packaging?

[giovp]

hi @willtownes , really looking forward to try this out! Is there any chance that this will get pip installable in the near future? Happy to contribute if you think it's useful.

[willtownes]

Yes it is on my to-do list, but I have to finish up some other stuff first so hoping to get to it in a couple weeks. I'd be delighted for any assistance as my python packaging experience is somewhat limited. If you like I could create a blank repository and give you edit permission. I don't want the package to live in this repository.

[giovp]

hi @willtownes thanks for prompt reply. I guess I could take a stab but I think there are a number of decisions you'd have to take re API . What I could do relatively fast it:

• make it pip installable pip instal -e .
• minimal pre-commit checks (black, flake8) for code quality
• maybe tox for tests? although I guess pytest already solution.

Then pypi publishing and docs (if planned) could be done later on.

I guess what I think would be cool to have eventually is something like this

import nsf
import scanpy as sc

# maybe some preprocessing e.g. data splitting etc.
nsf.model(adata, n_factors=..., kernel_type=..., idx_points=..., ...)
sc.pl.spatial(adata, color="factor1", ...)

let me know what you think.

[willtownes]

Sorry this took so long, the package is now under development at https://github.com/willtownes/spatial-factorization-py . Let's move the discussion to there.