xarray-contrib / xarray.dev

The Xarray landing page
https://xarray.dev
Apache License 2.0
12 stars 18 forks source link

[User story proposal] Describe Xarray for protein ML/AI work #664

Open dcherian opened 3 months ago

dcherian commented 3 months ago

This is a proposal for a sequencing/protein engineering use case story for Xarray. The story will be published on Xarray's blog (https://xarray.dev/blog).

Why?

@asford and colleagues are quite excited about using Xarray +torch for generative ML models and "protein engineering". It'd be nice to broadcast this out.

It's also interesting that they find coordinate variables quite useful and so DataArray is the right structure rather than Variable/NamedArray/pytorch.NamedTensor

What?

We are targeting a short and non-technical post that illustrates how Xarray is (or could be) used in your genomics context. Below is a template outline that you should feel free to modify:

  1. Who am I?
  2. What problem am I trying to solve?
  3. How does Xarray help?
  4. Why did I choose Xarray?
  5. Current pain points (or ways that Xarray could better serve your use case)
  6. Technology I use around Xarray
  7. Anything else to know?
  8. Links and references

Feel free to insert images or short code blocks if they help you tell your story.

How?

Xarray's blog using Markdown with some front-matter. You can copy one of the existing posts to get started or you can write in a google-doc-like-thing and ask us to do the markdown formatting. If you copy a previous post, use a name like user-story-genomics or whatever makes sense for your application.

xref: https://github.com/xarray-contrib/xarray.dev/issues/272

dcherian commented 3 months ago

@asford would you be willing to write a quick blogpost about why you love Xarray for your workloads?

asford commented 3 months ago

Sure, but this will probably be back-burner for at least a few weeks. I'll keep this in-stack with the discussed updates for torch support we had today.