Separate openai assistant related config items from llm_config

[IANTHEREAL]

Why are these changes needed?

openai assistant related config in llm_config will cause some function not working, like

import logging
import os

from autogen import AssistantAgent, UserProxyAgent, config_list_from_json
from autogen.agentchat.contrib.gpt_assistant_agent import GPTAssistantAgent

logger = logging.getLogger(__name__)
logger.setLevel(logging.WARNING)

assistant_id = os.environ.get("ASSISTANT_ID", None)

config_list = config_list_from_json("OAI_CONFIG_LIST", filter_dict={"tags": ["assistant"]})
llm_config = {"config_list": config_list, "assistant_id": assistant_id}

gpt_assistant = GPTAssistantAgent(
    name="assistant", instructions=AssistantAgent.DEFAULT_SYSTEM_MESSAGE, llm_config=llm_config
)

user_proxy = UserProxyAgent(
    name="user_proxy",
    code_execution_config={
        "work_dir": "coding",
        "use_docker": False,
    },  # Please set use_docker=True if docker is available to run the generated code. Using docker is safer than running the generated code directly.
    is_termination_msg=lambda msg: "TERMINATE" in msg["content"],
    human_input_mode="NEVER",
    max_consecutive_auto_reply=1,
)
result = user_proxy.initiate_chat(gpt_assistant, message="Print hello world", summary_method="reflection_with_llm")
print(result.summary)

Related issue number

Closes #1805

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

Attention: Patch coverage is 12.50000% with 21 lines in your changes are missing coverage. Please review.

Project coverage is 48.41%. Comparing base (8844f86) to head (6d79368).

Files	Patch %	Lines
autogen/agentchat/contrib/gpt_assistant_agent.py	12.50%	21 Missing ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##             main    #1964       +/-   ##
===========================================
+ Coverage   36.99%   48.41%   +11.42%     
===========================================
  Files          66       66               
  Lines        7015     7031       +16     
  Branches     1534     1666      +132     
===========================================
+ Hits         2595     3404      +809     
+ Misses       4194     3341      -853     
- Partials      226      286       +60     

Flag	Coverage Δ
unittests	48.28% <12.50%> (+11.29%)	⬆️

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.

[IANTHEREAL]

To fix the CI error, should we replace np.Inf with np.inf in the code?

ERROR samples/tools/finetuning/tests/test_conversable_agent_update_model.py - AttributeError: `np.Inf` was removed in the NumPy 2.0 release. Use `np.inf` instead.

[ekzhu]

To fix the CI error, should we replace np.Inf with np.inf in the code?

ERROR samples/tools/finetuning/tests/test_conversable_agent_update_model.py - AttributeError: `np.Inf` was removed in the NumPy 2.0 release. Use `np.inf` instead.

This is fixed in the latest release

[jtrugman]

LGTM - looks like we can extend this to add the feature of utilizing OpenAI Assistants API Threads as well 😄

[gagb]

Please double-check my understanding, other wise looks good! great pr @IANTHEREAL

[ekzhu]

There is an openai error: https://github.com/microsoft/autogen/actions/runs/8302792055/job/22725770776?pr=1964#step:7:46

[sonichi]

Could you make the PR from the upstream repo because we use pull_request now as the trigger for openai workflows?

[IANTHEREAL]

Could you make the PR from the upstream repo because we use pull_request now as the trigger for openai workflows?

@sonichi I don't catch up with you, do you mean make a new PR using microsoft/autogen as upstream?

[sonichi]

Could you make the PR from the upstream repo because we use pull_request now as the trigger for openai workflows?

@sonichi I don't catch up with you, do you mean make a new PR using microsoft/autogen as upstream?

Yes. create a branch in microsoft/autogen and make a PR from there. Otherwise the openai test will fail.