You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I played around a bit with the mentions_pipeline_notebook repo. The “requirements.txt” error was coming from the setup.py specifying software_mentions_client==0.1.0. If that changes to software_mentions_client==0.1.6 then we get past that bit.
That said, it then runs into an error that i think happens because python 3.7 is default (for repo2docker, at least with the version on pypy currently. This causes the require_python='>=3.8' to not pass, stopping the build.
Looks like the way forward would be to specify python > 3.8 in an environment.yml file (so that conda handles that install).
We might move the python dependencies into the requirements.txt file. I'm wondering whether a module for this notebook makes sense anyway? Can't see that anyone would install it as a module?
The text was updated successfully, but these errors were encountered:
We might move the python dependencies into the requirements.txt file. I'm wondering whether a module for this notebook makes sense anyway? Can't see that anyone would install it as a module?
Yes you're right ! I have created a setup.py more as a routine I think but this is not relevant here. We can just have a requirements.txt. I think the default python 3.7 is working well for these modules, so we can also just forget about a minimum requirement for python (I think the python version can be set in a file runtime.txt coming with the requirements.txt, or everything in environment.yml with conda)
I played around a bit with the mentions_pipeline_notebook repo. The “requirements.txt” error was coming from the setup.py specifying software_mentions_client==0.1.0. If that changes to software_mentions_client==0.1.6 then we get past that bit.
That said, it then runs into an error that i think happens because python 3.7 is default (for repo2docker, at least with the version on pypy currently. This causes the require_python='>=3.8' to not pass, stopping the build.
I poked around a bit: https://discourse.jupyter.org/t/can-repo2docker-resolve-python-requires-in-setup-py/18599 and it seems unlikely that repo2docker will ever read the python version out of setup.py (or other python packaging specifications).
Looks like the way forward would be to specify python > 3.8 in an environment.yml file (so that conda handles that install).
We might move the python dependencies into the requirements.txt file. I'm wondering whether a module for this notebook makes sense anyway? Can't see that anyone would install it as a module?
The text was updated successfully, but these errors were encountered: