Refactor Flow Management to be User-Scoped and Database-Backed #71
Description
The current FlowfileHandler is a global, in-memory object, meaning all users share the same flows and all data is lost on restart. This issue tracks the work to replace it with a robust, database-driven system.
- Create DB Schema: Define new SQLAlchemy models for Flow and Node. Each model must contain a user_id foreign key that links to the User model, ensuring a clear ownership record for every piece of data.
- Implement a Service Layer: Develop a new UserFlowService class. This class will be instantiated for each API request and will contain the authenticated user's ID. All methods within this service (get_flow, create_node, etc.) must use the user_id to scope database queries, guaranteeing that a user can only access their own data.
- Refactor API Endpoints: Update all API endpoints that currently depend on the global flow_file_handler. Use FastAPI's dependency injection to provide the UserFlowService to these endpoints. This ensures that the application logic is automatically operating within the context of the correct user.