please don't pin numpy versions

[agramfort]

just sharing my experience now...

I just tried to install modopt on my conda env. As you pin requirements and you don't provide conda packages
I had to use pip. Doing so just downgraded numpy on my env and simply broke all my installation.

please please avoid pinning numpy versions (use >= and not ==) and consider pypi and conda packages

🙏

[sfarrens]

Hi @agramfort this is something we will be fixing soon. Sorry for the inconvenience.

[mathurinm]

👍 pinning makes installation harder than it should be

let me know if I can help !

[sfarrens]

Hi @mathurinm thanks for your interest. There is actually a good reason for why we did this as a temporary solution, but we will be removing the pinned versions for the next release, which is currently planned for mid December.

[agramfort]

could you do it ASAP on master branch and we install master in benchopt? we don't need to wait for a release in benchopt

…

[sfarrens]

Hi @agramfort, absolutely. Apologies for the delay, I didn't realise you didn't need the formal release. I will do it this week and ping you.

[agramfort]

thanks !

…

[sfarrens]

@agramfort I have unpinned the package dependencies (at least those needed by users) in #189. This is currently available in the develop branch. Let me know if specifically need these changes in the master branch, if so I will merge this straight away. Otherwise, this will be merged for the next release (on or before 17/12/21).

[sfarrens]

Resolved in #189