Add configuration rules, refactor settings management, and enhance import functionality

- Introduced a new configuration file `.cursorrules` to define coding standards, error handling, testing requirements, and project-specific guidelines.
- Refactored `config.py` to implement a `Settings` dataclass for better management of environment variables, improving testability and maintainability.
- Updated the import duty schedule handler to utilize session management with `session_scope`, ensuring proper database session handling.
- Enhanced the import service to streamline the duty schedule import process, improving code organization and readability.
- Added new service layer functions to encapsulate business logic related to group duty pinning and duty schedule imports.
- Updated README documentation to reflect the new configuration structure and improved import functionality.

db/session.py

@@ -1,9 +1,11 @@
 """SQLAlchemy engine and session factory.
 
-Note: Engine and session factory are cached globally per process. Only one
-DATABASE_URL is effectively used for the process lifetime. Using a different
-URL later (e.g. in tests with in-memory SQLite) would still use the first
-engine. To support multiple URLs, cache by database_url (e.g. a dict keyed by URL).
+Engine and session factory are cached globally per process. Only one DATABASE_URL
+is effectively used for the process lifetime. Using a different URL later (e.g. in
+tests with in-memory SQLite) would still use the first engine. To use a different
+URL in tests, set env (e.g. DATABASE_URL) before the first import of this module, or
+clear _engine and _SessionLocal in test fixtures. Prefer session_scope() for all
+callers so sessions are always closed and rolled back on error.
 """
 
 from contextlib import contextmanager