The _get_current_difficulty() method returns a DifficultyLevel enum,
but Pydantic v2 requires the string value for serialization. This fix
converts the enum to its .value before passing to HallucinationObservation.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Animated particle system background with floating geometric objects
- Glassmorphism UI elements with gradient effects
- Interactive playground with live API testing
- Animated stat counters and smooth transitions
- Streamlined Python code for better readability
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Enhanced server/grader.py with improved reward calculation logic
- Expanded test coverage in test_grader.py, test_environment.py, test_dataset_loader.py
- Updated README.md with additional documentation
- Improved Dockerfile configuration
- Updated dependencies in requirements.txt and pyproject.toml
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Rewrite test_grader.py to use actual calculate_reward() function
(previously referenced non-existent HallucinationGrader class)
- Fix openenv.yaml port: 8000 -> 7860 (matches HF Spaces default)
- Add openai>=1.0.0 to requirements.txt (needed by inference.py)
- Update baseline scores in inference.py comments to match actual results
All 17 tests now pass.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
