search

oemof / oemof-tabular

Load oemof energy systems from tabular data sources.
https://oemof-tabular.readthedocs.io/
BSD 3-Clause "New" or "Revised" License
7 stars 5 forks source link

Replace all appearences of pkg_resources except one #101

Closed fwitte closed 1 year ago

fwitte commented 1 year ago

Ran the tests, all working fine except for the test_examples_datapackages_scripts_infer. There the original implementation already fails locally for me. So I did not touch that.

See #100

fwitte commented 1 year ago

Apparently importlib.resources.files is only available since python 3.9 (https://docs.python.org/3/library/importlib.resources.html).

fwitte commented 1 year ago

So we would need to drop support for python 3.8 and lower. Python 3.8 is the default version of ubuntu 20.04 LTS, which will be around for another two years, I guess... I do not really know, how to feel about that. In my opinion it might make sense to move the lower python version limit to 3.9, but that might break things for a couple of people. What do you think?

fwitte commented 1 year ago

So, all tests (except the one that did not work in the first place) are working with python 3.9 and 3.10. Python 3.11 has installation issues with cchardet (https://github.com/oemof/oemof-tabular/actions/runs/4891251104/jobs/8731565175).

So the question remains, whether to drop support of older versions. I'll let you guys decide on this, since I do not (yet) use the software...

p-snft commented 1 year ago

Python 3.8 is the default version of ubuntu 20.04 LTS

Honestly, that's the reason why I dist-upgraded.

henhuy commented 1 year ago

I also vote for dropping support for python 3.8

nailend commented 1 year ago

I also vote for dropping support for python 3.8

fwitte commented 1 year ago

Anybody has an idea, what the reason for the failing test is?

nailend commented 1 year ago

Anybody has an idea, what the reason for the failing test is?

which failing tests? I guess this got somehow solved in another issue?!

fwitte commented 1 year ago

Anybody has an idea, what the reason for the failing test is?

which failing tests? I guess this got somehow solved in another issue?!

... I oversaw it was fixed by merging dev into here 😀

Also, there should be some kind of warning before releasing this into the wild, right? Suddenly disappearing support for 3.8 might be inconvenient.

nailend commented 1 year ago

Also, there should be some kind of warning before releasing this into the wild, right? Suddenly disappearing support for 3.8 might be inconvenient.

Yes, that's right. I'm not sure how many active users there are, though. Following the release convention, we haven't even released a feature, only bug fixes :D, so there is no stable version yet. v0.0.3

How is this normally handled? You would raise a warning in another release? Or is it maybe fine to set a tag before and mention it in the warning?

fwitte commented 1 year ago

@nailend: since this has been merged as well there is no more point in the warning message :D. People with python < 3.9 will not be able to install the latest version anymore. I would suggest downgrade the python requirement again, make a release, upgrade it again and make another release 2 or so weeks later.