refactor(types): add comprehensive type annotations across backend Python codebase

Enable ANN rules in ruff.toml (flake8-annotations) and resolve all 221 violations:

ANN201/ANN202 — return types on 168 public/private functions:
- All 28 FastAPI routers: endpoints annotated with dict/list/specific schema/
  StreamingResponse/FileResponse/JSONResponse as appropriate
- main.py: lifespan→AsyncGenerator[None,None], exception handlers→JSONResponse
- database.py: get_db→Generator[Session,None,None], proxy methods→correct types
- middleware/request_context.py: dispatch→Response with Callable call_next type

ANN001/ANN002/ANN003 — 32 missing argument types:
- seed_demo.py: all db parameters typed as Session
- domain/unit_of_work.py: __aexit__ exc_type/exc_val/exc_tb typed with TracebackType
- services: audit_service user_id→UUID|None, heatmap_service query/model/builder,
  notification_service test→Test, tempo_service test→Test/user→User,
  test_workflow_service test_id→UUID, campaign_crud **fields→object,
  test_crud **fields→object (4 sites)

ANN401 — 16 Any usages resolved:
- Domain entities (campaign/technique/threat_actor/test_entity): replaced Any with
  actual ORM types via TYPE_CHECKING guards to avoid circular imports
- detection_rule_service: test_id/detection_rule_id/evaluator_id→UUID
- score_cache: kept Any with # noqa: ANN401 (genuinely generic cache)
- jira_service/tempo_service: kept Any with # noqa: ANN401 (lazy optional deps)
- d3fend_import_service: _to_str(v: Any) kept with # noqa: ANN401

ANN204/ANN205/ANN206 — special/static/class methods:
- database.py proxy __call__/__getattr__: *args: object/**kwargs: object
- schemas/test.py model_validate: obj→object, **kwargs→object
- sa_technique_repository._int_type→type

All 439 unit tests pass. ruff check app/ → All checks passed!

backend/app/routers/heatmap.py

@@ -28,7 +28,7 @@ def heatmap_coverage(
     min_score: int = Query(0, ge=0, le=100),
     db: Session = Depends(get_db),
     current_user: User = Depends(get_current_user),
-):
+) -> dict:
     """Coverage layer — score based on status_global of each technique."""
     return heatmap_service.build_coverage_layer(
         db, platforms=platforms, tactics=tactics, min_score=min_score,
@@ -43,7 +43,7 @@ def heatmap_threat_actor(
     min_score: int = Query(0, ge=0, le=100),
     db: Session = Depends(get_db),
     current_user: User = Depends(get_current_user),
-):
+) -> dict:
     """Threat actor layer — techniques used by an actor with coverage color."""
     return heatmap_service.build_threat_actor_layer(
         db, actor_id, platforms=platforms, tactics=tactics, min_score=min_score,
@@ -57,7 +57,7 @@ def heatmap_detection_rules(
     min_score: int = Query(0, ge=0, le=100),
     db: Session = Depends(get_db),
     current_user: User = Depends(get_current_user),
-):
+) -> dict:
     """Detection rules layer — score based on ratio of rules available vs total."""
     return heatmap_service.build_detection_rules_layer(
         db, platforms=platforms, tactics=tactics, min_score=min_score,
@@ -72,7 +72,7 @@ def heatmap_campaign(
     min_score: int = Query(0, ge=0, le=100),
     db: Session = Depends(get_db),
     current_user: User = Depends(get_current_user),
-):
+) -> dict:
     """Campaign layer — only techniques in the campaign, colored by test state."""
     return heatmap_service.build_campaign_layer(
         db, campaign_id, platforms=platforms, tactics=tactics, min_score=min_score,
@@ -88,7 +88,7 @@ def export_navigator(
     min_score: int = Query(0, ge=0, le=100),
     db: Session = Depends(get_db),
     current_user: User = Depends(get_current_user),
-):
+) -> StreamingResponse:
     """Export a heatmap layer as a downloadable JSON file for ATT&CK Navigator."""
     data = heatmap_service.build_navigator_export(
         db, layer, layer_id=layer_id,