate Python.
Import statements are a breeze with Kite, and it learns and suggests your favorite aliases over time.
Kite continues to tap into Jupyter kernel completions, which are useful when you need to access attributes such as columns on a DataFrame. With Kite, these kernel completions now show up automatically as you type.
Save time searching for Python docs with Kite’s Copilot desktop app. View Python docs with just one click or mouse-hover, plus find helpful examples and how-tos.
More features you’ll get with Kite in JupyterLab:
- Kite’s models work locally and independently of your Python kernel. This means if your kernel’s busy reading in data, you‘ll still get Kite completions while coding in other cells. Plus, no code is sent to a cloud server for processing.
- Kite works in .py files within JupyterLab. 2Kite also works in other editors, like PyCharm and VS Code. However, Kite does not support .ipynb files in PyCharm or VS Code. Unfortunately these editors provide proprietary notebook support which prevents us from building compatibility for the time being. We’re tracking this issue in our public Github repo here and here.
Kite makes coding with Python faster and more enjoyable
We’ve trained Kite’s deep learning models on over 25 million open-source Python files to ensure Kite works with your favorite libraries. With how fast data science tooling evolves, it’s critical to stay on top of new modules and APIs, and Kite’s completions help make that easier to do.
Kite’s deep learning models have learned the most popular patterns used by data scientists, plus they understand the context of your code. This means Kite can predict relevant chunks of code and put them in your completions. This can be useful in three ways:
- If you already know what you need to type, Kite helps you jump ahead to the next task.
- If you’re having trouble remembering a calculation or code pattern, Kite can remind you so you don’t need to search on Google.
- If you have never used a module or function before, you can get documentation faster with the Kite Copilot.
JupyterHub: Boost the whole team’s productivity with Kite
Does your team use JupyterHub? We offer additional features for JupyterHub teams:
- Deploy Kite on your JupyterHub server to instantly bring AI-powered completions and one-click documentation to the whole team.
- Add Kite’s largest ML models to a GPU-powered server for smarter, longer completions.
- Custom-tailor Kite’s models to your team’s codebase and APIs.
- Manage Kite licenses and billing through a unified system.
Email firstname.lastname@example.org to set up a demo and get extended free trials of Kite Pro for your whole team.
Thank you, Jupyter community!
This plugin would not be possible without the guidance of the Jupyter community and JupyterLab’s development team. In partnership with Quansight Labs, we worked with the JupyterLab development team to contribute to the Jupyter completions API and completions interface via four PRs and 87 commits (here are two examples). These improvements enable others in the Jupyter community to more easily build plugins for JupyterLab.
We will continue working with the community to help JupyterLab evolve and remain a powerful option for data scientists and developers.
The Kite Team
P.S. If you are a university student you can get Kite Pro for free by signing up with your school email address (instructions). And we’ll be expanding student access to high schoolers and others soon, just in time for the fall semester.