- Fixed field extraction logic in test_field_schema_validation.py to properly parse methods with docstrings
- Previous regex was too greedy and matched across multiple method definitions
- Now uses proper parenthesis and docstring matching to isolate method bodies
- Correctly handles both 'fields = [...]' and 'fields = common_fields + [...]' patterns
- Updated db_schema.sql to include missing columns:
- campaign_insights: added frequency, cpp, cost_per_action_type columns
- adset_insights: added account_currency column
- campaign_insights_by_country: added frequency, cpp, cost_per_action_type columns
- All field schema validation tests now pass
- Test dynamically extracts fields from scheduled_grabber.py source code
- Compares against actual database schema from db_schema.sql
- Properly filters metadata-only fields (campaign_id, campaign_name, etc.)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
