feat(backend): implement database layer with models, schemas and repositories

Complete backend core implementation (BE-001 to BE-005):

BE-001: Database Connection & Session Management
- Create src/core/database.py with async SQLAlchemy 2.0
- Configure engine with pool_size=20
- Implement get_db() FastAPI dependency

BE-002: SQLAlchemy Models (5 models)
- Base model with TimestampMixin
- Scenario: status enum, relationships, cost tracking
- ScenarioLog: message hash, PII detection, metrics
- ScenarioMetric: time-series with extra_data (JSONB)
- AwsPricing: service pricing with region support
- Report: format enum, file tracking, extra_data

BE-003: Pydantic Schemas
- Scenario: Create, Update, Response, List schemas
- Log: Ingest, Response schemas
- Metric: Summary, CostBreakdown, MetricsResponse
- Common: PaginatedResponse generic type

BE-004: Base Repository Pattern
- Generic BaseRepository[T] with CRUD operations
- Methods: get, get_multi, count, create, update, delete
- Dynamic filter support

BE-005: Scenario Repository
- Extends BaseRepository[Scenario]
- Specific methods: get_by_name, list_by_status, list_by_region
- Business methods: update_status, increment_total_requests, update_total_cost
- ScenarioStatus enum
- Singleton instance: scenario_repository

All models, schemas and repositories tested and working.

Tasks: BE-001, BE-002, BE-003, BE-004, BE-005 complete

src/core/__init__.py

@@ -0,0 +1,5 @@
+"""Core utilities and configurations."""
+
+from src.core.database import Base, engine, get_db, AsyncSessionLocal
+
+__all__ = ["Base", "engine", "get_db", "AsyncSessionLocal"]