feat: FASE 1 - Backend Core (modelli, auth, API)

Implementazione completa del backend FastAPI:
- Modelli SQLAlchemy: User, Recipe, RecipeVersion, RecipeTask,
  RecipeSubtask, Measurement, AccessLog, SystemSetting, RecipeVersionAudit
- Schemas Pydantic v2 per tutti i CRUD + statistiche SPC
- Middleware: API Key auth (X-API-Key) con role checking + access logging
- Router: auth, users, recipes, tasks, measurements, files, settings
- Services: auth (bcrypt+secrets), recipe (copy-on-write versioning),
  measurement (auto pass/fail con UTL/UWL/LWL/LTL)
- Alembic env.py con import modelli attivi
- Fix architect review: no double-commit, recipe_id subquery filter,
  user_id in access logs, type annotations corrette

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

server/models/__init__.py

@@ -0,0 +1,19 @@
+"""SQLAlchemy models for TieMeasureFlow."""
+from models.user import User
+from models.recipe import Recipe, RecipeVersion
+from models.task import RecipeTask, RecipeSubtask
+from models.measurement import Measurement
+from models.access_log import AccessLog
+from models.setting import SystemSetting, RecipeVersionAudit
+
+__all__ = [
+    "User",
+    "Recipe",
+    "RecipeVersion",
+    "RecipeTask",
+    "RecipeSubtask",
+    "Measurement",
+    "AccessLog",
+    "SystemSetting",
+    "RecipeVersionAudit",
+]