Hello, I'm looking for guidance on how to limit the

platform_system

recognized by Pants (and by extension, pip) within my project's configuration. I'm currently facing a challenge in upgrading TensorFlow from version 2.15 to 2.16. This upgrade is essential for my project as I need to move from Python 3.12 to Python 3.11, but TensorFlow 2.15 does not support Python 3.12. The issue becomes apparent when I attempt to generate the lockfile, as for

platform_system == "Linux"

and either

platform_machine == "aarch64"

or

platform_machine == "arch64"

, there's a dependency on

tensorflow-cpu-aws

, which hasn't been updated to support TensorFlow 2.16. However, the

aarch64

and

arch64

platforms are not relevant to my use case. Is there a way to bypass or exclude these specific platform checks to facilitate the TensorFlow update? Thank you in advance for your assistance!

Is this expressible with "normal" requirements?

sys_platform != 'darwin' or platform_machine != 'arm64'

etc?

I tried that, pants still tries to generate the lockfiles for every platform.

Running into this myself, did you find a resolution?

I have not really fixed this. Changing to the nightly tensorflow release worked, but that caused other issues instead (because transition from keras 2 to keras 3). So I am basically just waiting for 2.16.2 and hope it addresses the issue you posted.