Implement Unit Tests and LangFuse Integration

[Haydart]

Task 010: Implement Unit Tests and LangFuse Integration

Description

Create comprehensive unit tests for all components and integrate LangFuse observability for LLM monitoring.

Acceptance Criteria

• Create tests/unit/domain/model/test_trend_models.py for all Pydantic models
• Create tests/unit/services/test_simple_firebase_service_trends.py for Firebase methods
• Create tests/unit/services/test_trend_cache.py for cache service
• Create tests/unit/domain/processor/stages/test_trend_classification_stage.py
• Create tests/unit/domain/processor/stages/test_new_trend_creation_stage.py
• Create tests/unit/domain/processor/stages/test_trend_action_stage.py
• Create tests/unit/domain/processor/test_trend_detection_processor.py
• Create tests/unit/domain/executor/test_trend_firebase_executor.py
• Integrate LangFuse in TrendClassificationStage (follow existing pattern)
• Integrate LangFuse in NewTrendCreationStage (follow existing pattern)
• Mock LLM responses at HTTP layer (not service layer)
• All tests pass with pytest
• LangFuse tracks classification latency, token usage, costs

Technical Details

• Test location: tests/unit/
• Use existing test patterns from codebase
• Mock Firebase at client level, not service level
• Mock LLM at HTTP layer (litellm library level)
• No mock services - test real service logic
• Verbose tests for debugging capability
• LangFuse integration: copy pattern from existing LLM usage
• Track: model name, latency, input/output tokens, cost per call

Effort Estimate

• Size: L
• Hours: 12-16
• Parallel: false (depends on all implementation tasks)