Katkat
Back

Implement Task Management and Duplicate Detection

AnthropicSoftware Engineer · Technical Screen
MediumSoftware EngineerTechnical ScreenCoding & Algorithms

Last updated

This problem tests your ability to design an in-memory task management system with concurrency-aware state tracking and implement an efficient file-deduplication routine using hashing.

Part 1: Task Management Service

Implement a TaskManager class to manage tasks for multiple users. Each task has an id, user_id, priority (1-10), and status ('pending', 'completed').

  • add_task(user_id, priority): Assign a unique ID and store the task. Enforce a per-user quota of 5 active (non-completed) tasks.
  • list_tasks(user_id): Return tasks sorted by priority (descending).
  • get_history(task_id): Return an audit log of status changes for a given task.

Part 2: File Deduplication

You are given a root directory path. Implement a function to find duplicate files based on content. Two files are duplicates if they have identical contents. Assume you have access to a helper function compute_hash(file_path) that returns a string.

Example

# Task Management
tm = TaskManager()
tm.add_task(user_1, 5)
tm.list_tasks(user_1) # Returns list of tasks for user_1 sorted by priority

# Deduplication
# Input: '/root_dir'
# Output: [['/root_dir/a.txt', '/root_dir/sub/b.txt'], ['/root_dir/c.png']]
# Explanation: Files with identical hashes are grouped together.

Constraints

  • TaskManager: Perform operations in O(log N) or better where N is the number of tasks.
  • Deduplication: Optimize for memory; do not read the entire file into memory if it is large.

Hints

Hint 1

For the TaskManager, use a dictionary mapping user_ids to a min-heap or a sorted list to maintain priority order and active task counts.

Hint 2

For file deduplication, use a hash map (dictionary) where keys are file hashes and values are lists of file paths.

Hint 3

To optimize deduplication for large files, only compute hashes for files that share the same file size, as files with different sizes cannot be duplicates.

Follow-up questions

  • How would you make the TaskManager thread-safe in a multi-threaded environment?
  • How would you modify the deduplication routine to handle symbolic links without getting stuck in infinite loops?
  • If the task history grows very large, how would you optimize storage for `get_history`?