angelosalatino / cso-classifier

Python library that classifies content from scientific papers with the topics of the Computer Science Ontology (CSO).
https://cso.kmi.open.ac.uk
Apache License 2.0
84 stars 18 forks source link

Bump `igraph` version #13

Open pontakornth opened 1 year ago

pontakornth commented 1 year ago

Currently, CSO classifier is using an outdated version of python-igraph. That results in error in the installation step.

I come from the python-igraph repository https://github.com/igraph/python-igraph/issues/654

angelosalatino commented 1 year ago

Thanks a lot for the heads up, I will fix this over the week end

angelosalatino commented 1 year ago

I have fixed and tested the version on github.

I wil need some more time to create the builds and release them on pypi. If you want, please install this version here.

thoughtpoet commented 1 year ago

I get python-igraph failures in most scenarios (python 3.8, 3.9 3.10).

(base) thoughtpoet@Freds-MBP-2 py36 % python --version Python 3.6.15 (base) thoughtpoet@Freds-MBP-2 py36 % pip install cso-classifier
...
File "/var/folders/4c/4fgkkjz57sv92py6kdgpkw4c0000gn/T/easy_install-5fhnojyo/numpy-1.24.2/setup.py", line 20, in PY2 = sys.version_info[0] == 2 RuntimeError: Python version >= 3.8 required.

----------------------------------------

Command "python setup.py egg_info" failed with error code 1 in /private/var/folders/4c/4fgkkjz57sv92py6kdgpkw4c0000gn/T/pip-install-1hghb74r/gensim/

From downloads on 3.6.15, I get: ERROR: Cannot install cso-classifier==2.1, cso-classifier==2.2, cso-classifier==2.3, cso-classifier==2.3.1, cso-classifier==2.3.2 and cso-classifier==3.0 because these package versions have conflicting dependencies.

The conflict is caused by: cso-classifier 3.0 depends on spacy==3.0.5 cso-classifier 2.3.2 depends on pandas==0.24.2 cso-classifier 2.3.1 depends on pandas==0.24.2 cso-classifier 2.3 depends on pandas==0.24.2 cso-classifier 2.2 depends on pandas==0.24.2 cso-classifier 2.1 depends on pandas==0.24.2

I've tried with anaconda and then with pyenv. What's the best way to build? Do you have locked dependency versions that will work for any of the releases?
angelosalatino commented 1 year ago

which version are you trying to install? The one on via pypi or the one on github.

I am asking because the one on github is the latest version. I need to publish it on pypi. And this version on github, should have a better version of dependencies.

thoughtpoet commented 1 year ago

Both!!!! I have tried github for every version pip install failed on (3.6.3, 3.6.15, 3.8, 3.9, 3.10).

On Sun, Apr 16, 2023 at 5:36 AM Angelo Antonio Salatino < @.***> wrote:

which version are you trying to install? The one on via pypi or the one on github.

I am asking because the one on github is the latest version. I need to publish it on pypi. And this version on github, should have a better version of dependencies.

— Reply to this email directly, view it on GitHub https://github.com/angelosalatino/cso-classifier/issues/13#issuecomment-1510361765, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB5XWT4PDP452I2ZHBSQOB3XBPRWJANCNFSM6AAAAAAWMT3G6A . You are receiving this because you commented.Message ID: @.***>

angelosalatino commented 1 year ago

Which operative system are you on?

thoughtpoet commented 1 year ago

Sorry, I should have said that. MacOS Ventura 23.2.1 (Apple M1). I have tried pyenv for 3.8 and below, and conda environments for 3.8, 3.9, and 3.10.

On Sun, Apr 16, 2023 at 8:21 AM Angelo Antonio Salatino < @.***> wrote:

Which operative system are you on?

— Reply to this email directly, view it on GitHub https://github.com/angelosalatino/cso-classifier/issues/13#issuecomment-1510411980, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB5XWT7XLVQST3HZHLHPTRDXBQE77ANCNFSM6AAAAAAWMT3G6A . You are receiving this because you commented.Message ID: @.***>

angelosalatino commented 1 year ago

Ok. perfect. I will run some tests and keep you posted.

Best

thoughtpoet commented 1 year ago

Thanks, Antonio, I’d love to try this out.

On Sun, Apr 16, 2023 at 8:34 AM Angelo Antonio Salatino < @.***> wrote:

Ok. perfect. I will run some tests and keep you posted.

Best

— Reply to this email directly, view it on GitHub https://github.com/angelosalatino/cso-classifier/issues/13#issuecomment-1510414820, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB5XWTYOTGZUYPWQNTQN3H3XBQGSBANCNFSM6AAAAAAWMT3G6A . You are receiving this because you commented.Message ID: @.***>

angelosalatino commented 1 year ago

Hi Fred,

I had the chance to get my hands on a Mac OS Ventura on M1 (so similar to yours).

I tested both python 3.8 and python 3.7. I was able to install the cso classifier via pip install cso-classifier without issues. It seems I can't reproduce your errors.

Besides, based on the log you provided, I realised that you are installing the cso-classifier on a base environment. I suggest you to try on a brand new environment with the latest version of pip.

thoughtpoet commented 1 year ago

Thanks, I will look into it further. I appreciate your effort, and my situation seems to be environment-related. I did accidentally proceed from a base environment initially, but switching didn’t seem to solve my issues. I will let you know how it goes.

On Mon, Apr 17, 2023 at 11:43 AM Angelo A. Salatino < @.***> wrote:

Hi Fred,

I had the chance to get my hands on a Mac OS Ventura on M1 (so similar to yours).

I tested both python 3.8 and python 3.7. I was able to install the cso classifier via pip install cso-classifier without issues. It seems I can't reproduce your errors.

Besides, based on the log you provided, I realised that you are installing the cso-classifier on a base environment. I suggest you to try on a brand new environment with the latest version of pip.

— Reply to this email directly, view it on GitHub https://github.com/angelosalatino/cso-classifier/issues/13#issuecomment-1511899763, or unsubscribe https://github.com/notifications/unsubscribe-auth/AB5XWT3U62TSRN4KEEZ6MUDXBWFOFANCNFSM6AAAAAAWMT3G6A . You are receiving this because you commented.Message ID: @.***>