Rest Easy with Jupyrest: Deploy notebooks as web services
summary
Jupyrest claims it turns Jupyter notebooks into web services without rewriting code by treating notebooks as notebook functions with an ID an input JSON Schema and an output schema and by injecting a parameter cell and running the notebook end to end. It produces an executable HTML artifact plus JSON via a save_output call, converts JSON into Python objects for complex models, integrates with FastAPI and Azure Functions for serverless deployments, runs in production at Microsoft since 2020 for many Azure teams and fits automation and reporting but not low latency services.
AI-written summary · about
chapters (16)
- 0:00Jupyrest introduction
- 1:27What is serverless
- 2:10Jupyter notebooks benefits
- 2:51Jupyrest purpose
- 3:56Notebook function structure
- 6:07Demo: Notebook to REST API
- 9:21Debugging notebook failures
- 11:06Reusing Python classes
- 13:58Handling Python class inputs
- 15:46When not to use Jupyrest
- 16:51Incident response automation
- 18:15Practical use cases
- 19:18Open source resources
- 20:01Q&A: JSON schemas
- 21:23Q&A: Interactive widgets
- 22:25Q&A: Deployment methods
related talks
31mNotebooks as Serverless Functions
34mNotebooker: Production and Scheduling for your Jupyter Notebooks
25mOvercoming 5 Hurdles to Using Jupyter Notebooks for Data Science, by the JetBrai...
32mJupylates- spaced repetition for teaching with Jupyter
17mNotes on Notebooks: Is Jupyter the Bringer of Jollity?
28mFrom Jupyter Notebook to Publish-Ready Report: Effortless Sharing with Quarto
01h 00mFrom notebook to production using
41mPublishing Jupyter Notebooks with Quarto