Image Carousel Backend with Threaded Comments and Upvotes

Context

Design and implement the backend for an image carousel feature in a web app. Users can view images, add threaded comments per image, and upvote both images and comments. Use in-memory data structures for an efficient interview-ready solution. Assume a single-process service with concurrent requests (e.g., async HTTP server). Authentication exists and provides a stable userId.

Requirements

• Data modeling for:
  • Images
  • Comments (threaded: parent/child)
  • Upvotes (for both images and comments)
• RESTful endpoints to:
  1. Post a comment (top-level or reply)
  2. Upvote an image or a comment (idempotent)
  3. Remove an upvote (idempotent)
  4. Fetch the carousel with aggregated counts and paginated comments
• Behavior and guarantees:
  • Idempotency: repeated upvote/un-upvote requests must not double-count and should be safe to retry
  • Concurrency-safety for upvotes and comment creation
  • Pagination for images (carousel) and comments (paginate top-level comments per image)
  • Sorting by time or score (score = upvote count; tie-breaker by time)
  • Basic rate limiting per user/IP
• Deliverables:
  • Define data models
  • Specify REST endpoints with request/response shape
  • Choose in-memory data structures
  • Provide core functions (pseudo-code is fine) ensuring idempotency and concurrency-safety
  • Explain time and space complexity

Assumptions

• Single-process node handling concurrent requests; use per-key mutexes for critical sections
• Time fields are ISO timestamps; IDs are UUIDs
• Comments: paginate top-level comments only; return full nested replies for each top-level comment page
• Sorting options: time (newest first) or score (upvotes desc; break ties by newest)
• Rate limit: example 60 requests/min per userId (fallback to IP if unauthenticated)