Problem Overview

Implement a ToDo List system for AI agents that manages tasks with various states and handles task dependencies. The problem starts simple but progressively adds complexity through multiple parts.

Given structure:

• TaskStatus (Enum): States like READY, IN_PROGRESS, SUCCEEDED, FAILED, BLOCKED

• Task (Class): Represents a single task

• ToDoList (Class): Manages the collection of tasks

Key challenges:

• Regression Testing: Each new part has independent test cases. Your code modifications must NOT break previous parts.

• Code Volume: High implementation speed is required.

• State Management: Properly handle cascading state updates through dependency chains.

Important Notes

• There are multiple parts to this problem - ask the interviewer how many parts there are to better manage your time

• Each part builds on the previous one while maintaining backward compatibility

• Write your own test cases and ensure your code compiles and runs correctly

Part 1: Basic CRUD Operations

Implement the fundamental operations for task management.

Requirements

add_task(task_id: str, description: str) -> Task

• Create a new task with the given ID and description

• Retrieve a task by its ID

• Raise exception if task does not exist

update_status(task_id: str, new_status: TaskStatus) -> None

• Update the status of an existing task

• Raise exception if task does not exist

` todo_list = ToDoList()

Add a task

task1 = todo_list.add_task("task1", "First task") print(task1.status) # TaskStatus.READY

Get a task

retrieved = todo_list.get_task("task1") print(retrieved.task_id) # "task1"

Update status

todo_list.update_status("task1", TaskStatus.IN_PROGRESS) todo_list.update_status("task1", TaskStatus.SUCCEEDED) print(todo_list.get_task("task1").status) # TaskStatus.SUCCEEDED `

• How should we handle same-task dependencies (task depending on itself)?

• Should we validate state transitions (e.g., cannot go from SUCCEEDED to READY)?

• What happens if we try to add a duplicate task ID?

Add support for task dependencies with cascading state management.

When updating a task's status, trigger cascading updates: