Could we add Python Optimal Transport to the pangeo-notebook environment?

[tatsumonkman47]

I'm using the Python Optimal Transport (POT) library to calculate ND Wasserstein distances between large probability distributions generated from ocean gcm data on the leap-pangeo 2i2c hub. The is a very intensive but easily parallelizable calculation that I'd like to perform using Dask Gateway.

Would it be possible to add the Python Optimal Transport (POT) to the pangeo-notebook image so that I can use it with a dask cluster without having to manually install it on each client?

Thanks!

[weiji14]

Hi @tatsumonkman47, it looks like the pot package is on conda-forge already (with linux-aarch64/ppc64le builds even), and has a reasonable number of downloads and stars on GitHub, so I think it would be ok to add.

Would you like to open a Pull Request to add pot to https://github.com/pangeo-data/pangeo-docker-images/blob/master/pangeo-notebook/environment.yml?

[tatsumonkman47]

Thanks! I've opened a pull request

[weiji14]

FYI, the new tag 2025.07.09 of pangeo-notebook should contain pot, so I'll close this now. Let us know how it goes!

[tatsumonkman47]

Great! Thank you @weiji14 !

[tatsumonkman47]

Update (7/15/2025): POT appears to be working fine on the dask gateway clients