Add support for Pydantic model validation/serialization (fixes #8751)

[mathiasertl]

Allow validation of arguments using Pydantic, as well as result serialization. This PR would fix (IMHO) #8751.

With this PR, you can pass pydantic=True to a task decorator. If set, any argument type hinted as a Pydantic model will be validated (note: this means data validation and loading into the model class Pydantic speak) to that model, and any result matching a return type annotation will automatically be converted to a dict (using model_dump()).

With this feature, tasks look a lot more like FastAPI endpoints, where query arguments, request and response bodies are validated/serialized using Pydantic. The advantage is of course automatic data validation and safer typing. mypy will catch type hinting errors, and IDEs auto-complete values in the models.

[codecov]

Codecov Report

Attention: Patch coverage is 76.66667% with 7 lines in your changes are missing coverage. Please review.

Project coverage is 77.81%. Comparing base (4cf6ba3) to head (774e677).

Files	Patch %	Lines
celery/app/base.py	76.66%	7 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #9023      +/-   ##
==========================================
+ Coverage   77.80%   77.81%   +0.01%     
==========================================
  Files         150      150              
  Lines       18682    18711      +29     
  Branches     3194     3201       +7     
==========================================
+ Hits        14535    14560      +25     
- Misses       3854     3860       +6     
+ Partials      293      291       -2     

Flag	Coverage Δ
unittests	77.79% <76.66%> (+0.01%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[mathiasertl]

Note: I force-pushed a few minor changes to make sure that test coverage does not decrease.