InstaFeed

A simplified Instagram-like feed for iOS demonstrating dynamic mixed media handling, optimized scrolling performance, and proper caching strategies.

Assignment Requirements

Requirement	Implementation
Scrollable feed with posts	✅ List for efficient cell recycling
Single photo posts	✅ SingleMediaView with natural aspect ratio
Single video posts	✅ VideoPlayerView with auto-play support
Mixed photo + video posts	✅ TabView carousel with PageTabViewStyle
Auto-play videos when visible	✅ Visibility tracking via GeometryReader + PreferenceKey
Pause videos when scrolled out	✅ Hysteresis-based visibility algorithm in FeedViewModel
Caching mechanisms	✅ ImageCacheActor and VideoCacheActor using Swift actors
External data source	✅ Mock data from picsum.photos, pravatar.cc, and Google sample videos

Technical Stack

• iOS: SwiftUI
• Architecture: MVVM (Model-View-ViewModel)
• Concurrency: Swift Concurrency (async/await, actor)
• Video Playback: AVFoundation / AVKit

Architecture Overview

┌─────────────────────────────────────────────────────────────┐
│                        InstaFeedApp                         │
│                   (DependencyContainer)                     │
└─────────────────────────────────────────────────────────────┘
                              │
                    @Environment(\.dependencies)
                              │
┌─────────────────────────────────────────────────────────────┐
│                          FeedView                           │
│                       + FeedViewModel                       │
│              (visibility tracking, scroll logic)            │
└─────────────────────────────────────────────────────────────┘
                              │
┌─────────────────────────────────────────────────────────────┐
│                          PostView                           │
│                       + PostViewModel                       │
│            (PostHeaderView, PostMediaContentView,           │
│             PostActionsView, PostCaptionView)               │
└─────────────────────────────────────────────────────────────┘
                              │
           ┌──────────────────┴──────────────────┐
           │                                     │
┌─────────────────────┐             ┌─────────────────────┐
│  CachedAsyncImage   │             │   VideoPlayerView   │
│ + ImageLoaderService│             │ + VideoPlayerVM     │
└─────────────────────┘             └─────────────────────┘
           │                                     │
┌─────────────────────┐             ┌─────────────────────┐
│  ImageCacheActor    │             │  VideoCacheActor    │
│   (thread-safe)     │             │   (thread-safe)     │
└─────────────────────┘             └─────────────────────┘

Key Features

1. Scrollable Feed with Cell Recycling

• Uses List for efficient memory management through cell recycling
• Handles hundreds of posts without memory bloat

2. Mixed Media Support

• Single Image: Displays with natural aspect ratio
• Single Video: Auto-plays with natural aspect ratio
• Carousel: TabView with page indicator for multiple items

3. Video Auto-Play/Pause

• Tracks visibility using GeometryReader and PreferenceKey
• Hysteresis algorithm: Videos keep playing until < 30% visible, new videos start at > 50% visible
• Prevents flickering during slow scrolling

4. Caching Strategy

Cache	Type	Purpose
ImageCacheActor	actor	Thread-safe in-memory image cache with LRU eviction
VideoCacheActor	actor	Thread-safe AVAsset cache to avoid reloading metadata
VideoPlayerCacheActor	actor	Reuses AVPlayer instances per URL

5. Dependency Injection

• No singletons — all services created in DependencyContainer
• Injected via @Environment(\.dependencies)

How to Run

1. Open InstaFeed.xcodeproj in Xcode
2. Select the InstaFeed scheme
3. Choose a simulator or device
4. Press Cmd + R to build and run

Note: Internet connection required for loading mock media from external URLs.

Data Sources

The app uses external services for mock data:

• Images: picsum.photos (Lorem Picsum)
• Avatars: pravatar.cc (Random avatars)
• Videos: Google sample videos (Big Buck Bunny, Elephants Dream)

Future Improvements

• Disk caching for offline support
• Pull-to-refresh
• Infinite scroll / pagination
• Like/comment interactions
• Local video/image assets as alternative data source