Closed ATajadod94 closed 4 years ago
Seems like this issue is resolved in #28. As a side note perhaps we should pin the versions in requirements.txt
.
It appears I was wrong about numpy installation issue. While running python setup.py build_ext --inplace
there are a bunch of warnings before #28 (https://pastebin.com/nzwR8Czv) and after #28 (https://pastebin.com/jhSN8c2R) with the warning of only auto_ptr being removed. Note that it doesn't fail to build the extension.
However, there is no issue with python setup.py install
. It may be noted that they are not particularly numpy errors.
Dear all,
I tried to install tvb-gdist from pypi using pip install command under linux distribution (xenial) and I encounter the numpy not found problem that causes the installation to crash. I also tried to install it from latest source (github) thinking the problem was solved but the installation still crashes (see travis log file enclosed). Using windows i do not encounter any problem since there are available binaries in Pypi. Any idea how I could solve this issue ?
Can you please try pip install numpy
before installing tvb-gdist?
Hello anyan-b, thank you for your quick answer. If I explicitely install numpy and cython beforetvb-gdist everything works fine (see below). The thing is numpy and cython need to be installed before whereas scipy is gathered automatically. Hence, if I set tvb-gdist as a dependency of my project through the setup.py mechanism, the install crashes. A quick solution would be the user need to explicitely install numpy and cython first. But, if possible I would like to automatically trigger the install of tvb-gdist and its depencies using the pip install . or setup.py install mechanism. Any idea why the second way is crashing ? Thank you for your time
Hi @alexpron,
Made #39 for that. Please use the repo for now pip install git+https://github.com/the-virtual-brain/tvb-gdist
.
Thank you very much Ayan Banerjee for taking the time to solve this isue. I wish you all the best Yours faithfully Alexandre Pron
Le lun. 8 juin 2020 à 16:32, Ayan Banerjee notifications@github.com a écrit :
Hi @alexpron https://github.com/alexpron,
Made #39 https://github.com/the-virtual-brain/tvb-gdist/pull/39 for that. Please use the repo for now pip install git+ https://github.com/the-virtual-brain/tvb-gdist.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/the-virtual-brain/tvb-gdist/issues/19#issuecomment-640664493, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKY62L3EVFQALAV2DV7DLRTRVTZBNANCNFSM4FPTS2YA .
I just tested in a clean/empty env and I confirm that this is still happening for tvb-gdist version 2.0.1 (latest on Pypi) in any env where a whl does not exist and the tar.gz sources are used.
But as I just merged #38 from @ayan-b it works to temporarily use
pip install git+https://github.com/the-virtual-brain/tvb-gdist
In the next few days we will try to make a new release on Pypi, to officially fix there too. Thanks @alexpron for reporting and documenting this
I guess this is due to the import numpy statement inside the setup.py file, pyproject.toml file may be the most promising lead but for un know reason it does not work
Le lun. 8 juin 2020 à 22:37, Lia Domide notifications@github.com a écrit :
I just tested in a clean/empty env and I confirm that this is still happening for tvb-gdist version 2.0.1 (latest on Pypi) in any env where a whl does not exist and the tar.gz sources are used.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/the-virtual-brain/tvb-gdist/issues/19#issuecomment-640873172, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKY62L7B72WMETK27XROEFDRVVD2JANCNFSM4FPTS2YA .
Just retested now. Works perfectly fine in Ubuntu and MacOsX raw environnement. Thank you @ayan-b and @liadomide
Tried moving a notebook to binder and I wasn't able to do it because of the numpy requirement. Also had trouble with my local venv setup, but that was easy to solve