Doris Xin

Doris Xin

10/06/2022, 5:52 PM
Hi everyone! Wanted to share LineaPy, an OSS project my team created 🙂 Is translating dev code to data pipelines a time-consuming mess for you? Is it impossible to share data science work with teammates when everyone has their own environments and scripts and file systems? Frustrated that you have no idea how a model was trained, who trained it, and what data was used? LineaPy can help with all of these problems. LineaPy works locally out-of-the-box, but it’s much more powerful when used to collaborate. We’re excited to share this demo that shows you how to make LineaPy work for your team to support collaboration! To spin it up yourself, read our demo tutorial here: https://bit.ly/3LXBL2d See our

YouTube video

walkthrough of it. Or, see it live during our October 14th workshop! Register here: https://bit.ly/3C1Pwbw What sets LineaPy apart from other tools? LineaPy automatically refactors code and generates pipelines, saves the code and value for artifacts (models, datasets, scalars, charts, etc.) in one place for easy retrieval, and provides a central artifact store for all your data science work instead of having it spread across many notebooks and file systems. It does so without requiring you to change how you work. LineaPy captures everything automatically and performs program analysis for semantic understanding of your workflow to infer the structure of the data pipeline and refactor code accordingly.