fix: No context vars for async agents replies

[violonistahiles]

Why are these changes needed?

The Autogen currently utilizes the asyncio.get_event_loop().run_in_executor() method for generating async agent replies. However, this approach presents a challenge if the application decorates OpenAI calls (such as openai.resources.chat.completions.AsyncCompletions.create) for purposes like call tracing or statistics measurement. The problem is that context variables cannot be accessed in this setup.

To address this issue, suggested to use contextvars.copy_context().run() method. This change will allow context variables to be properly propagated and accessed.

Related issue number

Checks

• [ ] I've included any doc changes needed for https://microsoft.github.io/autogen/. See https://microsoft.github.io/autogen/docs/Contribute#documentation to build and test documentation locally.
• [ ] I've added tests (if relevant) corresponding to the changes introduced in this PR.
• [ ] I've made sure all auto checks have passed.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 29.30%. Comparing base (8a8fcd8) to head (61b70b9). Report is 4 commits behind head on 0.2.

Additional details and impacted files

@@           Coverage Diff           @@
##              0.2    #4640   +/-   ##
=======================================
  Coverage   29.30%   29.30%           
=======================================
  Files         117      117           
  Lines       13013    13021    +8     
  Branches     2469     2470    +1     
=======================================
+ Hits         3813     3816    +3     
- Misses       8854     8859    +5     
  Partials      346      346           

Flag	Coverage Δ
unittests	29.30% <100.00%> (+<0.01%)	:arrow_up:

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

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.