Jupyter Notebooks are perfectly suited for exploration, experimentation, and quickly sharing insights with others. But when you need to productionize that code, you will find that some features that make them great for exploration don’t translate well to writing robust, reproducible code that runs automatically. We’ll help you bridge that gap.
In this tutorial we will take an existing notebook, refactor it into production ready code, write comprehensive tests, and build an API. In each part of the tutorial we’ll start with an introductory presentation, then participants will put the principles into practice with a hands-on exercise. We’ll provide the example notebook and starter code for each exercise in a GitHub repository available before the tutorial.
The tutorial will have four parts: 1. Strategies for refactoring your code from a notebook to a standalone script, and tools you can use to make the journey easier. 2. An introduction to writing tests: what a basic test should include, the difference between a unit test and an integration test, and how to run tests using pytest. 3. How to build a basic API: what is an endpoint, common types of REST endpoint, and how to build an API using FastAPI. If time permits, we’ll demonstrate how to deploy this endpoint to a cloud hosting service. 4. These tasks can be done efficiently using AI-assisted code generation. We’ll cover strategies to use an AI-powered code editor effectively to take your code from notebooks to production.
This tutorial is for anyone who works predominantly in notebooks but wants to learn best practices for putting their code in production. Participants will need to know how to clone a GitHub repository, as well as basic knowledge of some common libraries like pandas.