Remove dependency on the Pytz library

[spillner]

Pytz no longer supports the latest Python releases, and not much of it was actually used by OWSLib anyway.

Recreated the previous PR #921 to permit branching my source repository, to better manage the rebase and elide false-positive merge conflicts reported by GitHub.

[coveralls]

coverage: 60.113% (-0.02%) from 60.128% when pulling 5e8f39ade909cf0ecbf95c74d41a636267fc0999 on spillner:remove-pytz into d4e7080531cd4805ad23c7cd7829e53f15afdbcc on geopython:master.

[geographika]

Thanks @spillner for updating the pull request. Would it be possible to add a very simple test to test_util.py to parse a date using the new TimeZone_UTC class and check the output?

I'm +1 on merging once that is in place.

[geographika]

Tested the branch locally - all working as expected:

>>> import pytz
>>> from datetime import datetime, timedelta
>>> dt = datetime.utcnow()
>>> dt.replace(tzinfo=pytz.utc)
datetime.datetime(2024, 10, 8, 10, 7, 59, 167706, tzinfo=<UTC>)
>>> from owslib.util import TimeZone_UTC
>>> datetime.utcnow().replace(tzinfo=TimeZone_UTC())
datetime.datetime(2024, 10, 8, 10, 9, 31, 262084, tzinfo=<owslib.util.TimeZone_UTC object at 0x0000015D9D187CD0>)

[spillner]

Added two unit tests, one for datetime normalization as a standalone operation, and one for proper operation of extract_time() as a whole.

I also took advantage of the shift to 3.10 and newer to drop the TimeZone_UTC class in favor of the standard datetime.timezone.utc. If anyone needs to restore support for 3.8 and earlier, you could just roll back the very last commit in this series.

[geographika]

@spillner - thanks very much for this. Removing a dependency and being able to use the standard library is excellent.