- Updated `conftest.py` to set fake API keys only if real ones are not available, enhancing test flexibility.
- Refactored the `mock_llm_calls` fixture to conditionally mock LLM calls based on the use of real APIs, improving test isolation and reliability.
- Ensured that when real APIs are used, the mocks return empty values to allow tests to run without interference.
These changes enhance the robustness of the testing framework and ensure better handling of API dependencies.
- Added environment variables for fake API keys in `conftest.py` to facilitate testing.
- Refactored mock functions in tests to utilize global mocks for LLM calls, improving consistency and reducing redundancy.
- Updated various test cases to rely on the new mocking structure, ensuring accurate simulation of LLM responses and fact extraction.
- Removed direct calls to model clients in tests, replacing them with mocks to streamline test execution and maintain isolation.
hyusap