`bpl` dependency issue with UV

[JaiPancholi]

When running the airsenal_run_prediction via uv run command, I get the following error:

from bpl import ExtendedDixonColesMatchPredictor, NeutralDixonColesMatchPredictor
ModuleNotFoundError: No module named 'bpl'

Environment

• Operating System: macOS 15.4.1 (24E263)
• Python Version: 3.12
• Library Version: 1.11.0
• Environment Type: uv
• Installation Method: uv add airsenal

[jack89roberts]

Thanks for making the issue @JaiPancholi . Did you clone the repo and install from source or try to install AIrsenal from PyPi directly? If from PyPI, can you try cloning the repo instead and installing from source? And either way if you run uv sync first do you get any errors/changes?

[xchen2763]

Hi, I encountered the same issue when installing from PyPi and running airsenal_run_prediction:

[xchen2763]

Hi, I encountered the same issue when installing from PyPi and running airsenal_run_prediction:

Running well when installing from source with uv

[jack89roberts]

Thanks for letting us know. I don't think this is something I can fix, as I don't own the PyPI build, so I've assigned it to Nick and I'll update the readme in the meantime.

[jack89roberts]

I think with you fixing the bpl-next PyPI package last week airsenal on PyPI should now too @nbarlowATI if we update it (I think you own it so I can't try myself)?

I'm not sure if anything else is needed to make something `uv tool install airsenal'-able too, because that would then make airsenal commands globally available too without people needing to worry about managing environments etc. themselves.

[nbarlowATI]

I just uploaded AIrsenal 1.16.0 (the current main) to PyPi. Running in a clean test environment, it looks like we can now just do e.g. uv run airsenal_run_prediction etc. and it will use the pypi version (and find bpl-next) ok.