fix: make ChatGPT_API_with_finish_reason return consistent tuple

[luojiyin1987]

Summary

Fixes #62 - Make ChatGPT_API_with_finish_reason return consistent tuple type on error.

Problem

The function returned (content, finish_reason) on success but "Error" string when max retries reached, causing unpacking errors.

Change

• Return ("", "error") tuple instead of "Error" string
• Maintains consistent return type for all code paths

Test

Manual testing of error scenarios.

[BukeLy]

Code review

No issues found. Checked for bugs and CLAUDE.md compliance.

The fix correctly addresses the inconsistent return type: all callers use response, finish_reason = ChatGPT_API_with_finish_reason(...) tuple unpacking, so the previous return "Error" string would cause a ValueError. Returning ("", "error") maintains type consistency across all code paths.

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}