I installed this branch's pipenv environment like this:

pipenv --rm
pipenv install

And confirmed that I was getting the right version of Python:

$ pipenv run python -V
Python 3.11.9

I ran into problems running pipenv using a Python 3.11.9 conda environment for some reason, but I think as long as the pipenv command above is reporting Python 3.11.9, everything is working as expected.

I then ran the Data API from this branch and pointed local instances of the following webapps against it:

• Northern Climate Reports
• Arctic-EDS
• Historical Sea Ice Atlas
• ARDAC Explorer

And everything is working as it should! This covers a ton of test cases all at once, so I think all functionality survived the Python upgrade (and removal of the netcdf4 package too).

I searched through the code and the only references I can see to the netcdf4 package are in the Pipfile, Pipfile.lock, and requirements.txt files. So, I think this was probably a package we stopped using and forgot to remove from the Pipfile. That would explain why all functionality seems to be intact without it.

So, everything looks good with this branch! But we may need to also generate a new requirements.txt file with the following command to make sure our Elastic Beanstalk instance uses package versions compatible with Python 3.11.9:

pip freeze > requirements.txt

I'll defer to @BobTorgerson on this since he has more experience with this kind of thing.