Build smart recommendation engine

[SH20RAJ]

🎯 Goal

Implement intelligent recommendations based on usage patterns and cleanup history.

🚀 Features

• Analyze cleanup history to find patterns
• Identify fast-growing categories
• Suggest optimal cleanup frequency
• Provide actionable recommendations
• Calculate priority levels (HIGH/MEDIUM/LOW)

📊 Example Output

💡 Smart Recommendations (based on your usage)

HIGH PRIORITY (Do Now)
├─ Xcode DerivedData: 12GB (grows 2GB/week on average)
│  └─ Recommendation: Clean now, then schedule weekly cleanup
│
└─ node_modules: 8.2GB across 12 projects
   └─ Recommendation: Consider switching to pnpm or yarn workspaces

MEDIUM PRIORITY (This Week)
├─ Docker: 4.2GB unused images
│  └─ 23 images haven't been used in 30+ days

LOW PRIORITY (Optional)
└─ npm cache: 3.5GB
   └─ Last used 2 days ago, can wait

🔧 Implementation Details

• Track cleanup history in ~/.cleanmac/history.json
• Calculate growth rates per category
• Analyze access patterns
• Generate contextual recommendations
• Learn from user behavior

✅ Acceptance Criteria

• Recommendations are accurate
• Priority levels make sense
• Growth rate calculations are correct
• Users can enable/disable suggestions

🔗 Related

Part of Phase 4