refactor: replace universal_migration.py with Alembic

Remove the 7,791-line universal_migration.py and 16 incomplete individual
Alembic migrations. Replace with a single initial schema migration using
Base.metadata.create_all(checkfirst=True).

Changes:
- Add programmatic Alembic runner (app/database/migrations.py) with
  auto-stamp logic for existing databases transitioning from
  universal_migration
- Extract ensure_default_web_api_token() to web_api_token_service.py
- Extract sync_postgres_sequences() to database.py with SQL injection
  prevention via _quote_ident()
- Add HMAC token hashing support with backward-compatible dual-hash
  fallback and automatic rehashing
- Remove dead init_db() function and unused imports
- Add Makefile targets: migrate, migration, migrate-stamp, migrate-history
- Fix fileConfig() destroying structlog config (disable_existing_loggers)
- Remove duplicate migrations/alembic/alembic.ini with credentials
- Add script.py.mako template for future migration generation
- Update startup flow: alembic upgrade → sync sequences → ensure token
- Harden database.py: ParamSpec for retry decorator, safe URL logging,
  echo='debug' mode, execute_with_retry validation
- Update documentation references

31 files changed, 302 insertions(+), 9,226 deletions(-)

Makefile

@@ -40,6 +40,22 @@ fix: ## Исправить код (ruff check --fix + format)
 	uv run ruff check . --fix
 	uv run ruff format .
 
+.PHONY: migrate
+migrate: ## Применить миграции (alembic upgrade head)
+	uv run alembic upgrade head
+
+.PHONY: migration
+migration: ## Создать миграцию (usage: make migration m="description")
+	uv run alembic revision --autogenerate -m "$(m)"
+
+.PHONY: migrate-stamp
+migrate-stamp: ## Пометить БД как актуальную (для существующих БД)
+	uv run alembic stamp head
+
+.PHONY: migrate-history
+migrate-history: ## Показать историю миграций
+	uv run alembic history --verbose
+
 .PHONY: help
 help: ## Показать список доступных команд
 	@echo ""