OpenAI o1 model support

[marklysze]

Why are these changes needed?

OpenAI has just released o1-mini and o1-preview, these changes add support for the model.

I don't have level 5 API access to test against the API, if anyone can test with the models that would be great.

Notes:

• OpenAI is depreciating max_tokens and max_completion_tokens should be used. Code maps max_tokens to max_completion_tokens.
• Beta limitations

Tests still TODO.

Related issue number

Checks

• I've included any doc changes needed for https://autogen-ai.github.io/autogen/. See https://autogen-ai.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.

[BabyCNM]

LGTM. I cannot test o1 yet because of limited access. However, the implementation looks great.

A lint question: should we rename "map_params" to "_map_params" for readability?

[BabyCNM]

The 'num_tokens' change and test updates are good to have.

[sonichi]

@marklysze could you check the build error? It indicates some problem with the token counting.

[yiranwu0]

@marklysze could you check the build error? It indicates some problem with the token counting.

This should be related to the change in count_token function.

[marklysze]

@marklysze could you check the build error? It indicates some problem with the token counting.

This should be related to the change in count_token function.

All good, committing correction shortly

Update: committed

[sonichi]

@qingyun-wu could you make sure there's an o1 endpoint in the config list used by the openai CI?
@BabyCNM has your change request been addressed?