chore(versions): Upgrade minimum python version

[jorwoods]

(Redo of #1461, which got messed up when I was trying to rebase development)

As of October, 2024, Python 3.8 is out of supportend of life. Upgrading syntax to target Python 3.9. Adds builds for Python 3.13.

[jorwoods]

@bcantoni Apologies, the other PR got closed when I was trying to fix my branches. This one has the same changes successfully rebased on development. I have noticed that the tests appear to run considerably faster with this branch and the syntax upgraded.

[jorwoods]

Copying comment from the original PR here:

I'm a little concerned about the volume of changes, especially with all those format strings. Was that a scripted change you made? (I.e. how confident are we nothing broke?)

This was a scripted change. In a year when the minimum version is bumped to 3.10, the volume of change should be lower. It is a lot of change, but I have noticed that some areas of the codebase use f-strings, some use .format, and some use the older python2 style %s string substitution. So while it is a lot of change, hopefully it brings more consistency in these small details across the codebase. Personally, I find the f-strings easier to read. f-strings are also faster than .format or %s, but not in a way that will be meaningful with the main purpose being around API calls.

By upgrading the syntax to target 3.9, will the library still work for 3.8 or will this be a breaking change?

This would be a breaking change, but not because of the f-strings. The main syntax change I see that will cause it to not work for 3.8 would be the change from Dict to dict, List to list, Set to set, and Tuple to tuple.

My comment about 3.8 being out of support is that the Python Software Foundation will stop releasing security fixes for 3.8 in October 2024, as it is considered "End of Life." Obviously we could continue to support older versions of python past their end of life, but I think it would hold back adopting newer syntax, features, and quality of life changes.

[jorwoods]

I reviewed all of the changes made by the script. The changes fall into a few main categories:

• Converting .format and %s to f-strings
• Removing unnecessary explicit inheritance from object
• Removing unnecessary explicit parameters being passed to super
• Switching to shorthand to define sets instead of creating a collection then passing to the set constructor
• Changing errors to OSError.

I tested the error handling locally, and it shouldn't have any impact to end users. The following reproducible example demonstrates these are interchangeable now.

try: raiseIOErrorexceptOSError: print('caught IOError') exceptException: print('whoops') try: raiseOSErrorexceptIOError: print('caught OSError') exceptException: print('whoops') try: raiseEnvironmentErrorexceptOSError: print('caught EnvironmentError') exceptException: print('whoops') try: raiseOSErrorexceptEnvironmentError: print('caught OSError') exceptException: print('whoops')

[jorwoods]

@jacalata I went through and fixed the f-strings the script missed. Should be good to go now.