Problem Statement

We are developing a movie content management system - ZipReel for a streaming platform, where users can search and access movie information efficiently. The system needs to handle frequent searches and provide fast responses through a multi-level caching mechanism.

As an engineer, your task is to build a feature that allows:

1. Registration of movies and users in the system
2. Implementation of a multi-level cache to optimize search operations
3. Efficient retrieval of movie information based on various parameters

This system aims to provide lightning-fast search results via intelligent caching.

Explanations

What is the multi-level cache?

• L1 Cache: User-specific recent searches (max 5 entries per user)
• L2 Cache: Global popular searches (max 20 entries)
• Primary Store: All movie data in-memory

What is a cache hit/miss?

• When searching, system checks caches in order: L1 -> L2 -> Primary Store
• Cache hit: Data found in any cache level
• Cache miss: Data not found, requiring search in base data

Requirements

System should be able to:

1. Add movies with following attributes:
   • Movie ID (unique)
   • Title
   • Genre
   • Release Year
   • Rating
2. Register users with:
   • User ID (unique)
   • Name
   • Preferred Genre
3. Implement multi-level caching:
   • L1: User-specific cache
     • Maximum 5 entries per user
     • Implements LRU (Least Recently Used) eviction
   • L2: Global popular searches cache
     • Maximum 20 entries
     • Implements LFU (Least Frequently Used) eviction
   • Primary Store: Complete movie database in memory
4. Handle search operations:
   • Search by title
   • Search by genre
   • Search by release year
   • Multiple filter combination

Search Process:

1. Check L1 cache (user-specific)
2. If not found, check L2 cache (global popular)
3. If not found, search in Primary Store (main database)
4. Update caches based on search result

Bonus Requirement:

System should be able to view cache level wise cache hits i.e for L1, L2 and Primary Store and should display the total number of searches in the system.

Commands

ADD_MOVIE <id> <title> <genre> <year> <rating>

Example: ADD_MOVIE 1 "Inception" "Sci-Fi" 2010 9.5

Output: Movie 'Inception' added successfully

ADD_USER <id> <name> <preferred_genre>

Example: ADD_USER 1 "John" "Action"

Output: User 'John' added successfully

SEARCH <user_id> <search_type> <search_value>

Example: SEARCH 1 GENRE "Action"

Output: [List of movies with cache level indicator] Format: {movie_title} (Found in {cache_level})

SEARCH_MULTI <user_id> <genre> <year> <min_rating>

Example: SEARCH_MULTI 1 "Action" 2020 8.0

Output: [List of filtered movies with cache level indicator]

VIEW_CACHE_STATS

Example: VIEW_CACHE_STATS

Output:

L1 Cache Hits: X
L2 Cache Hits: Y
Primary Store Hits: Z
Total Searches: N

CLEAR_CACHE <cache_level>

Example: CLEAR_CACHE L1

Output: L1 cache cleared successfully

Guidelines:

1. Input can be read from a file or STDIN or coded in a driver method. [No API and No UI]
2. Output can be written to a file or STDOUT. [No API]
3. Store all interim/output data in-memory data structures. The usage of databases is not allowed.
4. Language should be Java only.

Expectations:

1. The code should be demo-able and functionally correct
2. Proper exception handling for edge cases:
   • Invalid movie/user IDs
   • Duplicate entries
   • Invalid search parameters
3. Code should be modular and follow OOP principles
4. Efficient implementation of cache eviction policies
5. Clear separation of concerns between:
   • Data storage
   • Cache management
   • Search operations
6. Proper unit tests for critical components are good to have.

What are the ways we can approach this problem. Design Patterns we can use and why?