- Set environment variables before importing DEFAULT_CONFIG to prevent hanging
- Add MockResult class for proper tool_calls handling
- Add error handling in market_analyst for Mock objects
- Remove temporary test files
- bind_tools() now returns a mock chain
- chain.invoke() returns the result with tool_calls attribute
- Fixes TypeError: object of type 'Mock' has no len()
- Apply Black formatting to all test files
- Fix Mock objects to include tool_calls attribute for len() checks
- Add proper __name__ attributes to mock toolkit methods for @tool decorator
- Create mock_toolkit_fix helper for consistent toolkit mocking
All tests should now pass with proper mocking setup.
- Updated mock_toolkit fixture to create proper mock functions with __name__ attributes
- Fixed TypeError/ValueError issues where Mock objects were passed to tool decorators
- Downgraded numpy to 1.26.4 and pandas to 2.1.4 to resolve import performance issues
- Added test scripts to verify mock fixes are working correctly
The mock functions now properly implement:
- __name__ attribute for tool decorator compatibility
- name attribute for tool name extraction
- Callable interface with proper return values
- Mock tracking capabilities (called, call_count, etc.)
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
佐藤優一