Please remove the unneded makedepends=python-sphinxcontrib-napoleon
.
Also please drop python-funcsigs
dependency. It is not required.
Funcsigs is a backport of a Python 3.3+ standard feature (PEP362) to Python 2.6, 2.7, and 3.2.
Git Clone URL: | https://aur.archlinux.org/python-nipype.git (read-only, click to copy) |
---|---|
Package Base: | python-nipype |
Description: | Neuroimaging in python pipelines and interfaces |
Upstream URL: | https://github.com/nipy/nipype |
Keywords: | mri neuroimaging nipy |
Licenses: | Apache |
Submitter: | fishburn |
Maintainer: | liamtimms |
Last Packager: | liamtimms |
Votes: | 0 |
Popularity: | 0.000000 |
First Submitted: | 2018-03-09 18:42 (UTC) |
Last Updated: | 2022-09-30 16:01 (UTC) |
Please remove the unneded makedepends=python-sphinxcontrib-napoleon
.
Also please drop python-funcsigs
dependency. It is not required.
Funcsigs is a backport of a Python 3.3+ standard feature (PEP362) to Python 2.6, 2.7, and 3.2.
Very open to a co-maintainer on this package as it has a lot of dependencies and is relatively complex.
I cannot update this package until the new dependency: https://github.com/effigies/looseversion has been packaged as well. Please attempt to package that dependency before reporting this package as out of date.
A nonzero amount of the check function is fairly broken right now but you can get around it using makepkg --nocheck
. I am not completely removing it because I want users to be aware of any failing checks before using the package. I am testing -git
versions of a few packages to see if I can get around these issues.
@keiichiiownsu12 I can confirm this on my side too. My first guess is that it's because we are on Python 3.9 on Arch but nipype is only officially for up to 3.8. We are technically outside the bounds of official support. However, it may be worth talking to upstream. I can file a report in a few days or you can work on it now if you like. In the meantime, you could try disabling the testing and see if it actually affects your workflows.
When installing nipype 1.5.1, the testing phase produces >30 repeated errors involving the following:
nipype/interfaces/dipy/__init__.py:5: in <module>
from .preprocess import Resample, Denoise
nipype/interfaces/dipy/preprocess.py:27: in <module>
globals()[new_name] = dipy_to_nipype_interface(new_name, obj)
nipype/interfaces/dipy/base.py:194: in dipy_to_nipype_interface
args + (val,) for args, val in zip(parser.optional_parameters, default_values)
E TypeError: 'NoneType' object is not iterable
(Same error across ~30 scripts). Have not seen mention of this upstream...
Important note to anyone attempting to use this package: there is an issue with dependencies which requires manual intervention. https://github.com/nipy/nipype/issues/3055 due to this, Nipype may install without issue but fail when running workflows. The workaround is as follows; 1) install python-prov as you would a normal AUR package, this will call the python-rdflib dependency, 2) make the python-neurdflib package with makepkg -s
, 3) use the normally not recommended --overwrite /usr/lib/python* python-neurdflib*.pkg.tar.xz
flag on pacman -U
to force neurdflib to install its additions over rdflib, 4) install python-nipype. This is the unfortunate state until rdflib responds to the neurdflib pull request and may cause issues in the future.
@wedjat Thanks. No plans to create anymore nipy packages in the near future. Feel free to add!
Thanks for uploading this and nilearn! I thought of making them myself and submitting next week. Are you planning on adding the rest of nipy packges, like nitime, dipy etc.?
Pinned Comments
liamtimms commented on 2023-02-15 21:17 (UTC)
Very open to a co-maintainer on this package as it has a lot of dependencies and is relatively complex.