feat: add retry logic and improve logging in OpenAIClient

[danielchalef]

Important

Enhance OpenAIClient error handling with retry logic and improved logging in generate_response().

• Error Handling:
  • Modify generate_response() in openai_client.py to include retry logic for exceptions, with a maximum of 2 retries.
  • Handle RateLimitError, RefusalError, APITimeoutError, APIConnectionError, and InternalServerError differently, with specific exceptions not triggering retries.
• Logging:
  • Log detailed error messages and retry attempts in generate_response().
  • Update exception message in _generate_response() to include response_object.model_dump() details.
• Constants:
  • Add MAX_RETRIES as a class-level constant in OpenAIClient.

^{This description was created by for 769bf5f. It will automatically update as commits are pushed.}

[ellipsis-dev]

👍 Looks good to me! Reviewed everything up to 769bf5f in 1 minute and 26 seconds

More details

• Looked at 78 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 4 drafted comments based on config settings.

1. graphiti_core/llm_client/openai_client.py:22

• Draft comment:
  AsyncOpenAI is imported but not used. Consider removing it to clean up the imports.
• Reason this comment was not posted:
  Confidence changes required: 10%
  The import of AsyncOpenAI is not used in the code, which is unnecessary and should be removed.

2. graphiti_core/llm_client/openai_client.py:156

• Draft comment:
  Appending error messages to messages on each retry can lead to an ever-growing list, potentially affecting performance or logic. Consider resetting or managing the list size.
• Reason this comment was not posted:
  Confidence changes required: 80%
  The retry logic in generate_response appends an error message to the messages list on each retry, which could lead to an ever-growing list and potentially affect performance or logic. This should be addressed.

3. graphiti_core/llm_client/openai_client.py:120

• Draft comment:
  The generate_response method is handling retries and error logging. Consider breaking it down into smaller methods to adhere to the Single Responsibility Principle.
• Reason this comment was not posted:
  Confidence changes required: 80%
  The code is mostly well-structured, but there are some areas that could be improved for better readability and maintainability.

4. graphiti_core/llm_client/openai_client.py:108

• Draft comment:
  The exception message 'Invalid response from LLM' could be more descriptive. Consider including more context about the error.
• Reason this comment was not posted:
  Confidence changes required: 60%
  The code is mostly well-structured, but there are some areas that could be improved for better readability and maintainability.

Workflow ID: wflow_fzHjZt0cQMCnieDs

---

You can customize Ellipsis with 👍 / 👎 feedback, review rules, user-specific overrides, quiet mode, and more.