feat(client): add support for openai responses.parse

[jerryjrxie]

https://platform.openai.com/docs/guides/structured-outputs?api-mode=responses

---

[!IMPORTANT] Add support for parsing OpenAI responses using parse method in both sync and async contexts, with tests verifying functionality.

• Behavior:
  • Add support for openai.resources.responses.Responses.parse and openai.resources.responses.AsyncResponses.parse in langfuse/openai.py.
  • Handles synchronous and asynchronous parsing of OpenAI responses.
• Tests:
  • Add test_responses_parse() in tests/test_openai.py to verify parsing functionality, including checking the parsed output and Langfuse tracking.
• Misc:
  • Update OpenAiDefinition list in langfuse/openai.py to include new parse methods.

^{This description was created by for 31f2c3a1fec471658fd5834135907af376e1a36d. You can customize this summary. It will automatically update as commits are pushed.}

Greptile Summary

Disclaimer: Experimental PR review

Added support for OpenAI's responses.parse functionality in the Python SDK, enabling structured output parsing with integrated Langfuse tracking.

• Added OpenAiDefinition entries for responses.parse and AsyncResponses.parse methods with OpenAI SDK v1.66.0+ requirement in langfuse/openai.py
• Implemented response handling logic in langfuse/openai.py to extract and process parsed outputs while maintaining existing tracking patterns
• Added comprehensive test case test_responses_parse() in tests/test_openai.py to validate structured output parsing into Pydantic models
• Preserved backward compatibility and consistent error handling with existing OpenAI integration features

_{💡 (1/5) You can manually trigger the bot by mentioning @greptileai in a comment!}

[CLAassistant]

All committers have signed the CLA.

[hassiebp]

Thanks for your contribution! This is already part of our latest Python SDK 👍🏾