Quick question from the Awkward Array project

Ecosystem
1

Hi all,

I’m the maintainer of Awkward Array, and I noticed that your team has been exploring or integrating Awkward alongside AnnData. It’s incredibly exciting to see the library finding a home in single-cell workflows.

We’re planning our next development roadmap (and looking at some grant funding to back it up), so I wanted to reach out and see if we could support your work better.

How are you currently utilizing Awkward in your pipelines? If there’s any specific integration you’re interested in, or if you’ve hit any weird friction points, I’d love to hear about it.

Feel free to just drop a few sentences here if you have the time!

Best regards,

Ianna

2

Hi there!

I am not too familiar with the ins-and-outs of awkward array (that is more a @grst question), but I am a file-format person and a core zarr developer.

So the biggest thing for me would be formalizing an on-disk spec and adding it to Zarr Conventions · GitHub so that tools outside anndata can recognize it (our spec for awkward arrays is here: On-disk format — anndata 0.12.16 documentation)

There was some recent discussion of this but cannot find it anymore.

Ilan

2 Likes
3

There was some recent discussion of this but cannot find it anymore.

Maybe this thread?

Irregular chunked Zarr Arrays for High-Energy Physics Analysis · zarr-developers/zarr-python · Discussion #3948 · GitHub

1 Like
4

I don’t have any requests at the moment. For my use-case (single-cell immune receptor data) it works well. As far as I can tell this is also still the predominant use-case in the scverse ecosystem. I have seen some uses also with spatial data, but it’s not widespread.

1 Like