fix many deprecation warnings

[jkittner]

the warnings came mainly from marshmallow and cpython itself

• you cannot provide metadata via kwargs to scheams in marshmallow any longer
• validator must not return bools any longer - just raise an exception upon failure
• some minor cpython stuff e.g. in the ast module

In total that should have removed a couple hundred warnings - and I enabled them again, so we can actually see if something is deprecated.

There is a warning left with multiprocessing I cannot find the root cause of.

[jkittner]

I restored the ignores to the previous state:

we had 378 warnings on 3.12 before this patch. We are now down to 224 and with #364 we will be down to 216.

What remains are these warnings:

tests/drivers/test_raster_drivers.py: 14 warnings
tests/scripts/test_optimize_rasters.py: 182 warnings
tests/server/test_cors.py: 3 warnings
  /usr/lib/python3.12/multiprocessing/popen_fork.py:66: DeprecationWarning: This process (pid=152991) is multi-threaded, use of fork() may lead to deadlocks in the child.
    self.pid = os.fork()

tests/drivers/test_sqlite_remote.py: 17 warnings
  /home/kittnjdr/workspace/terracotta/.tox/py312/lib/python3.12/site-packages/botocore/auth.py:425: DeprecationWarning: datetime.datetime.utcnow() is deprecated and scheduled for removal in a future version. Use timezone-aware objects to represent datetimes in UTC: datetime.datetime.now(datetime.UTC).
    datetime_now = datetime.datetime.utcnow()

The boto one is clearly not our fault and needs fixing upstream - the former ones likely are and I believe are due to ThreadPoolExecutor?

[dionhaefner]

As long as tests are passing I'm not too worried about deprecation warnings. The majority of them will be fixed by other people (3rd party maintainers) when they become a problem. For the remainder, we'll deal with them when they actually break something; at that point there are usually enough guides on how to update that this transition is much easier to deal with.

[dionhaefner]

So yes, the former is likely a consequence of the fact that we use a ProcessPoolExecutor in code that is multithreaded, but when the time comes and this actually gets deprecated there will likely be an official alternative to use (that e.g. avoids using fork under the hood).

[dionhaefner]

Good fixes, thanks for the cleanup!

[jkittner]

btw. marshmallow 4.0.0 release 10 hours ago (https://pypi.org/project/marshmallow/4.0.0/) so version 0.9.0 currently does not install cleanly w/o specifying marshmallow<4.0.0. This patch/the main branch should have already fixed that.