search

moble / quaternionic

Interpret numpy arrays as quaternionic arrays with numba acceleration
MIT License
84 stars 7 forks source link

conflicting importlib-metadata requirements #20

Closed ds-steventondeur closed 3 years ago

ds-steventondeur commented 3 years ago

Hi,

I got here via the 'numpy-quaternion' package and I'm trying to change my code to using quaternionic as the latter seems to be the preferred approach (at least quaternionic seems to have a bit better documentation).

The 'importlib-metadata' dependency is causing a problem though (I'm on python 3.6; hope to upgrade soon but not there yet):

When running python -m pip install quaternionic in my .venv, I get the following error message:

ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. keyring 23.0.0 requires importlib-metadata>=3.6, but you have importlib-metadata 1.7.0 which is incompatible.

A workaround for this problem is to temporarily comment out any reference to our Azure repo in the venv's pip.ini file and then upgrade importlib-metadata back to a high enough version (>=3.6), but I was wondering why this dependency on importlib-metadata exists in quaternionic; any chance we can get rid of it/upgrade it to a higher version?

Kind regards, steven

moble commented 3 years ago

I got here via the 'numpy-quaternion' package and I'm trying to change my code to using quaternionic as the latter seems to be the preferred approach (at least quaternionic seems to have a bit better documentation).

The latter certainly takes advantage of newer capabilities, which makes the documentation easier to write. And I myself have mostly switched to quaternionic at this point. But I should point out that it's not as battle-tested as the former, just because it's newer and far fewer people use it.

any chance we can get rid of it/upgrade it to a higher version?

Absolutely. The new version should be on PyPI in about half an hour, and conda-forge later in the day.

ds-steventondeur commented 3 years ago

Thanks a bunch! Lucky Luke, fixing issues faster than his shadow.