-
Notifications
You must be signed in to change notification settings - Fork 4
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
using robustica with numpy>=2.0 #6
Comments
Hi @handwerkerd , thanks for noting this and posting the issue. I will look into it asap! Best, Miquel |
Thank you @MiqG! I've been looking into this a bit. I tried to locally install [Update] It seems like this is a partially known issue: scikit-learn-contrib/scikit-learn-extra#177 and that issue makes it seem like a solution might take them time. [Update2] Based on the above thread, if I uninstall and then reinstall scikit-learn-extra using
this error disappears and then I can also successfully import robustica. I think this means, to get a new version of your code up pipy that can use |
Oh I see! I just updated the README to inform users about that. Thanks for digging into the bug @handwerkerd ! |
@BahmanTahayori was able to remove scikit-learn-extra and this overall issue is resolved primarily by removing two options in In general, I'd like to get a sense if you have the bandwidth to fix this in your repo or if, for our project, we should go forward with our own fork that removes |
…des whiten default from whiten=True to whiten=arbitrary-variancefollowing latest scikit-learn updates RobustICA class.
Hi @handwerkerd , this is easily fixable! In this case, rather than removing it completely, I think making I just committed some changes that in my end allow installing and running the package without requiring Could you give it a try as well installing the latest version of the repo? If it works on your end as well, I will make the release of a new version and update pypi: pip install git+https://github.com/CRG-CNAG/robustica.git My intention is to keep the repo as active and up to date as possible. Of course contributions are very welcome! |
This works. Thank you! I'm also glad to hear you plan to keep this active & up to date. I will mention two things that might need edits:
|
Great, I think I'll be able to implement this by next week. |
Done! |
Thank you for taking care of this! |
I'm part of a group that is adding robustica into https://github.com/ME-ICA/tedana
I just came across an issue that packages that use numpy and were last compiled before the release of numpy v2.0 don't run when numpy>=2.0. That means
pip install robustica
no longer works if an environment uses a newer version of numpy.This is the message I am seeing:
I'm assuming, if you release a new version of
robustica
using an environment with a newer numpy version, it should work with newer and older versions of numpy. This is a bit out of my expertise, but I don't think any code changes are necessary.I'm not sure this is relevant here, but if you want to see the details of the PR where we're adding
robustica
, it's here: ME-ICA/tedana#1125 (comment)Thank you for creating and openly sharing robustica!
The text was updated successfully, but these errors were encountered: