Add retry mechanism to LLMFactory and BaseLLM

[Undertone0809]

Related to #651

Add retry mechanism to LLMFactory and LiteLLM classes to increase robustness.

• LLMFactory:

  • Add max_retry parameter to LLMFactory.build method in promptulate/llms/factory.py.
  • Pass max_retry parameter to LiteLLM constructor.

• LiteLLM:

  • Add max_retry attribute to LiteLLM class in promptulate/llms/_litellm.py.
  • Modify _predict method to include retry logic.

• AIChat:

  • Add max_retry parameter to AIChat constructor in promptulate/chat.py.
  • Pass max_retry parameter to LLMFactory.build method.

• Tests:

  • Add tests in tests/llms/test_factory.py to verify retry mechanism.
  • Mock LiteLLM to simulate failure and retry behavior.

---

For more details, open the Copilot Workspace session.

Summary by CodeRabbit

• New Features

  • Introduced a max_retry parameter across various functionalities, enhancing user control over retry attempts during LLM initialization and chat operations.
  • Improved error handling and robustness in the LiteLLM prediction process with a retry mechanism.

• Bug Fixes

  • Enhanced resilience in LLM interactions by implementing a retry logic for failed attempts.

• Tests

  • Added new tests to validate the retry mechanisms for the LiteLLM model, ensuring robust error handling.